article
stringlengths 0
745k
| abstract
stringlengths 0
92.7k
|
|---|---|
let @xmath6 be an elliptic curve over a finite field @xmath7 , and let @xmath8 be a cyclic subgroup of prime order @xmath9 .
when implementing cryptographic protocols using @xmath10 , the fundamental operation is _
scalar multiplication _ ( or _ exponentiation _ ) : given @xmath11 in @xmath10 and @xmath12 in @xmath13 , compute @xmath14p : = \underbrace{p\oplus\cdots\oplus p}_{m \text { times } } $ ] .
the literature on general scalar multiplication algorithms is vast , and we will not explore it in detail here ( see ( * ? ? ? * , 11.2 ) and ( * ? ? ?
* chapter 9 ) for introductions to exponentiation and multiexponentiation algorithms ) . for our purposes
, it suffices to note that the dominant factor in scalar multiplication time using conventional algorithms is the bitlength of @xmath12 . as a basic example , if @xmath10 is a generic cyclic abelian group , then we may compute @xmath14p$ ] using a variant of the binary method , which requires at most @xmath15 doublings and ( in the worst case ) about as many addings in @xmath10 .
but elliptic curves are not generic groups : they have a rich and concrete geometric structure , which should be exploited for fun and profit .
for example , endomorphisms of elliptic curves may be used to accelerate generic scalar multiplication algorithms , and thus to accelerate basic operations in curve - based cryptosystems .
suppose @xmath6 is equipped with an efficient endomorphism @xmath16 , defined over @xmath7 . by _
efficient _ , we mean that we can compute the image @xmath17 of any point @xmath11 in @xmath18 for the cost of @xmath19 operations in @xmath7 . in practice
, we want this to cost no more than a few doublings in @xmath18 .
assume @xmath20 , or equivalently , that @xmath16 restricts to an endomorphism of @xmath10 .
, then @xmath21({\mathbb{f}}_{q } ) = { \mathcal{g}}+\psi({\mathcal{g } } ) \cong ( { \mathbb{z}}/n{\mathbb{z}})^2
$ ] , so @xmath22 and @xmath23 ; such @xmath6 are cryptographically inefficient , and discrete logs in @xmath10 are vulnerable to the menezes okamoto vanstone reduction @xcite .
however , these @xmath10 do arise naturally in pairing - based cryptography ; in that context the assumption should be verified carefully .
] now @xmath10 is a finite cyclic group , isomorphic to @xmath24 ; and every endomorphism of @xmath24 is just an integer multiplication modulo @xmath9 .
hence , @xmath16 acts on @xmath10 as multiplication by some integer eigenvalue @xmath25 : that is , @xmath26_{\mathcal{g}}.\ ] ] the eigenvalue @xmath25 must be a root in @xmath24 of the characteristic polynomial of @xmath16 . returning to the problem of scalar multiplication : we want to compute @xmath14p$ ] . rewriting @xmath12 as @xmath27 for some @xmath28 and @xmath29
, we can compute @xmath14p$ ] using the relation @xmath30p = [ a]p + [ b\lambda_\psi]p = [ a]p + [ b]\psi(p)\ ] ] and a two - dimensional multiexponentation such as straus s algorithm @xcite , which requires has a loop length of @xmath31 ( ie , @xmath31 doubles and as many adds ; recall that @xmath32 ) .
if @xmath25 is not too small , then we can easily find @xmath33 such that @xmath31 is roughly half of @xmath34 .
( we remove the `` if '' and the `` roughly '' for our @xmath16 in [ sec : decompositions ] . )
the endomorphism lets us replace conventional @xmath34-bit scalar multiplications with @xmath35-bit multiexponentiations . in terms of basic binary methods , we are halving the loop length , cutting the number of doublings in half .
of course , in practice we are not halving the execution time .
the precise speedup ratio depends on a variety of factors , including the choice of exponentiation and multiexponentiation algorithms , the cost of computing @xmath16 , the shortness of @xmath28 and @xmath29 on the average , and the cost of doublings and addings in terms of bit operations to say nothing of the cryptographic protocol , which may prohibit some other conventional speedups .
for example : in @xcite , galbraith , lin , and scott report experiments where cryptographic operations on gls curves required between 70% and 83% of the time required for the previous best practice curves with the variation depending on the architecture , the underyling point arithmetic , and the protocol . to put this technique into practice
, we need a source of cryptographic elliptic curves equipped with efficient endomorphisms . to date , in the large characteristic case or @xmath36
; so we will not discuss @xmath37-adic / frobenius expansion - style techniques here . ] , there have been essentially only two constructions : 1 .
the classic _ gallant lambert vanstone _ ( glv ) construction @xcite .
elliptic curves over number fields with explicit complex multiplication by cm - orders with small discriminants are reduced modulo suitable primes @xmath3 ; an explicit endomorphism on the cm curve reduces to an efficient endomorphism over the finite field .
the more recent _ galbraith lin
scott _ ( gls ) construction @xcite .
here , curves over @xmath38 are viewed over @xmath0 ; the @xmath3-power sub - frobenius induces an extremely efficient endomorphism on the quadratic twist ( which can have prime order ) .
these two constructions have since been combined to give 3- and 4-dimensional variants @xcite , and extended to hyperelliptic curves in a variety of ways @xcite .
however , basic glv and gls remain the archetypal constructions . in this work ,
we propose a new source of elliptic curves over @xmath0 with efficient endomorphisms : quadratic @xmath1-curves .
[ def : qq - curves ] a _
quadratic @xmath1-curve of degree @xmath39 _ is an elliptic curve @xmath6 _ without _ complex multiplication , defined over a quadratic number field @xmath40 , such that there exists an isogeny of degree @xmath39 from @xmath6 to its galois conjugate @xmath41 , where @xmath42 .
( the galois conjugate @xmath41 is the curve formed by applying @xmath43 to all of the coefficients of @xmath6 . )
@xmath1-curves are well - established objects of interest in number theory , where they have formed a natural setting for generalizations of the shimura taniyama conjecture .
survey @xcite gives an excellent introduction to this beautiful theory .
our application of quadratic @xmath1-curves is rather more prosaic : given a @xmath39-isogeny @xmath44 over a quadratic field , we reduce modulo an inert prime @xmath3 to obtain an isogeny @xmath45 over @xmath0 .
we then exploit the fact that the @xmath3-power frobenius isogeny maps @xmath41 back onto @xmath6 ; composing with the reduced @xmath39-isogeny , we obtain an endomorphism of @xmath6 of degree @xmath46 . for efficiency reasons
, @xmath39 must be small ; it turns out that for small values of @xmath39 , we can write down one - parameter families of @xmath1-curves ( our approach below was inspired by the explicit techniques of hasegawa @xcite ) .
we thus obtain one - parameter families of elliptic curves over @xmath0 equipped with efficient non - integer endomorphisms . for these endomorphisms we can give convenient explicit formul for short scalar decompositions ( see [ sec : decompositions ] ) . for concrete examples , we concentrate on the cases @xmath47 and @xmath48 ( in [ sec : degree-2 ] and [ sec : degree-3 ] , respectively ) , where the endomorphism is more efficient than a single doubling ( we briefly discuss higher degrees in [ sec : further ] ) . for maximum generality and flexibility , we define our curves in short weierstrass form ; but we include transformations to montgomery , twisted edwards , and doche icart kohel models where appropriate in [ sec : models ] . like glv ,
our method involves reducing curves defined over number fields to obtain curves over finite fields with explicit complex multiplication .
however , we emphasise a profound difference : in our method , the curves over the number fields generally _ do not have complex multiplication themselves_. glv curves are necessarily isolated examples and the really useful examples are extremely limited in number ( see ( * ? ? ? * app .
a ) for a list of curves ) .
the scarcity of glv curves is the kronecker class number @xmath49 , which is in @xmath50 .
of course , for the tiny values of @xmath51 in question , the asymptotics of @xmath49 are irrelevant ; for the six @xmath51 corresponding to endomorphisms of degree at most 3 , we have @xmath52 , so there is only one @xmath53-invariant . for @xmath54 ( corresponding to @xmath55 )
there are two or four twists over @xmath38 ; for @xmath56 ( corresponding to @xmath57 ) we have two or six , and otherwise we have only two . in particular , there are at most 18 distinct curves over @xmath38 with a non - integer endomorphism of degree at most 3 . ]
is their achilles heel : as noted in @xcite , if @xmath3 is fixed then there is no guarantee that there will exist a glv curve with prime ( or almost - prime ) order over @xmath58 .
consider the situation discussed in @xcite : the most efficient glv curves have cm discriminants @xmath59 and @xmath60 .
if we are working at a 128-bit security level , then the choice @xmath5 allows particularly fast arithmetic in @xmath38 .
but the largest prime factor of the order of a curve over @xmath38 with cm discriminant @xmath60 ( resp .
@xmath59 ) has @xmath61 ( resp .
@xmath62 ) bits : using these curves wastes 9 ( resp .
13 ) potential bits of security .
in fact , we are lucky with @xmath56 and @xmath60 : for all of the other discriminants offering endomorphisms of degree at most 3 , we can do no better than a 95-bit prime factor , which represents a catastrophic 80-bit loss of relative security .
in contrast , our construction yields true families of curves , covering @xmath63 isomorphism classes over @xmath0 .
this gives us a vastly higher probability of finding prime ( or almost - prime)-order curves over practically important fields . like gls
, we construct curves over @xmath0 equipped with an inseparable endomorphism .
while these curves are not defined over the prime field , the fact that the extension degree is only 2 means that weil descent attacks offer no advantage when solving dlp instances ( see @xcite ) . and like gls , our families offer around @xmath3 distinct isomorphism classes of curves , making it easy to find secure group orders when @xmath3 is fixed .
but unlike gls , our curves have @xmath53-invariants in @xmath0 : they are not isomorphic to or twists of subfield curves .
this allows us to find twist - secure curves , which are resistant to the fouque
lercier ral valette fault attack @xcite .
as we will see in [ sec : gls ] , our construction reduces to gls in the degenerate case @xmath64 ( that is , where @xmath65 is an isomorphism ) .
our construction is therefore a sort of generalized gls though it is not the higher - degree generalization anticipated by galbraith , lin , and scott themselves , which composes the sub - frobenius with a non - rational separable homomorphism and its dual homomorphism ( cf .
* theorem 1 ) ) . in [ sec : decompositions ] , we prove that we can immediately obtain decompositions of the same bitlength as gls for curves over the same fields : the decompositions produced by our proposition [ prop : bitlength ] are identical to the gls decompositions of ( * ? ? ?
* lemma 2 ) when @xmath64 , up to sign .
for this reason , we do not provide extensive implementation details in this paper : while our endomorphisms cost a few more @xmath7-operations to evaluate than the gls endomorphism , this evaluation is typically carried out only once per scalar multiplication .
this evaluation is the only difference between a gls scalar multiplication and one of ours : the subsequent multiexponentiations have exactly the same length as in gls , and the underlying curve and field arithmetic is the same , too .
throughout , we work over fields of characteristic not 2 or 3 .
let @xmath66 be an elliptic curve over such a field @xmath40 . for every automorphism @xmath43 of @xmath40
, we define the conjugate curve @xmath67 if @xmath68 is an isogeny , then we obtain a conjugate isogeny @xmath69 by applying @xmath43 to the defining equations of @xmath70 , @xmath6 , and @xmath71 . for every @xmath72 in @xmath73 , we define a twisting isomorphism @xmath74 by @xmath75 the twist @xmath76 is defined over @xmath77 , and @xmath78 is defined over @xmath79 . for every @xmath40-endomorphism @xmath16 of @xmath6 , there is a @xmath77-endomorphism @xmath80 of @xmath76 . observe that @xmath81 for any @xmath82 in @xmath40 , and @xmath83 $ ] .
also , @xmath84 for all automorphisms @xmath43 of @xmath85 .
if @xmath86 is a nonsquare in @xmath40 , then @xmath87 is called a _
quadratic twist_. if @xmath88 , then @xmath89 and @xmath90 are @xmath7-isomorphic for all nonsquares @xmath91 , @xmath92 in @xmath7 ( the isomorphism @xmath93 is defined over @xmath7 because @xmath94 must be a square ) .
when the choice of nonsquare is not important , we denote the quadratic twist by @xmath95 .
similarly , if @xmath16 is an @xmath7-endomorphism of @xmath6 , then @xmath96 is the corresponding @xmath7-endomorphism of @xmath95 .
( _ conjugates are marked by left - superscripts , twists by right - superscripts_. ) if @xmath88 , then @xmath97 denotes the @xmath98-power frobenius endomorphism of @xmath6 . recall that the characteristic polynomial of @xmath97 has the form @xmath99 the integer @xmath100 is the _ trace _ of @xmath6 ; we have @xmath101 and @xmath102 .
we write @xmath103 for the @xmath3-th powering automorphism of @xmath104 .
note that @xmath103 is almost trivial to compute on @xmath105 , because @xmath106 for all @xmath28 and @xmath29 in @xmath58 .
suppose @xmath107 is a quadratic @xmath1-curve of prime degree @xmath39 ( cf .
definition [ def : qq - curves ] ) , where @xmath108 is a discriminant prime to @xmath39 , and let @xmath109 be the corresponding @xmath39-isogeny . in general , @xmath65 is only defined over a quadratic extension @xmath110 of @xmath111 .
we can compute @xmath112 from @xmath108 and @xmath113 using ( * ? ? ?
* proposition 3.1 ) , but after a suitable twist we can always reduce to the case where @xmath114 ( see ( * ? ? ? * remark after lemma 3.2 ) ) .
the families of explicit @xmath1-curves of degree @xmath39 that we treat below have their isogenies defined over @xmath115 ; so to simplify matters , from now on we will assume @xmath65 is defined over @xmath115 .
let @xmath3 be a prime of good reduction for @xmath116 that is inert in @xmath111 and prime to @xmath39 .
if @xmath117 is the ring of integers of @xmath111 , then @xmath118 looking at the galois groups of our fields , we have a series of injections @xmath119 the image of @xmath103 in @xmath120 is @xmath43 , because @xmath3 is inert in @xmath111 . when extending @xmath43 to an automorphism of @xmath115 , we extend it to be the image of @xmath103 : that is , @xmath121 for all @xmath122 , and @xmath123 .
( recall that the legendre symbol @xmath124 is @xmath125 if @xmath126 is a square mod @xmath3 , @xmath127 if @xmath126 is not a square mod @xmath3 , and @xmath128 if @xmath3 divides @xmath126 . )
now let @xmath6 be the reduction modulo @xmath3 of @xmath116 .
the curve @xmath129 reduces to @xmath130 , while the @xmath39-isogeny @xmath131 reduces to a @xmath39-isogeny @xmath132 defined over @xmath0 .
applying @xmath43 to @xmath65 , we obtain a second @xmath39-isogeny @xmath133 travelling in the opposite direction , which reduces mod @xmath3 to a conjugate isogeny @xmath134 over @xmath0 .
composing @xmath135 with @xmath65 yields endomorphisms @xmath136 of @xmath116 and @xmath137 of @xmath129 , each of degree @xmath138 . but
( by definition ) @xmath116 and @xmath129 do not have complex multiplication , so all of their endomorphisms are integer multiplications ; and since the only integer multiplications of degree @xmath138 are @xmath139 $ ] and @xmath140 $ ] , we can conclude that @xmath141_{{\widetilde{\mathcal{e } } } } \quad \text { and } \quad { \widetilde{\phi}}\circ{{{}^{\sigma}\!\!{{\widetilde{\phi } } } } } = [ \epsilon_p d]_{{{{}^{\sigma}\!\!{{\widetilde{\mathcal{e } } } } } } } , \quad \text { where } \epsilon_p \in \{\pm1\ } .\ ] ] technically , @xmath135 and @xmath142 are_up to sign_the dual isogenies of @xmath65 and @xmath70 , respectively .
the sign @xmath143 depends on @xmath3 ( as well as on @xmath65 ) : if @xmath37 is the extension of @xmath43 to @xmath115 that is _ not _ the image of @xmath103 , then @xmath144_{{\widetilde{\mathcal{e}}}}$ ] .
reducing mod @xmath3 , we see that @xmath145_{{\mathcal{e } } } \quad \text { and } \quad \phi\circ{{{}^{(p)}\!\!{\phi } } } = [ \epsilon_p d]_{{{{}^{(p)}\!\!{{\mathcal{e } } } } } } .\ ] ] the map @xmath146 defines @xmath3-isogenies @xmath147 clearly , @xmath148 ( resp .
@xmath149 ) is the @xmath36-power frobenius endomorphism of @xmath6 ( resp .
@xmath130 ) .
composing @xmath150 with @xmath70 yields a degree-@xmath151 endomorphism @xmath152 if @xmath39 is very small say , less than @xmath153then @xmath16 is efficient because @xmath70 is defined by polynomials of degree about @xmath39 , and @xmath150 acts as a simple conjugation on coordinates in @xmath0 , as in eq . .
( the efficiency of @xmath16 depends primarily on its separable degree , @xmath39 , and not on the inseparable part @xmath3 . )
we also obtain an endomorphism @xmath96 on the quadratic twist @xmath95 of @xmath6 . indeed , if @xmath154 , then @xmath155 , and @xmath96 is defined over @xmath0 .
[ prop : eigenvalue ] with the notation above : @xmath156\pi_{{\mathcal{e } } } \qquad & \text { and } \qquad ( \psi')^2 = [ -\epsilon_p d]\pi_{{\mathcal{e } } ' } .
\intertext { there exists an integer \(r\ ) satisfying \
( dr^2 = 2p + \epsilon_p { { \mathrm{tr}({{\mathcal{e } } } ) } } \ ) such that } \psi = \tfrac{1}{r}\left(\pi_{{\mathcal{e } } } + \epsilon_p p\right ) \qquad & \text { and } \qquad \psi ' = \tfrac{-1}{r}\left(\pi_{{\mathcal{e } } ' } - \epsilon_p p\right ) .
\end{aligned}\ ] ] the characteristic polynomial of both @xmath16 and @xmath96 is @xmath157 clearly , @xmath158 .
hence @xmath159{{{}^{(p)}\!\!{\pi_0 } } } = [ \epsilon_p d]\pi_0{{{}^{(p)}\!\!{\pi_0 } } } = [ \epsilon_p d]\pi_{{\mathcal{e } } } .\ ] ] choosing a nonsquare @xmath86 in @xmath0 , so @xmath160 and @xmath155 , we find @xmath161\pi_{{\mathcal{e}}}\circ{\delta({\mu^{-1/2 } } ) } \\ & = { \delta({\mu^{(1-p^2)/2}})}[\epsilon_p d]\pi_{{\mathcal{e } } ' } = { \delta({-1})}[\epsilon_p d]\pi_{{\mathcal{e } } ' } = [ -\epsilon_p d]\pi_{{\mathcal{e } } ' } \end{aligned}\ ] ] using the relations @xmath162 and @xmath163 , we verify that the expressions for @xmath16 and @xmath96 give the two square roots of @xmath164 in @xmath165 , and @xmath166 in @xmath167 , and that the claimed characteristic polynomial is satisfied .
now we just need a source of quadratic @xmath1-curves of small degree .
elkies @xcite shows that all @xmath1-curves correspond to rational points on certain modular curves : let @xmath168 be the quotient of the modular curve @xmath169 by all of its atkin lehner involutions , let @xmath40 be a quadratic field , and let @xmath43 be the involution of @xmath40 over @xmath1 .
if @xmath170 is a point in @xmath171 and @xmath172 is a preimage of @xmath170 in @xmath173 , then @xmath172 parametrizes ( up to @xmath174-isomorphism ) a @xmath39-isogeny @xmath175 over @xmath40 .
luckily enough , for very small @xmath39 , the curves @xmath169 and @xmath168 have genus zero so not only do we get plenty of rational points on @xmath168 , we get a whole one - parameter family of @xmath1-curves of degree @xmath39 .
hasegawa gives explicit universal curves for @xmath47 , @xmath48 , and @xmath176 in ( * ? ? ?
* theorem 2.2 ) : for each squarefree integer @xmath177 , every @xmath1-curve of degree @xmath178 over @xmath111 is @xmath174-isomorphic to a rational specialization of one of these families .
hasegawa s curves for @xmath47 and @xmath48 ( @xmath179 in [ sec : degree-2 ] and @xmath180 in [ sec : degree-3 ] ) suffice not only to illustrate our ideas , but also to give useful practical examples .
before moving on to concrete constructions , we will show that the endomorphisms developed in [ sec : construction ] yield short scalar decompositions .
proposition [ prop : bitlength ] below gives explicit formul for producing decompositions of at most @xmath181 bits .
suppose @xmath10 is a cyclic subgroup of @xmath182 such that @xmath183 , and let @xmath184 .
proposition [ prop : eigenvalue ] shows that @xmath16 acts as a square root of @xmath185 on @xmath10 : its eigenvalue is @xmath186 we want to compute a decomposition @xmath27 so as to efficiently compute @xmath30p = [ a]p + [ b\lambda_\psi]p = [ a]p + [ b]\psi(p ) .\ ] ] the decomposition of @xmath12 is not unique : far from it .
the set of all decompositions @xmath33 of @xmath12 is the coset @xmath187 , where @xmath188 is the lattice of decompositions of @xmath128 ( that is , of @xmath33 such that @xmath189 ) .
we want to find a decomposition where @xmath28 and @xmath29 have minimal bitlength : that is , where @xmath190 is as small as possible .
the standard technique is to ( pre)-compute a short basis of @xmath191 , then use babai rounding @xcite to transform each scalar @xmath12 into a short decomposition @xmath33 .
the following lemma outlines this process ; for further detail and analysis , see @xcite and @xcite .
[ lemma : decomp ] let @xmath192 be linearly independent vectors in @xmath191 .
let @xmath12 be an integer , and set @xmath193 where @xmath194 is the ( unique ) solution in @xmath195 to the linear system @xmath196 .
then @xmath197 this is just ( * ? ? ?
* lemma 2 ) ( under the infinity norm ) .
we see that better decompositions of @xmath12 correspond to shorter bases for @xmath191 .
if @xmath198 is not unusually small , then we can compute a basis for @xmath191 of size @xmath199 using the gauss reduction or euclidean algorithms ( cf .
@xcite and @xcite ) .
are derived in @xcite , but they are suboptimal for our endomorphisms . in cryptographic contexts ,
proposition [ prop : bitlength ] gives better results .
] the basis depends only on @xmath9 and @xmath25 , so it can be precomputed . in our case ,
lattice reduction is unnecessary : we can immediately write down two linearly independent vectors in @xmath191 that are `` short enough '' , and thus give explicit formulae for @xmath33 in terms of @xmath12 .
these decompositions have length @xmath200 , which is near - optimal in cryptographic contexts : if @xmath201 , then @xmath202 .
[ prop : bitlength ] with the notation above : given an integer @xmath12 , let @xmath203 then , assuming @xmath204 and @xmath205 , we have @xmath206 eq . yields @xmath207 and @xmath208 , so the vectors @xmath209 and @xmath210 are in @xmath191 ( they generate a sublattice of determinant @xmath211 ) . applying lemma [ lemma : decomp ] with @xmath212 and @xmath213
, we see that @xmath214 and @xmath215 . but
@xmath216 ( since @xmath217 ) and @xmath204 , so @xmath218 .
the result follows on taking logs , and noting that @xmath219 ( since @xmath2 ) .
let @xmath108 be a squarefree integer .
hasegawa defines a one - parameter family of elliptic curves over @xmath111 by @xmath220 where @xmath221 is a free parameter taking values in @xmath1 ( * ? ? ?
* theorem 2.2 ) .
the discriminant of @xmath179 is @xmath222 , so the curve @xmath179 has good reduction at every @xmath2 with @xmath223 , for every @xmath221 in @xmath1 .
the curve @xmath179 has a rational @xmath224-torsion point @xmath225 , which generates the kernel of a @xmath224-isogeny @xmath226 defined over @xmath227 .
we construct @xmath228 explicitly : vlu s formulae @xcite define the ( normalized ) quotient @xmath229 , and then the isomorphism @xmath230 is the quadratic twist @xmath231 .
composing , we obtain an expression for the isogeny as a rational map : @xmath232 conjugating and composing , we recognise that @xmath233 is @xmath234 $ ] if @xmath235 , and @xmath236 $ ] if @xmath237 : that is , the sign function for @xmath238 is @xmath239 [ th : d2 ] let @xmath2 be a prime , and define @xmath143 as in eq . .
let @xmath108 be a nonsquare is ( theoretically ) irrelevant , since all quadratic extensions of @xmath58 are isomorphic .
if @xmath108 and @xmath240 are two nonsquares in @xmath58 , then @xmath241 for some @xmath28 in @xmath58 , so @xmath242 and @xmath243 are identical .
we are therefore free to choose any practically convenient value for @xmath108 , such as one permitting faster arithmetic in @xmath244 . ] in @xmath38 , so @xmath245 .
let @xmath246 be the mapping defined by @xmath247 for each @xmath221 in @xmath38 , let @xmath248 be the elliptic curve over @xmath0 defined by @xmath249 then @xmath248 has an efficient @xmath0-endomorphism @xmath250 of degree @xmath251 , such that @xmath252\pi_{{\mathcal{e}}_{2,\delta , s}}\ ] ] for some integer @xmath253 satisfying @xmath254 . the characteristic polynomial of @xmath255 is @xmath256 .
the twisted endomorphism @xmath257 on @xmath258 satisfies @xmath259 , and @xmath260\pi_{{\mathcal{e}}_{2,\delta , s}'}$ ] , and @xmath261 .
reduce @xmath179 and @xmath228 mod @xmath3 and compose with @xmath150 as in [ sec : construction ] , then apply proposition [ prop : eigenvalue ] using eq . .
if @xmath262 is a cyclic subgroup of order @xmath9 such that @xmath263 , then the eigenvalue of @xmath255 on @xmath10 is @xmath264 applying proposition [ prop : bitlength ] , we can decompose scalar multiplications in @xmath10 as @xmath14p = [ a]p + [ b]\psi_{2,\delta , s}(p)$ ] where @xmath28 and @xmath29 have at most @xmath265 bits .
[ prop : j - d2 ] theorem [ th : d2 ] yields at least @xmath266 non - isomorphic curves ( and at least @xmath267 non-@xmath0-isomorphic curves , if we count the quadratic twists ) equipped with efficient endomorphisms .
it suffices to show that the @xmath53-invariant @xmath268 takes at least @xmath266 distinct values in @xmath0 as @xmath221 ranges over @xmath38 .
if @xmath269 with @xmath270 , then @xmath271 and @xmath272 satisfy @xmath273 , where @xmath274 and @xmath275 are polynomials over @xmath38 .
if @xmath271 and @xmath272 are in @xmath38 , then we must have @xmath276 . solving the simultaneous equations , discarding the solutions that can never be in @xmath38 , and dividing by two ( since @xmath277 and @xmath278 represent the same collision ) yields at most 3 collisions @xmath279 with @xmath280 in @xmath38 .
we observe that @xmath281 , so we do not gain any more isomorphism classes in proposition [ prop : j - d2 ] by including the codomain curves .
let @xmath108 be a squarefree discriminant ; hasegawa defines a one - parameter family of elliptic curves over @xmath111 by @xmath282 where @xmath221 is a free parameter taking values in @xmath1 . as for the curves in [ sec : degree-2 ] , the curve @xmath180 has good reduction at every @xmath2 with @xmath223 , for every @xmath221 in @xmath1 .
the curve @xmath180 has a subgroup of order @xmath48 defined by the polynomial @xmath283 , consisting of @xmath128 and @xmath284 .
exactly as in [ sec : degree-2 ] , taking the vlu quotient and twisting by @xmath285 yields an explicit 3-isogeny @xmath286 ; its sign function is @xmath287 [ th : d3 ] let @xmath2 be a prime , and define @xmath143 as in eq . .
let @xmath108 be a nonsquare , the particular value of @xmath108 is theoretically irrelevant .
] in @xmath58 , so @xmath245 .
let @xmath288 be the mapping defined by @xmath289 for each @xmath221 in @xmath38 , we let @xmath290 be the elliptic curve over @xmath0 defined by @xmath291 then @xmath290 has an efficient @xmath0-endomorphism @xmath292 of degree @xmath293 , such that @xmath294\pi_{{\mathcal{e}}_{3,\delta , s } } \qquad \text { and } \qquad \psi_{3,\delta , s } = \frac{1}{r}\left(\pi + \epsilon_p p\right)\ ] ] for some integer @xmath253 satisfying @xmath295 .
the characteristic polynomial of @xmath296 is @xmath297 .
the twisted endomorphism @xmath298 on @xmath299 satisfies @xmath300\pi_{{\mathcal{e}}_{3,\delta , s}'}$ ] , and @xmath301 , and @xmath302 .
reduce @xmath180 and @xmath303 mod @xmath3 , compose with @xmath150 as in [ sec : construction ] , and apply proposition [ prop : eigenvalue ] using eq . .
[ prop : j - d3 ] theorem [ th : d3 ] yields at least @xmath304 non - isomorphic curves ( and counting quadratic twists , at least @xmath305 non-@xmath0-isomorphic curves ) equipped with efficient endomorphisms .
the proof is exactly as for proposition [ prop : j - d2 ] .
we will now exhibit some curves with our families with cryptographic parameter sizes , and secure and twist - secure group orders .
we computed the curve orders below using magma s implementation of the schoof elkies atkin algorithm @xcite .
first , consider the degree-2 curves of [ sec : degree-2 ] . by definition , @xmath248 and its quadratic twist
@xmath258 have points of order 2 over @xmath0 : they generate the kernels of our endomorphisms . if @xmath306 , then @xmath307 implies @xmath308 , so when @xmath306 either @xmath309 or @xmath310 is divisible by 3 .
however , when @xmath311 we can hope to find curves of order twice a prime whose twist also has order twice a prime .
let @xmath312 and @xmath313 . for @xmath314 , we find a twist - secure curve : @xmath315 and @xmath316 where @xmath317 are 159-bit primes .
proposition [ prop : bitlength ] lets us replace 160-bit scalar multiplications in @xmath318 and @xmath319 with 80-bit multiexponentiations .
now , consider the degree-3 curves of [ sec : degree-3 ] .
the order of @xmath320 is always divisible by @xmath48 : the kernel of @xmath296 is generated by the rational point @xmath321 . however , on the quadratic twist , the nontrivial points in the kernel of @xmath298 are _ not _ defined over @xmath0 ( they are conjugates over @xmath0 ) , so @xmath322 can have prime order .
let @xmath4 ; then @xmath323 is a nonsquare in @xmath38 .
the parameter value @xmath324 yields @xmath325 where @xmath9 is a 253-bit prime and @xmath326 is a 254-bit prime .
using proposition [ prop : bitlength ] , any scalar multiplication in @xmath327 or @xmath328 can be computed via a 127-bit multiexponentiation .
let @xmath5 ; then @xmath329 is a nonsquare in @xmath38 .
the parameter @xmath330 + yields @xmath331 and @xmath332 , where @xmath9 and @xmath326 are 509- and 510-bit primes , respectively .
proposition [ prop : bitlength ] transforms any 510-bit scalar multiplication in @xmath333 or @xmath334 into a 255-bit multiexponentiation .
the curve @xmath248 has a montgomery model over @xmath0 if and only if @xmath335 is a square in @xmath0 ( by ( * ? ? ?
* proposition 1 ) ) : in that case , setting @xmath336 the birational mapping @xmath337 takes us from @xmath248 to the projective montgomery model @xmath338 ( if @xmath335 is not a square , then @xmath339 is @xmath0-isomorphic to the quadratic twist @xmath258 . )
these models offer a particularly efficient arithmetic , where we use only the @xmath340 and @xmath341 coordinates @xcite .
the endomorphism is defined ( on the @xmath340 and @xmath341 coordinates ) by @xmath342 every montgomery model corresponds to a twisted edwards model ( and vice versa ) @xcite . with @xmath343 and @xmath344 , the birational maps @xmath345 take us between the montgomery model of eq . and the twisted edwards model @xmath346 doubling - oriented doche icart
kohel models of elliptic curves are defined by equations of the form @xmath347 these curves have a rational @xmath224-isogeny @xmath70 , with kernel @xmath348 ; in this form , we can double more quickly by using the decomposition @xmath349 = { { { \phi}^{\dagger}}}{\phi}$ ] ( see @xcite for details ) .
our curves @xmath248 come equipped with a rational @xmath224-isogeny , so it is natural to try putting them in doche
kohel form .
the isomorphism @xmath350 takes us from @xmath248 into a doubling - oriented doche icart kohel model @xmath351 while @xmath352 is defined over @xmath0 , the isomorphism is only defined over @xmath353 ; so
if @xmath354 is not a square in @xmath0 then @xmath352 is @xmath0-isomorphic to @xmath258 .
the endomorphism @xmath355 is @xmath356-isomorphic to the doche
icart kohel isogeny , since they have the same kernel .
the eigenvalue of @xmath255 on cryptographic subgroups is @xmath357 , so computing @xmath14p$ ] as @xmath358p + [ b]\psi_{2,\delta , s}^\mathrm{dik}$ ] with doche
kohel doubling for @xmath358 $ ] and @xmath359 $ ] is like using a @xmath360-adic expansion of @xmath12 .
similarly , we can exploit the rational @xmath48-isogeny on @xmath290 for doche
kohel tripling ( see @xcite ) .
the isomorphism @xmath361 , with @xmath362 and @xmath363 , takes us from @xmath290 to the tripling - oriented doche icart - kohel model @xmath364
returning to the framework of [ sec : construction ] , suppose @xmath116 is a curve defined over @xmath1 , and base - extended to @xmath365 : then @xmath366 , and we can apply the construction of [ sec : construction ] taking @xmath109 to be the identity map . reducing modulo an inert prime @xmath3 , the endomorphism @xmath16 is nothing but @xmath150 ( which is an endomorphism , since @xmath6 is a subfield curve ) .
we have @xmath367 , so the eigenvalue of @xmath16 is @xmath368 on cryptographic subgroups of @xmath182 . clearly , this endomorphism is of no use to us for scalar decompositions .
however , looking at the quadratic twist @xmath95 , the twisted endomorphism @xmath96 satisfies @xmath369 ; the eigenvalue of @xmath96 on cryptographic subgroups is a square root of @xmath127 .
we have recovered the galbraith
scott endomorphism ( cf .
* theorem 2 ) ) . more generally , suppose @xmath109 is a @xmath174-isomorphism : that is , an isogeny of degree @xmath125 . if @xmath116 does not have cm , then @xmath370 , so @xmath371\pi_{{\mathcal{e}}}$ ] with @xmath372 .
this situation is isomorphic to gls .
in fact , @xmath373 implies @xmath374 , so @xmath375 is in @xmath1 , and @xmath116 is isomorphic to ( or a quadratic twist of ) a curve defined over @xmath1 .
we note that in the case @xmath64 , we have @xmath376 in proposition [ prop : eigenvalue ] , and the basis constructed in the proof of proposition [ prop : bitlength ] is ( up to sign ) the same as the basis of ( * ? ? ?
* lemma 3 ) .
while @xmath377 may have prime order , @xmath182 can not : the points fixed by @xmath150 form a subgroup of order @xmath378 , where @xmath379 ( the complementary subgroup , where @xmath150 has eigenvalue @xmath127 , has order @xmath380 ) .
we see that the largest prime divisor of @xmath211 can be no larger than @xmath381 .
if we are in a position to apply the fouque lercier ral valette fault attack @xcite for example , if montgomery ladders are used for scalar multiplication and multiexponentiation then we can solve dlp instances in @xmath377 in @xmath382 group operations ( in the worst case ! ) . while @xmath382 is still exponentially difficult , it falls far short of the ideal @xmath381 for general curves over @xmath0 .
gls curves should therefore be avoided where the fault attack can be put into practice .
we conclude with some brief remarks on @xmath1-curves of other small degrees .
hasegawa provides a universal curve for @xmath383 ( and any @xmath108 ) in ( * ? ? ?
* theorem 2.2 ) , and our results for @xmath47 and @xmath384 carry over to @xmath383 in an identical fashion , though the endomorphism is slightly less efficient in this case ( its defining polynomials are sextic ) . for @xmath385 , hasegawa notes that it is impossible to give a universal @xmath1-curve for every discriminant @xmath108 : there exists a quadratic @xmath1-curve of degree @xmath386 over @xmath111 if and only if @xmath387 for every prime @xmath388 dividing @xmath108 ( * ? ? ?
* proposition 2.3 ) .
but this is no problem when reducing modulo @xmath3 , if we are prepared to give up the freedom of choosing @xmath108 : we could take @xmath389 for @xmath390 and @xmath391 for @xmath392 , and then use the curves defined in ( * ? ? ?
* table 6 ) .
composite degree @xmath1-curves ( such as @xmath393 and @xmath153 ) promise more interesting results , as do exceptional cm specializations of the universal curves ; we will return to these cases in future work .
degrees greater than 10 yield less efficient endomorphisms , and so are less interesting from a practical point of view .
x l. babai , _ on lovasz lattice reduction and the nearest lattice point problem_. combinatorica * 6 * ( 1986 ) 113 d. j. bernstein , p. birkner , m. joye , t. lange , and c. peters , _ twisted edwards curves_. in s. vaudenay ( ed . ) , _ africacrypt 2008_. lecture notes in comput .
* 5023 * ( 2008 ) 389405 j. w. bos , c. costello , h. hisil , and k. lauter , _ fast cryptography in genus 2_. in t. johansson and p. q. nguyen ( eds ) , _ advances in cryptology
eurocrypt 2013 _ , lecture notes in comput .
* 7881 * ( 2013 ) 194210 w. bosma , j. j. cannon , c. fieker , and a. steel ( eds ) , _ handbook of magma functions_. edition 2.19 ( 2013 ) h. cohen and g. frey ( eds ) , _ handbook of elliptic and hyperelliptic curve cryptography_. chapman & hall / crc ( 2006 ) j. cremona , j. c. lario , j. quer , and k. ribet ( eds ) , _ modular curves and abelian varieties_. birkhuser , basel ( 2004 ) c. doche , t. icart , and d. r. kohel , _ efficient scalar multiplication by isogeny decompositions_. in m. yung , y. dodis , a. kiayias , and t. malkin ( eds ) , _ public key cryptography 2006 _ , lecture notes in comput .
sci * 3958 * ( 2006 ) 191206 n. d. elkies , _ on elliptic @xmath394-curves_. in @xcite , pp .
8192 j. s. ellenberg , _
@xmath1-curves and galois representations_. in @xcite , pp .
93103 p .- a .
fouque , r. lercier , d. ral , and f. valette , _ fault attack on elliptic curve with montgomery ladder_. fdtc 08 , ieee - cs ( 2008 ) 9298 s. d. galbraith , _ mathematics of public key cryptography_. cambridge university press ( 2012 ) s. d. galbraith , x. lin , and m. scott ,
_ endomorphisms for faster elliptic curve cryptography on a large class of curves_. j. crypt . *
24 * # 3 ( 2011 ) 446469 r. p. gallant
, r. j. lambert , and s. a. vanstone , _ faster point multiplication on elliptic curves with efficient endomorphisms_. in j. kilian ( ed . ) , advances in cryptology : crypto 2001 , lecture notes in comput .
* 2139 * ( 2001 ) 190200 j. gonzalez , _ isogenies of polyquadratic @xmath1-curves to their galois conjugates_. arch .
* 77 * ( 2001 ) 38390 y. hasegawa , _
@xmath1-curves over quadratic fields_. manuscripta math . *
94 * # 1 ( 1997 ) 34764 h. hisil , k. wong , g. carter , and e. dawson , _ twisted edwards curves revisited_. in j. pieprzyk ( ed . ) , _ advances in cryptology asiacrypt 2008 _ , lecture notes in comput .
sci * 5350 * ( 2008 ) 326343 d. r. kohel and b. smith , _ efficiently computable endomorphisms for hyperelliptic curves_. in f. hess , s. pauli , and m. pohst ( eds ) , _ algorithmic number theory : ants - vii _ , lecture notes in comput .
* 4076 * ( 2006 ) 495509 p. longa and f. sica , _ four - dimensional gallant lambert vanstone scalar multiplication_. in x. wang and k. sako ( eds ) , _ advances in cryptology asiacrypt
2012_. lecture notes in comput .
* 7658 * ( 2012 ) 718739 .
full version : http://eprint.iacr.org/2011/608 the magma computational algebra system . http://magma.maths.usyd.edu.au/ p. l. montgomery , _ speeding the pollard and elliptic curve methods of factorization_. math . comp . * 48 * # 177 ( 1987 ) 243264 a. menezes , t. okamoto , and s. a. vanstone , _ reducing elliptic curve logarithms to logarithms in a finite field_. ieee trans . inform . theory * 39 * # 5 ( 1993 ) 163946 k. okeya , h. kurumatani , and k. sakurai , _ elliptic curves with the montgomery - form and their cryptographic applications_. in h. imai and y. zheng ( eds ) , _ public key cryptography : pkc 2000 _ , lecture notes in comput .
* 1751 * ( 2000 ) 23857 r. schoof , _ elliptic curves over finite fields and the computation of square roots mod @xmath3_. math .
* 44 * ( 1985 ) 74563 f. sica , m. ciet , and j .- j .
quisquater , _ analysis of the gallant - lambert - vanstone method based on efficient endomorphisms : elliptic and hyperelliptic curves_. in k. nyberg and h. m. heys ( eds ) , _ selected areas in cryptography : sac 2002 _ , lecture notes in comput .
* 2595 * ( 2003 ) 2136 j. h. silverman , _ the arithmetic of elliptic curves _
texts in math .
* 106 * ( 2e ) , springer ( 2009 ) e. g. straus , _ addition chains of vectors_. amer .
monthly * 71 * # 7 ( 1964 ) 8068 k. takashima , _ a new type of fast endomorphisms on jacobians of hyperelliptic curves and their cryptographic application_. ieice trans .
fundamentals * e89-a * # 1 ( 2006 ) 124133 j. vlu , _ isognies entre courbes elliptiques_. c. r. math . acad .
paris * 273 * ( 1971 ) 23841 e. verheul , _ evidence that xtr is more secure than supersingular elliptic curve cryptosystems_. j. crypt .
* 17 * ( 2004 ) 27796 z. zhou , z. hu , m. xu , and w. song , _
efficient 3-dimensional glv method for faster point multiplication on some gls elliptic curves_. inf .
* 110 * # 22 ( 2010 ) 10031006 | we construct new families of elliptic curves over @xmath0 with efficiently computable endomorphisms , which can be used to accelerate elliptic curve - based cryptosystems in the same way as gallant lambert vanstone ( glv ) and galbraith lin scott ( gls ) endomorphisms .
our construction is based on reducing @xmath1-curves curves over quadratic number fields without complex multiplication , but with isogenies to their galois conjugates modulo inert primes . as a first application of the general theory
we construct , for every @xmath2 , two one - parameter families of elliptic curves over @xmath0 equipped with endomorphisms that are faster than doubling . like gls ( which appears as a degenerate case of our construction )
, we offer the advantage over glv of selecting from a much wider range of curves , and thus finding secure group orders when @xmath3 is fixed .
unlike gls , we also offer the possibility of constructing twist - secure curves . among our examples are prime - order curves equipped with fast endomorphisms , with almost - prime - order twists , over @xmath0 for @xmath4 and @xmath5 .
elliptic curve cryptography , endomorphisms , glv , gls , exponentiation , scalar multiplication , @xmath1-curves . |
the two - gluon exchange model @xcite has for a long time offered a semi - quantitative understanding of the pomeron in qcd .
the model originated as a view of elastic scattering .
it has been applied extensively to vector meson photoproduction , and has recently been used to predict jet production in double pomeron interactions @xcite . in this paper , we use the model to calculate exclusive @xmath1 photoproduction at low @xmath2 .
the diagrams for @xmath0 are shown in figure 1 .
the upper half of each diagram can be calculated in a straight - forward manner , since only known electroweak and gluonic couplings appear .
the lower half of each diagram involves the non - perturbative color structure of the proton , but the required discontinuity can be obtained by parametrizing it and constraining the parameters to fit @xmath3 elastic scattering .
the calculation of @xmath1 photoproduction thus introduces no new parameters .
a similar calculation can be done for @xmath4 , where @xmath5 provides a large momentum scale ; but that calculation is not as clean because it brings in the @xmath6 hadronic wave function .
reggeization of the gluons and interactions between them ( e.g. , as given by ladder diagrams ) must be responsible for the gradual energy dependence of pomeron exchange @xcite .
we will ignore these effects here . from a practical standpoint , the energy dependence of @xmath0
will be dominated by the minimum longitudinal momentum transfer that is kinematically required when the energy is not far above threshold .
the predicted cross section is necessarily small , because the large mass @xmath7 sets a short distance scale @xmath8 for the typical transverse separation between the @xmath9 pair that form the @xmath1 .
this leads to a factor @xmath10 in the final amplitude , which is proportional to the color dipole moment of the pair .
a couple of enhancement factors @xmath11 arise in the calculation , as explained below ; but the final predicted cross section turns out to be too small to observe at the hera @xmath12 collider in its present form .
the imaginary part of the amplitude is equal to @xmath13 times the discontinuity given by the four diagrams of figure 1 .
the upper half of diagram ( a ) , for example , corresponds to a factor @xmath14 \ , [ 2 \pi \delta(k_2 ^ 2 - m^2 ) ] } { [ ( p_1 - k_1)^2 - m^2 ] \ , [ ( p_3 - k_2)^2 - m^2 ] } \ ; { \rm tr}_a \label{eq : eq1}\end{aligned}\ ] ] in the discontinuity , where @xmath15 \ , \gamma_\mu \ , ( m + i \gamma \ ! \cdot \ !
k_2 ) \ , ( a + b \gamma_5 ) \ , \nonumber \\ & & \ ; \ ; \quad \gamma \ ! \cdot \ !
\epsilon^*(p_3 ) \ , [ m - i \gamma \ ! \cdot \!(k_1 - q_2 ) ] \ , \gamma_\nu \ , ( m - i \gamma \ ! \cdot \!k_1 ) \ } \ , \ ; . \label{eq : eq2}\end{aligned}\ ] ] to obtain the high energy limit of the amplitude , we use light - cone coordinates @xmath16 where @xmath17 and @xmath18 are large and @xmath19 .
one of the two delta functions in eq .
( [ eq : eq1 ] ) is saved for the eventual integration over gluon momentum @xmath20 , which carries no large @xmath21 or @xmath22 component as a result of this and a similar delta function from the lower half of the diagram .
the other delta function reduces eq .
( [ eq : eq1 ] ) to a three - dimensional integral over transverse momentum and light - cone momentum fraction @xmath23 : @xmath24 where @xmath25 / x \ , \nonumber \\ d_2 & = & ( p_3 - k_2)^2 - m^2 \cong [ ( k_1 - q_2 - x p_3)_\perp^2 + m^2 - x(1-x)p_3 ^ 2 ] / ( 1 - x ) \ ; . \label{eq : eq4}\end{aligned}\ ] ] for @xmath26 we can choose a gauge in which @xmath27 has only transverse components
. then @xmath28 can be assumed to have only transverse components as well , because the production of longitudinal @xmath1 can be shown to be suppressed by a factor @xmath29 in the amplitude . upon summing over all four diagrams ,
the contribution from @xmath30 terms is found to vanish in the large @xmath31 limit . combining the denominators in eq .
( [ eq : eq3 ] ) using the feynman parameter trick @xmath32^{-2}$ ] allows the @xmath33 integral to be performed , and leads to @xmath34 where @xmath35 / c \nonumber \\ & - & [ 1 - 2x(1-x ) ] \ln c ) + 4 \ , v \ !
\epsilon_\perp(p_1 ) \ , v \ !
\epsilon_\perp^*(p_3 ) \ , x \ , ( 1-x)\ , y\ , ( 1-y)/ c \ } \nonumber \\ \
, \null \nonumber \\ % blank line in equation c & = & m^2 + y \ , ( 1-y ) \ , v^2 - x \ , ( 1-x ) \ , [ y \ , p_3 ^ 2 + ( 1-y ) \ , p_1 ^ 2 ] \nonumber \\ \ , \null \nonumber \\ % blank line in equation v & = & q_\perp + x \delta_\perp \ ; .
\label{eq : eq7}\end{aligned}\ ] ] here @xmath36 is approximately zero , @xmath37 , and @xmath38 is the transverse momentum transfer . the @xmath39 term in eq .
( [ eq : eq6 ] ) comes from the two `` off - diagonal '' diagrams in which one gluon is attached to the quark line and the other to the anti - quark .
the @xmath40 term comes from the diagonal diagrams in which both gluons are attached to the same line .
there is a strong cancellation between these two contributions when @xmath41 or @xmath42 is small , because the two - gluon system is color neutral and the @xmath9 system is spatially compact due to the large @xmath1 mass .
however , the diagonal term will dominate in the full amplitude because the range of eventual integration over @xmath43 is controlled by @xmath7 .
we can estimate the integral in eq .
( [ eq : eq7 ] ) using the approximation @xmath44 , @xmath45 , @xmath46 .
this limit is obtained by splitting the integration into six separate regions according to @xmath47 or @xmath48 and @xmath49 , @xmath50 , or @xmath51 .
the result is independent of @xmath52 and @xmath53 in the limit @xmath54 , @xmath55 and is given by @xmath56 \nonumber \\ & & \ , + \ , [ q_{1\perp } \ !
\epsilon_\perp(p_1 ) \ , q_{2\perp } \ !
\epsilon^*_\perp(p_3 ) \ , + \ , q_{2\perp } \ !
\epsilon_\perp(p_1 ) \ , q_{1\perp } \ !
\epsilon^*_\perp(p_3 ) ] / m^2 \ } \nonumber \\ \ , \null \nonumber \\ % blank line in equation f(a ) & = & a \ , [ \ln ( m^2/m_z^2 ) - i\pi - \frac{1}{2 } + b \ln \frac{b+1}{b-1 } \ , ] \nonumber \\ \ , \null \nonumber \\ % blank line in equation b & = & \sqrt{1 + 4/a } \label{eq : eq8}\end{aligned}\ ] ] we want the limit of high energy at small momentum transfer , where the helicity non - flip amplitude dominates .
we can therefore take @xmath57 to obtain @xmath58 it is a good approximation to neglect the quark mass @xmath59 , since eq .
( [ eq : eq8 ] ) is not singular in the limit @xmath60 .
this approximation is reasonably good numerically even for quark masses up to the charm quark mass , in the important regions of @xmath61 , @xmath62 and @xmath63 , because the corrections to it do not contain the large factor @xmath64 . in this way
, we find @xmath65 \nonumber \\
g(q^2 ) & = & q^2 \ , [ \ln q^2 / m_z^2 - i\pi ] \ ; .
\label{eq : eq10}\end{aligned}\ ] ] eq .
( [ eq : eq10 ] ) was derived for the region where @xmath66 , @xmath45 , and @xmath67 are small compared to @xmath68 .
we can provide it with approximately correct behavior when these quantities become comparable to @xmath68 by inserting an additional factor @xmath69^{-1}$ ] into eq .
( [ eq : eq10 ] ) .
the adequacy of the resulting approximation to @xmath70 , for the purpose of calculating the full amplitude , has been checked by numerical integration .
the coefficient @xmath71 in eq .
( [ eq : eq10 ] ) is given by @xmath72 this includes a factor @xmath73 from the sum over quark flavors @xmath74 and @xmath75 , and a factor @xmath76 from the sum over quark colors .
i take @xmath77 for the strong coupling at the low average momentum transfer scale that occurs here .
this is slightly optimistic , since the amplitude receives contributions from a spectrum of gluon transverse momenta extending all the way up to @xmath78 .
it leads to @xmath79 .
a reasonable parametrization for the discontinuity of the gluon - proton amplitude represented by the bottom half of figure 1 has been given in ref .
@xcite : @xmath80 where @xmath81 } \ ; , \label{eq : eq13}\end{aligned}\ ] ] with @xmath82 and @xmath83 . a selection of reasonable parameter choices that have been tuned to fit @xmath3 elastic scattering
is given in table i. dependence of the results on the choice of parameter set provides an estimate of systematic errors .
table i + parameters and predicted cross section @xmath84 and slope @xmath85 for @xmath0 . [ cols="^,^,^,^,^,^ " , ] [ table1 ] the full amplitude for @xmath0 in the two gluon exchange picture of figure 1 ( including a factor @xmath86 from the sum over gluon colors ) is given by @xmath87 } \label{eq : eq14}\end{aligned}\ ] ] with @xmath88 .
a finite gluon mass @xmath89 is included in the propagators to suppress contributions from long distance , as an approximation to color confinement @xcite . in elastic scattering , this is necessary to avoid unphysical behavior of the elastic slope in the limit @xmath90 .
the @xmath91 integral can be done numerically at each momentum transfer @xmath92 .
the result is enhanced by two powers of @xmath64 : one coming directly from @xmath70 ( see eq .
( [ eq : eq10 ] ) ) and the other coming from the integration over @xmath91 , since the integrand behaves as @xmath93 for large @xmath94 up to @xmath95 .
the differential cross section is given by @xmath96 in the high energy limit .
it is independent of @xmath31 in that limit .
the predicted integrated cross section @xmath97 is listed in table i for each choice of parameters .
the result is approximately @xmath98 , with an uncertainty of about a factor of 2 coming from the range of parameters that provide a plausible description of elastic scattering .
( it is interesting to note that this calculated result is not far from the crude estimate of @xmath99 that can be made by scaling the observed cross section for @xmath100 at hera @xcite by the dipole size factor @xmath101 . )
the dependence of @xmath102 on momentum transfer @xmath103 is roughly exponential , so we can characterize it by a slope parameter @xmath85 such that @xmath104 .
the value of @xmath85 , defined by reproducing @xmath84 and @xmath102 at @xmath105 , is also listed in table i. it is approximately equal to the one half the slope of elastic scattering , since it arises essentially from the wave function effect associated with the proton remaining intact , while in @xmath3 elastic scattering , _ both _ protons must remain intact .
equivalently , one may say that the extent of the interaction in impact parameter space is set by that of the proton .
one can hope to look for @xmath1 photoproduction in electroproduction @xmath106 , which will be dominated by low @xmath2 transverse photons producing @xmath1 s with the same helicity as the photon .
the cross section is given by @xmath107 where @xmath108 \label{eq : eq17}\end{aligned}\ ] ] is the flux of transverse photons @xcite .
defining the four - momenta in @xmath109 as @xmath110 , @xmath111 , @xmath112 , @xmath113 , @xmath114 respectively , the usual kinematic variables are @xmath115 , @xmath116 , and @xmath117 .
the role of @xmath31 in sect .
[ sec : photoproduction ] is now played by @xmath118 .
the minimum value of @xmath2 is given by @xmath119 .
the maximum @xmath2 occurs when the electron is scattered through @xmath120 , but most of the cross section comes from the region of very small @xmath2 .
( we have assumed the limit @xmath121 in calculating the cross section for @xmath0 in sect .
[ sec : photoproduction ] , but if that condition is relaxed , the model shows no dramatic dependence on @xmath2 . )
we assume that the dependence on @xmath122 is dominated by the effect of the minimum momentum transfer @xmath123 . specifically , we assume the approximate dependence @xmath124 with @xmath125 , which was found in the large @xmath126 limit , to be approximately valid at finite @xmath126 where the minimum value of @xmath127 is different from @xmath128 because of the longitudinal momentum transfer necessary to produce the high - mass state
. this assumption might appear optimistic because producing the on - shell intermediate states in figure 1 involves an increase in mass of the states at both ends of each gluon line , which requires a noticeably larger minimum longitudinal momentum transfer .
however , the scale of transverse momentum transfer for each gluon exchange extends all the way to @xmath129 , so the longitudinal momentum transfer associated with the individual gluon exchanges need not be dominant . integrating eq .
( [ eq : eq16 ] ) at the hera energy @xmath130 leads to @xmath131 .
combining this with the results listed in table i yields a predicted cross section for diffractive @xmath1 production at hera of approximately 1 femtobarn .
a realistic experimental search would have to rely on one of the clean decay modes @xmath132 or @xmath133 , which would suppress the rate by a branching fraction @xmath134 .
the resulting predicted rate is too small by 2 3 orders of magnitude to be observed at the hera @xmath12 collider in its present form .
the quasi - elastic photoproduction process @xmath0 has been calculated using the two - gluon exchange model of the pomeron .
the calculation is rather clean because the electroweak part of it is completely well defined and the soft proton part can be normalized to elastic scattering .
the process would lead to @xmath1 production in @xmath12 collisions , which would have a clean and dramatic signature : one @xmath135 or @xmath136 pair with invariant mass equal to the @xmath1 mass , and hence with very large individual transverse momenta ; and nothing else in the detector , since the scattered electron and proton would generally disappear down the beam pipes .
unfortunately , the predicted rate is too small to be observed with the luminosity currently available at hera .
possible upgrades to hera that would increase its luminosity by 2 3 orders of magnitude could make the process observable .
because of the higher available energy , the related diffractive process @xmath137 should be observable with a @xmath138 option at a @xmath139 next linear collider @xcite .
the cross section can be estimated from sect .
[ sec : photoproduction ] and regge factorization as @xmath140 for @xmath138 energies far above threshold .
this work was supported in part by u.s . national science foundation grant number phy-9507683 .
cudell and b.u .
nguyen , nucl .
* b420 * , 669 ( 1994 ) ; j.r .
cudell and b. margolis , nucl .
* b336 * , 1 ( 1990 ) ; j.r .
cudell and i. royen , `` elastic vector meson production at hera '' , e - print archive : hep - ph/9609490 . | the two - gluon exchange model of the pomeron is used to compute the photoproduction reaction @xmath0 .
the predicted cross section is too small to be observed at hera , but may be detectable at an eventual next linear collider .
hep - ph/9612356 + msuhep61222 + * diffractive photoproduction of @xmath1 * 0.30 in * jon pumplin * 0.1 in physics and astronomy department + michigan state university + east lansing mi 48824 , u.s.a . 0.5 in |
the realization of quasi one - dimensional ultracold boson gases with tunable interaction parameters @xcite and the succeeding advances in atom chip trap technology @xcite have renewed interest in theoretical models of one - dimensional bosons with short - ranged interactions .
of particular relevance to workers in this field is the lieb - liniger model @xcite in which contact interactions are described by dirac delta functions .
the suitability of this model in describing the low temperature properties of these quasi one - dimensional bosonic systems has been further strengthened by olshanii s analysis of the low energy scattering of atoms under tight transverse harmonic confinement : the longitudinal @xmath5-wave scattering amplitudes are indeed reproduced by a one - dimensional pseudopotential proportional to a dirac delta function @xcite .
the resulting demonstration that the magnitude of this effective delta interaction can be explicitly calculated from the three - dimensional atomic scattering length and the dimensions of the confining external trap @xcite further strengthens the link between theoretical one - dimensional models and quasi one - dimensional experiments .
still , the free lieb - liniger model is quite an idealization for actual experiments because the atoms are generally longitudinally confined by an external potential and thus much effort has been devoted to studying the effects of confinement of interacting bosons @xcite . introducing an external harmonic potential to the free lieb - linger model of spinless bosons
leads to the many - particle schrdinger eigenvalue equation @xmath6\psi^\text{b } \\-
\frac{\hbar^2}{ma}\sum_{i < j } \delta(x_j - x_i)\psi^\text{b}\negmedspace , \label{ham0}\end{gathered}\ ] ] where @xmath7 is the mass of each of the @xmath1 atoms , @xmath8 is the angular frequency of the trap , and @xmath9 is the one - dimensional scattering length .
the superscript ` b ' refers to the bosonic nature of the labeled quantities .
measuring energy in units of @xmath10 and length in oscillator units @xmath11 gives the dimensionless eigenvalue equation @xmath12\psi^\text{b } , \label{ham}\ ] ] where we have introduced the dimensionless interaction strength @xmath13 .
we consider here the repulsive case @xmath14 ( negative scattering length ) to be specific . in the absence of an harmonic potential
the corresponding eigenvalue equation is solvable by the bethe ansatz and consequently much is known about the ground state and elementary excitations of this system @xcite , as well as the properties of the various correlation functions at zero and finite temperatures @xcite .
however , for the important case of harmonic confinement an exact solution to this problem for general values of the interaction strength @xmath0 is lacking .
the sole exceptions are the two - particle case that is separable in relative and central coordinates @xcite , and the tonks - girardeau ( tg ) limit of infinite repulsion @xmath15 in which the exact @xmath1-particle wavefunctions are absolute values of slater determinants @xcite .
these results , especially the latter , are prototypical examples of the fermion - boson duality derived by girardeau @xcite and later generalized by cheon and shigehara @xcite for one - dimensional systems of particles having contact interactions . for finite values of the interaction parameter @xmath0 , expressions for the atomic density and collective oscillation frequencies
have been calculated using local density approximations @xcite and time - dependent density functional theory @xcite , while formal expressions for the self - consistent hartree - fock equations for the single - particle density matrix have been obtained for general trapping potentials @xcite .
the analogous problem of a system of confined interacting fermions has also been treated by similar methods @xcite . in this work
, we will use the mentioned fermion - boson relation to develop perturbative @xmath2 corrections to the ground state energy and chemical potential of an harmonically confined interacting boson gas about the tonks - girardeau solution .
the details of the specific fermion - boson mapping we employ here that utilizes a non - local pseudopotential @xcite are given in section [ fbmapping ] . in section [ perturb ]
we present our perturbation calculations for the general case of an @xmath1-particle system obtaining a closed form analytical result that is calculable for any @xmath1 .
we analyze few - body cases and discuss the thermodynamic limit @xmath16 of our solution in section [ cases ] .
we summarize our results and give concluding remarks in section [ conclusion ] .
in one dimension it has been demonstrated that a bosonic model with pairwise contact interactions of strength @xmath0 can be mapped into a fermionic model with pairwise interactions of strength @xmath2 @xcite .
specifically , given a fermionic wavefunction @xmath17 that satisfies the eigenvalue equation @xmath18 \phi^\text{f } = e^\text{f}\phi^\text{f } , \label{hamf}\ ] ] an appropriate choice of a pseudopotential operator @xmath19 allows us to make the following correspondence between the bosonic wavefunction @xmath20 and its fermionic counterpart @xmath17 @xmath21 where @xmath22 is a function that is completely antisymmetric under any transposition @xmath23 . since the complex squares of the wavefunctions @xmath20 and @xmath17 are identical , the energy eigenvalues @xmath24 and @xmath25 are equal . for our specific problem ,
a suitable fermionic pseudopotential operator has matrix elements in the coordinate representation given by @xcite @xmath26\ , dr_{ij},\ ] ] where @xmath27 and @xmath28 are relative and central coordinates , and @xmath29 and @xmath30 are the coordinate space wavefunctions corresponding to the fermionic state kets @xmath31 and @xmath32 . in the infinite repulsion limit
, the bosonic eigenvalue equation is solved by the absolute value of the ground state slater determinant @xcite @xmath33 where @xmath34 is the fermionic ground state wavefunction , @xmath35 are the coordinates of the atoms and @xmath36 are the @xmath1 lowest energy single - particle harmonic oscillator eigenfunctions @xmath37 .
the @xmath38 appearing here are the usual hermite polynomials .
the corresponding energy of this tg ground state ( in units of @xmath10 ) is @xmath39 for a finite and large repulsion strength we may therefore use the quantity @xmath40 as a perturbation parameter for the fermionic problem so that ordinary first - order perturbation theory gives the desired correction @xmath41 in the infinite repulsion limit is given by the horizontal dotted line . ]
in this section we work mainly in the fermionic sector and omit the superscripts ` f ' for brevity . here , our objective is to explicitly evaluate the leading correction @xmath42 to the ground state energy @xmath43 . since @xmath44 is a slater determinant and the fermionic interaction operator @xmath45 is a sum of two - body operators @xmath46 having matrix elements @xmath47^*}{\partial r}\\ \times \frac{\partial\bigl[\psi_m(x_1)\psi_n(x_2)\bigr]}{\partial r}\biggr\}\ , dr,\end{gathered}\ ] ] we may calculate the perturbation @xmath48 using the slater - condon rule @xmath49 @xcite . prior to calculating the derivatives appearing inside the integral , we must be careful to write the coordinates @xmath50 and @xmath51 in terms of the relative and central coordinates @xmath52 and @xmath53 .
the symmetry of the integrand allows us to write and define @xmath54 where @xmath55}{\partial r}\biggr\}^2\ , dr.\ ] ] thus , the leading correction becomes @xmath56 and is always negative as expected . after some manipulation
, we obtain a finite series expression for the energy correction that may be evaluated for any number of atoms @xmath1 : @xmath57 .
\end{aligned}\ ] ] ( a)(b )
the special case of @xmath58 particles is separable in relative and central coordinates and the resulting eigenvalue problem for the ground state energy is exactly solvable . upon imposing vanishing boundary conditions on the two - body wavefunction at infinity , we find that the ground state energy @xmath43 of the trapped two boson system satisfies the transcendental equation @xmath59 \tan\bigl[\tfrac{1}{2}(1-e_{0})\pi\bigr]}{\gamma\bigl[e_{0}/2\bigr ] } = -\frac{c}{\sqrt{2}},\ ] ] where @xmath60 $ ] .
our perturbative result gives @xmath61 and coincides with the leading correction term in the @xmath2 series expansion of this exact solution about the tg ground state energy @xmath62 .
also , an essentially identical result was obtained by sen @xcite for the reduced single - particle problem using a different local pseudopotential involving the second derivative of the delta function @xmath63 .
a comparison of our first - order perturbation result and the exact two - particle ground state energy is shown in figure [ 2body ] and we observe good agreement between the two results in the strongly interacting regime @xmath64 . before discussing the situation for more than two particles , let us restore units and rewrite the perturbed ground state energy as @xmath65 + \mathcal{o}(1/c^2),\ ] ] where @xmath66 is a dimensionless function of @xmath1 . for values of @xmath1 up to @xmath67 the magnitude of the scaled first - order correction @xmath68 is plotted in figure [ energyn]a as a function of @xmath1 on a double logarithmic plot .
inspection of this graph suggests a simple power law scaling for the first - order correction with large @xmath1 .
leading order asymptotic analysis reveals that the partial sums in eq .
scale as @xmath69 for large @xmath1 so that @xmath70 : @xmath71,\quad n \to\infty,\ ] ] where @xmath72 is a constant number .
indeed , for as few as @xmath73 particles the factor @xmath66 is quite well approximated ( within 1% ) by the function @xmath74 with @xmath75 .
in other words , the correction factor @xmath66 reaches its asymptotic scaling behavior for systems as small as @xmath76 . to obtain a thermodynamic limit with an extensive ground state energy
, we observe that in addition to sending the number of particles to infinity we must also require the trapping frequency @xmath8 to vanish as @xmath77 .
as we shall see , this condition allows us to reproduce the thomas - fermi results near the tonks - girardeau limit .
this scaling requirement implies that the quantity @xmath78 approaches a constant in the thermodynamic limit , in contrast to the thermodynamic limit in the case of a lieb - liniger gas confined in a flat - bottomed box , in which the linear dimension of the system is taken to scale proportionally with particle number @xcite . looking back at our asymptotic expression @xmath79 , we find that the quantity @xmath80 approaches a constant value as @xmath16 in our prescribed thermodynamic limit .
this is precisely the condition used by ma and yang @xcite to obtain the ground state energy of the trapped interacting boson gas in the thomas - fermi formalism , which we reproduce here in figure [ energyn]b .
we find that our first - order @xmath2 result is reliable for @xmath81 , which means that for a typical experimental setup with hundreds of atoms first - order perturbation theory and the thomas - fermi result coincide only in the extreme repulsion limit @xmath82 .
if we now define the tonks - girardeau limit chemical potential as @xmath83 and the scaled interaction parameter as @xmath84 , we obtain the zero temperature chemical potential @xmath85.\ ] ] the first term in this expression corresponds to the chemical potential of free fermions in a one - dimensional harmonic trap while the last term gives the reduction in the chemical potential due to the finite repulsion correction and serves as a measure of the departure of the system from the unitarity limit .
in this work we have calculated the first - order finite repulsion correction to the ground state energy of harmonically trapped bosons having contact interactions for any finite number of particles @xmath1 .
for @xmath76 we found that for a fixed interaction strength this correction scales as a power law @xmath3 , which allowed us to describe a thermodynamic limit that reproduces known results from thomas - fermi approaches .
this contribution clarifies the smooth transition of the ground state properties of an harmonically confined interacting boson system as the particle number goes from finite @xmath1 to infinity near the tonks - girardeau limit .
we have demonstrated that in this strongly interacting regime , to at least leading order in @xmath2 , finite number effects are negligible in current experimental situtations that have @xmath86 atoms .
a natural extension of this work would involve higher order corrections to the ground state energy and many - body wavefunction , as was done recently for a wedge - shaped trapping potential @xcite .
if we take the set of all fermionic slater determinants as an expansion basis for ordinary perturbation theory about the tg limit , we discover that the perturbing pseudopotential couples the ground state to an infinite number of excited states .
we therefore expect a complicated analytical result for the second order energy correction resulting in a numerical problem that may require a truncation of the corresponding hilbert space .
however , on the basis of the agreement between our asymptotic results and the thomas - fermi calculation ( figure [ energyn]b ) , we conjecture that the second order correction scales as @xmath87 in the thermodynamic limit .
during the course of this work many insightful discussions were shared with a. g. abanov , d. schneble , and f. franchini .
f. p. is financially supported by a teaching assistantship from the department of physics and astronomy of the state university of new york at stony brook .
v. k. acknowledges financial support by nsf grant no .
dms-0905744 .
36ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevlett.87.130402 [ * * , ( ) ] link:\doibase 10.1038/nature02530 [ * * , ( ) ] link:\doibase 10.1126/science.1100700 [ * * , ( ) ] link:\doibase 10.1007/s00340 - 005 - 2083-z [ * * , ( ) ] link:\doibase 10.1103/physreva.75.063604 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.100.090402 [ * * , ( ) ] link:\doibase 10.1088/0953 - 4075/43/15/155002 [ * * , ( ) ] link:\doibase 10.1103/physrev.130.1605 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop _ _ ( , , ) \doibase 10.1023/a:1018705520999 [ * * , ( ) ] link:\doibase 10.1103/physreva.63.033601 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) link:\doibase 10.1103/physrevlett.85.1146 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreva.67.015602 [ * * , ( ) ] link:\doibase 10.1088/0953 - 4075/37/7/073 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.92.133202 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.93.050402 [ * * , ( ) ] @noop * * , ( ) link:\doibase 10.1103/physreva.70.032508 [ * * , ( ) ] link:\doibase 10.1103/physreva.73.033609 [ * * , ( ) ] @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) @noop * * , ( ) | we calculate the first - order perturbation correction to the ground state energy and chemical potential of a harmonically trapped boson gas with contact interactions about the infinite repulsion tonks - girardeau limit . with @xmath0 denoting the interaction strength , we find that for a large number of particles @xmath1 the @xmath2 correction to the ground state energy increases as @xmath3 , in contrast to the unperturbed tonks - girardeau value that is proportional to @xmath4 .
we describe a thermodynamic scaling limit for the trapping frequency that yields an extensive ground state energy and reproduces the zero temperature thermodynamics obtained by a local density approximation .
= 10000 |
when modeling physical systems , the carrier space ( space of states or space of events ) is usually equipped with some background mathematical structure ( for instance , vector space structure and euclidean metric in elementary mechanics , hilbert space structure and commutation relations for quantum mechanics , and so on ) . with the advent of general relativity and the ambition to have a theory of the universe as a whole it is necessary to minimize the use of mathematical structures as given a priori .
actually the einstein general approach to physics has this goal : _ one of the imperfections of the original relativistic theory of gravitation was that as a field theory it was not complete : it introduced the independent postulate that the law of motion of a particle is given by an equation of geodesic .
a complete theory knows only fields and not the concepts of particle and motion . for
these must not exist independently of the field , but are to be treated as part of it _ @xcite . in a footnote of the same paper , einstein and rosen wrote , on the stress - energy tensor representing the source in the einstein equations : _ it was clear from the very beginning that this was only a provisory complexion of the theory in the sense of a phenomenological interpretation . _ in any case ,
no matter what the motivation may be , it seems desirable to revisit elementary mechanics starting with few assumptions and adding new ones when they are required to resolve ambiguities , if any . in this note
we would like to consider the minimal assumptions which may give rise to special relativity and galilean relativity and make clear which assumptions will discriminate between them and what are the compatibilities with the space structure .
it should be remembered that the problem of deriving special relativity without supposing _ a priori _ the constancy of the light speed has been discussed by many authors ( some of them are quoted in ref .
@xcite ) . to better compare different relativity theories
, we shall start with a four - dimensional smooth manifold m as a carrier space for the description of the evolution of point particles ( _ i.e. _ , we start with the space of events ) . by using a coordinate system
say @xmath0 , we consider the equations of motion in the form @xmath1 as a simplifying assumption we require our carrier space to be diffeomorphic to @xmath2 .
we are now in a position to define a _
free system _ on our space .
the motion is said to be a free motion if there is a ( global ) coordinate system @xmath3 , such that the equations of motion acquire the form @xmath4 the parameter @xmath5 is attached to the particular dynamical system we start with and has nothing to do with space - time variables .
any such coordinate system @xmath6 will define an affine structure on the carrier space inducing the one from @xmath2 .
it follows that the notion of straight line , or , more generally , of affine space on the space of events is _ frame dependent_. in addition , in each frame we must select , first of all , a coordinate , say @xmath7 , to be associated with a notion of _
evolution_. any solution @xmath8 of our free equation which represents the _ world - line _ of an existing object for the given frame must be such that @xmath9 along the _ world - line _ ( in what follows we shall use @xmath10 ) .
after the choice of a single _ world - line _ has been made , the translation group can be used to move it and thereby to construct a congruence of _ world - lines _ which can be thought of as solutions of a vector field @xmath11 on the carrier space . because the translation group is a symmetry group for ( 2 ) ,
these _ world - lines _ are all solutions of ( 2 ) .
we now introduce a closed 1-form @xmath12 , invariant under the translation group , such that @xmath13 .
this 1-form @xmath12 defines a family of 3-planes for @xmath2 ( a foliation ) , transversal to the congruence defined by @xmath11 .
as the _ world - lines _ associated with @xmath11 are solutions of ( 2 ) , se say that @xmath14 is an inertial frame . for any such frame
, we can define _ time - like _ vectors as those @xmath15 which satisfy @xmath16
. they will be future pointing if @xmath17 .
we can say that a point @xmath18 is in the past of @xmath19 if @xmath18 can be connected to @xmath20 by a curve whose tangent vectors are future oriented .
similarly we can define a point in the future of @xmath20 .
these comments are meant to stress that @xmath12 and @xmath11 allow us to define most of the standard non metric structures on space - time ( if we had given a metric structure it would have been possible to define @xmath12 from @xmath11 and the metric ) .
any _ time - like world - line _ , equipped with a map associating in a monotonic way a real number to each event on the _ world - line ( clock ) _ will be called an _ observer_. in our approach a clock can be associate to each _ time - like world - line _ by considering the pull - back of @xmath12 to it . specifically , this can be done by defining @xmath21 by setting @xmath22 . the induced map @xmath23 is clearly monotonic because of @xmath13 .
the pair @xmath14 defines a family of observers with temporal evolution along @xmath11 and _ rest frames _ associated with @xmath24 .
two inertial systems @xmath25 and @xmath26 are compatible if @xmath27 and @xmath28 .
two compatible systems will perceive the _ world - lines _ of each other as representing some physical entities ; for this reason the requirement @xmath29 will be called the _ mutual objective existence _ condition .
we shall make the choice @xmath30
. for any pair of inertial frames , say @xmath31 , @xmath32 , we can define relative clocks by setting @xmath33 given by @xmath34 here @xmath35 is determined by the intersection point of the two observers in @xmath31 and @xmath32 respectively .
if the two observers belong to the same inertial frame , @xmath35 is determined by their intersection with any common rest frame .
this view point makes clear that the notion of inertial system is related to the dynamical evolution of some chosen _ comparison system _ via the selection of a congruence of solutions .
it should be stressed that with the assumption that our carrier space is diffeomorphic with @xmath2 , all _ free systems _ are diffeomorphic to each other . by using the linear structure induced by an inertial reference frame , we find that the inhomogeneous general linear group @xmath36 acts transitively on the set of solutions of equation ( 1 ) .
now we shall look for relativity transformations , _
i.e. _ transformations connecting pairs of inertial systems , say @xmath37 and @xmath38 , and require that they form a subgroup of @xmath36 .
by selecting one fiducial inertial system @xmath39 we parametrize all others connected to it by the elements of the relativity group we are searching for .
the restrictions on the accepted relativity transformations means that a _ particle _ at rest in one inertial frame can not be perceived by another one as a particle _ existing _ all over a real line only at a given instant of time , and . from the point of view of the transformation group ( _ i.e. _ transformations connecting physically equivalent systems )
we have to exclude the possibility of exchanging _ time- _ and _ space- _ axes .
for instance , we should exclude from admissible _ relativity transformations _ those closing on a subgroup of @xmath36 which contains rotations in the _ time - space _ planes .
we shall now describe how to construct these _ relativity ( sub)-groups_. the main idea of our procedure consists of looking for the transformation groups which transform an inertial system into another one compatible with the _ mutual objective existence _ condition .
this is translated in the requirement that a _ transformed @xmath12 _ , say @xmath40 should not contain @xmath11 in the kernel , _
i.e. _ @xmath41 and @xmath42 . before carrying on this program we shall consider some preliminary aspects ( sections 2 - 4 ) .
given a second order differential equation on space - time , say @xmath43 we ask under which conditions we can find a new coordinate system in which the equation becomes @xmath4 clearly , by performing a change of coordinates @xmath44 , we find : @xmath45 _ i.e. _ @xmath46 if we compute the curvature associated with this connection , we find @xmath47 as a matter of fact this condition turns out to be also sufficient to go from ( 3 ) to ( 4 ) . as in general
we are not requiring the force @xmath48 in ( 11 ) to be quadratic in the velocities , this sufficiency condition should be stated in the framework of generalized connections associated with any second order vector field @xcite @xcite : @xmath49=0.\ ] ] in the particular case our starting system is in the form of a spray @xmath50 we have @xmath51 thus , relation ( 5 ) is necessary and sufficient condition for a second order differential equation to represent a free system on @xmath2 . in coming section
we shall describe a constructive procedure to find these special reference frames .
starting with a second order equation on @xmath52 , say @xmath53 we can define a _ natural coordinate system _ for any neighbourhood @xmath54 in the following way .
we consider a fiducial point @xmath55 and the tangent space @xmath56 .
with any vector @xmath57 we associate a point in @xmath52 by looking for the ( unique ) solution of ( 1 ) originating at @xmath55 with initial velocity @xmath58 .
the flow @xmath59 , associated with ( 1 ) , defines a map by restriction @xmath60 where @xmath61 is the point in @xmath62 reached after a _ time _ @xmath5 via the solution of ( 1 ) with initial conditions @xmath63 .
in particular the application @xmath64 may be considered as a generalization to a generic second order equation of the standard exponential map for geodesic equations . for simplicity , we assume that ( 1 ) defines a complete vector field on @xmath65 . in this hypothesis the particular map we have constructed
is defined for any @xmath57 injectively ( we recall that @xmath52 is diffeomorphic to @xmath2 by assumption and our considerations apply to a neighbourhood @xmath62 of @xmath55 ) . in particular we may define an addition rule on @xmath62 by setting @xmath66 .
the vector space structure induced on @xmath62 for each , complete , second order vector field on @xmath52 , depends on @xmath55 . when these vector space structures on @xmath62 are linearly related , _
i.e. _ transition functions are linear maps , our starting equation ( 1 ) reduces to a _ free particle _ equation . when moving from a point @xmath55 to a point @xmath67 we go from the vector space structure associated with @xmath56 to the one associated to @xmath68 , this can be done _ transporting _ one vector space onto the other along the _ solution curve _
connecting @xmath67 to @xmath55 . depending on @xmath55 and @xmath67
this connection map may fail to be _ linear _ up to some order depending on the _ extension _ of the neighbourhood , the set of points we may reach from @xmath69 by using _
solution curves _ while keeping the _ linearity _ violated to no more than some preassigned power @xmath70 in the parameters will be a @xmath70-order local inertial frame .
when equation ( 5 ) is satisfied , this procedure defines a global linear inertial frame .
we start with an inertial frame , _
i.e. _ a reference inertial frame described by @xmath14 , giving rise to a global coordinate system @xmath71 . we notice that a particular form for @xmath12 would be @xmath72 .
it is worth stressing that a different choice of the congruence ( _ i.e. _ a different choice of @xmath11 ) in general may give rise to a different class of inertial systems .
having found one coordinate system in which the equation of motions have the form ( 2 ) , how many of them exist ?
it is clear we have to look for all coordinate systems @xmath73 such that @xmath74 as @xmath75 by using @xmath76 we find @xmath77 _ i.e. _ , @xmath78 must be linear in @xmath79 ( it can be however any function of constants of the motion for system ( 2 ) ) .
we consider therefore @xmath80 in terms of the initial conditions , we have @xmath81 where @xmath82 stays for @xmath83 . _
_ as constants of the motion , say , for our comparison dynamics satisfy @xmath84 , it is clear that @xmath78 can be any function of them in addition to the explicit dependence on ( @xmath85 ) . here
, we only consider the dependence on constants of the motion ( @xmath86 ) and do not consider a possible dependence on other constants of motion .
this is a simplifying assumption useful for computations because @xmath87 parametrize the position - velocity phase space .
because our approach to inertial frames is a dynamical one , there is no reason to restrict our transformations to be _ point transformations _
( _ i.e. _ to tangent bundle automorphisms , to use the language from differential geometry ) . therefore , _ new _ velocities need not be linear functions of the _ old _ velocities .
the congruence of curves corresponding to @xmath88 will be given by @xmath89 where ( @xmath90 ) can be any smooth function of @xmath91 .
at this point we look for linear transformations on a given pair @xmath14 with the requirement that any transformed one still satisfies @xmath92 , @xmath93 , _ i.e. _ our transformations generate new inertial frames satisfying the _ mutual objective existence_. from here we shall be able to construct the relativity groups which are compatible with our requirement . as usual
, we can dispose of the translation part and concentrate our analysis on the linear homogeneous part . by using a passive point of view
, we can consider our transformations from @xmath2 to @xmath2 , preserving the origin .
we shall therefore consider linear transformations @xmath94 , with the additional requirement @xmath95 , to implement @xmath92 , @xmath93 .
here we think of the choice @xmath96 .
we have a preliminary lemma : _ any invertible linear transformation , _
@xmath97 , _ which , along with its inverse , preserves the time - like character can be decomposed into the product of a linear transformation in the _ @xmath98_-plane and two space - like transformations_. we denote by @xmath99 our generic transformation and apply it to a standard vector : @xmath100 where @xmath101 by assumption .
now , by using a space - transformation @xmath102 , we can transform @xmath103 therefore @xmath104 by using a linear transformation @xmath105 in the @xmath98-plane , acting as the identity in the remaining components and preserving the _ time - like _ character , we find @xmath106 by using the arbitrariness of the starting time - like vector , we find that @xmath107 is a space - transformation and we get the decomposition @xmath108 this decomposition lemma allows us to deal first with transformations in the @xmath98 plane to find out which ones are compatible with the requirement on the _ objective existence _ condition and then to compose them .
this analysis will be done in the following sections .
having reduced our problem to a two - dimensional one , it is easy to visualize the situation .
a vector @xmath11 is given and we should consider all those linear transformations in @xmath109 which never take a vector transversal to @xmath11 ( defined by @xmath24 ) into one parallel to @xmath11 . if we think of a one - parameter group of transformations connecting two allowed frames , for any infinitesimal generator @xmath110 , there will never be a value of the parameter @xmath111 for which @xmath112 is parallel to @xmath11 . at this point
it is convenient to decompose any element of @xmath113 into the product of an element in @xmath114 and a dilation . because dilations are in the center of @xmath113 they can be dealt with separately .
thus , we may restrict our analysis to @xmath114 .
for this analysis it is very convenient to notice that @xmath114 is the same as @xmath115,_i.e . _ the group of canonical transformations in a two - dimensional phase - space . from this point of view
, a matrix @xmath110 is associated with some hamiltonian function whose level sets contain the orbits of the one - parameter group associated with @xmath110 @xcite .
it is now clear that the family of inertial frames we obtain from a given one with the action of the one - parameter group @xmath116 , is associated with a quadratic form , the hamiltonian function _ generating _ @xmath110 , up to a numerical factor .
thus we are led to analyse quadratic forms on @xmath109 in connection with the _ placement _ of @xmath117 in space - time , keeping in mind that their level sets contain the orbits of the _ relativity transformation group _ we are searching for .
the infinitesimal transformation @xmath110 associated with a transformation matrix @xmath118 which tends to the unit of the group when its parameters tend to zero , has the form @xmath119 the evolution of the vector @xmath120 is described by @xmath121 where @xmath122 if @xmath110 does not admit real eigenvalues , it will define a rotation - like transformation ( _ i.e. _ level sets of @xmath123 are ellipses ) and will violate the _ mutual objective existence _ ( in what follows _ m.o.e .
_ ) condition .
we assume therefore that @xmath110 has real eigenvalues .
we get : @xmath124 eigen - directions are defined by @xmath125 and , in terms of them , we may write @xmath126 the solution of eq . ( 7 ) corresponding to the initial conditions @xmath127
is given by @xmath128 moreover , the ratio @xmath129 , when @xmath130 , tends to @xmath131 , independently of the value of the initial conditions . as a consequence , these quantities are characteristic invariants of the transformation .
the quantities @xmath132 are also invariant .
finally , with the substitution @xmath133 , the transformation matrix in ( 9 ) takes the form @xmath134 here @xmath135 plays the role of parameter of the transformation ; no ambiguity arises if we continue to indicate it with @xmath136 .
the level set corresponding to @xmath137 determines the asymptotes of the branches of the hyperbolas defined by @xmath138 and @xmath139 .
the asymptotes coincide with the _ eigen - directions _ of @xmath99 .
it is now clear that @xmath99 , associated to @xmath123 , will define an acceptable relativity transformation only if @xmath24 intersects the branches corresponding to @xmath140 and @xmath11 the branches corresponding to @xmath141 ( or viceversa , and then we redefine the group parameter so that @xmath24 always intersects the negative branch ) .
this implies that @xmath99 ( see eq .
( 6 ) ) can not transform a vector whose second component is zero in a vector whose first component is zero . a priori , for a given pair @xmath14 , we shall find several quadratic functions @xmath123 satisfying previous requirements , therefore we may think that a combination of them would be also acceptable .
which combinations may be admissible will be discussed in next subsection .
we analyse the trajectories associated with a generic hamiltonian function ( 8) @xmath142 in the @xmath143 plane , with reference to fig . 1 . * _ a ) _ * @xmath144 ; no compatibility with the _ m.o.e .
_ condition ; this is the rotation - like case .
@xmath145 ; the hamiltonian reduces to @xmath146 : no compatibility with the _ m.o.e .
_ condition ; @xmath147 ; both eigen - directions are contained between @xmath11 and @xmath24 ; no compatibility with the _ m.o.e .
_ condition ; @xmath148 ; this condition selects candidates to be acceptable lorentz transformations . _
i ) _ two transformations @xmath99 and @xmath149 of the form ( 10 ) belong to the same transformation group only if they have identical invariants ; if this is not the case , the _ m.o.e .
_ existence is violated by the transformation we get composing some powers of them ; in fact , if , _ e.g. _ , @xmath150 , a _ world - line _ , admissible for @xmath149 , may be rotated by some power of @xmath99 into @xmath24 ( for the same reason lorentz - type transformations will not be compatible with carroll and/or galilei transformations ) .
_ ii ) _ in addition , the product is commutative if and only if @xmath151 and @xmath152 , that is to say if and only if the _ m.o.e .
_ condition is satisfied .
_ iii ) _ the addition rule for the unique parameter is : @xmath153 . * _ e ) _ * @xmath154 ; in this _ carroll _ case @xcite
@xmath155 _ i ) _ @xmath11 is an invariant asymptote ; no transformation can move @xmath11 into @xmath24 ; _ ii ) _ _ equi - locality _ is absolute ( @xmath156 ) . * _ f ) _ * @xmath157 ; in this galileian case @xmath158 _ i ) _ @xmath24 is an invariant asymptote , _ ii ) _ the product of two transformations of the same type,(with different values for @xmath159 and @xmath136 ) satisfies also the _ m.o.e . _ condition ; it iii ) _ simultaneity _ is absolute ( @xmath160 ) . * _ g ) _ * @xmath161 ; this aristotelian transformation preserves both _
equi - locality _ and _ simultaneity _ and is compatible with lorentz , galilei and carroll .
inertial frames connected by a relativity transformation should be considered to be equivalent ; therefore we make a further requirement .
we impose that inertial frames connected by an allowed relativity transformation should have identical relative clocks , _
i.e. _ we require that the two maps @xmath162 and @xmath163 coincide for any two inertial frames we obtain starting with @xmath14 .
this requirement will impose @xmath164 . therefore our analysis is now greatly simplified .
the argument for @xmath164 is simple ; eq .
( 9 ) implies that previous condition is satisfied if @xmath165 that is to say if @xmath166 .
this equation , which is independent of the restriction to a bidimensional space - time , implies that dilation along the time - axis should be excluded from our relativityu transformations , _
i.e. _ our infinitesimal transformations should not contain @xmath167 .
going back to a bi - dimensional space - time , we find that the hamiltonian ( 8) assumes the form @xmath168 while @xmath99 reduces to @xmath169 we conclude that our ( 1,1 ) _ space - time _ imay be equipped with an invariant quadratic form given by @xmath170 in the ( generic ) lorentz case and by @xmath171 in the galileian case .
the physical interpretation of these transformations and the related problems , like the clock synchronization one , have been extensively discussed in the literature @xcite , @xcite , @xcite .
we conclude this section by remembering that if space and time coordinates of the universe are deformed in such a manner that all the space - time coincidences are conserved , then the universe remains unchanged @xcite .
notice that this requirement is satisfied already at the level in which only the _ m.o.e .
_ is supposed .
we have seen that , by using the decomposition lemma in section 5 , we have been able to select transformations compatible with our requirement of equality of relative clocks and _ mutualobjective existence _ in the @xmath98-space time . here
we would like to find out which are the implications in four dimensions , if we compose it with space transformations .
we first consider the @xmath172-galilei group .
it is clear that because our transformations preserve a space - slicing ( _ i.e. _ an absolute notion of simultaneity ) , any linear transformation along the space part will be an acceptable relativity transformation .
we find for the homogeneous _
generalized _ galilei group the semidirect product @xmath173 where @xmath174 stays for the three dimensional space of velocities . by defining the action on a vector space - time
@xmath175 we find @xmath176=[(b , a\vec{x}+b\vec{v})]\ ] ] along with the composition rule @xmath177 the rotation group of the standard galilei group is being replaced by the general linear group in three dimensions , i. e. the _ m.o.e .
_ condition does not require any preferred notion of distance along the space part of space - time .
a symmetric @xmath178tensor of the type @xmath171 is invariant under the action of @xmath179 .
now we consider the other compatible group , _
i.e. _ the _ lorentz type _ transformations . here , to carry on computations , we find more convenient to use infinitesimal transformations in terms of vector fields . for each plane involving a time coordinate and
a space coordinate , our @xmath172 analysis provides us with the following vector fields : @xmath180 it is convenient to redefine coordinates by setting @xmath181 ; we find then @xmath182 now we look for space transformations whose commutator with @xmath183 s does not violate the _ m.o.e .
_ condition .
we find @xmath184= a^{k}_{j}y^{j}\frac{\partial}{\partial y_{0 } } -y_{0}a^{j}_{k}\frac{\partial}{\partial y^{j}}=\nonumber\\ & = a^{k}_{j}\left(y^{j}\frac{\partial}{\partial y_{0}}+ y_{0}\frac{\partial}{\partial y^{j}}\right ) -(a^{k}_{j}+a^{j}_{k})y_{0}\frac{\partial}{\partial y^{j}},\nonumber\end{aligned}\ ] ] _ i.e. _ we find a combination of _ boosts _ and carrol transformations .
to preserve the _
_ condition we have to require @xmath185 ; in conclusion , the most general space transformations compatible with _ boosts _ to preserve the objective conditions are just rotations .
we find that our allowed relativity group is the lorentz group .
therefore our space time gets equipped with a generalized minkowskian metric .
the associated symmetric @xmath178tensor has the form @xmath186 .
some further intermediate situations are possible .
they correspond to the use of galilei - type trasnformations in some _ one time - ome space _ planes and lorentz type in the remaining ones .
these situatiomns are obtained when some of the coefficients of the rprevious quadratic form are being set equal to zero , say @xmath187 and @xmath188 being different from zero , or @xmath189 , @xmath188 being different from zero .
we have found that the notion of _ mutual objective existence _ along with the identity of relative clocks is enough to select only the galilei and lorentz transformations in one space and one time setting .
these transformations preserve a quadratic form which is degenerate in the galileian case and not degenerate in the lorentz case . when going to @xmath190 , we find that the _ mutual objective existence _ condition determines a minkowski - type metric in the case of the lorentz group , while the galilei - type group does not impose restrictions on the space structure .
some intermediate situations are also possible .
it is possible to have lorentz - type behaviour in some directions and galilei - type behaviour in the complementary directions .
these mixed situations can not be ruled out only on the basis of the _ mutual objective existence _ condition and some additional physical insight is needed .
we hope we have made clear how a _ general relativity ideology _ may be useful in dealing also with _ special relativity _
, where only an affine structure for space - time is needed .
the authors are grateful to w. m. tulczyjew for useful discussions , to e.j .
saletan for reading the manuscript and to the partecipants of the _ fryday seminars _ where these results were presented in several occasions and helped us to clarify several subtle points .
thanks are also due to alessandro preziosi ( 9 years old , grandchild of one of the authors ) for his computer drawing .
h. reichenbach , _ philosophie der raum - zeit - lehre _ , walter de gruyter , berlin and liepzig , 1928 , translated by m. reichenbach and j. freund as _ the phylosophy of space and time _ , dover , new york , n.y . | _ a free system , considered to be a comparison system , allows for the notion of objective existence and inertial frame .
transformations connecting inertial frames are shown to be either lorentz or generalized galilei . _ = 14truecm = 23truecm _ key words _ : inertial frames , relativity transformations , projective and metric spaces .
pacs : 02.40 - euclidean and projective geometries 03.30 - special relativity .
+ mailing address : + prof .
bruno preziosi + dipartimento di scienze fisiche , universit di napoli `` federico ii '' + mostra doltremare , pad.20 + i-80125 napoli ( italy ) + tel.n . : ( + 39 ) 81 - 7253416 , fax n. : ( + 39 ) 81 - 2394508 + e - mail : [email protected] + |
the stability of a soliton solution , which we denote here as @xmath97e^{-i\mu t}\end{aligned}\ ] ] splitting the real and imaginary parts , i.e. @xmath98re@xmath99 and @xmath98re@xmath99 , was established using the linear stability analysis . to this end , following the standard algorithm , we make the ansatz ( @xmath100 ) @xmath101 e^{-i\mu t}\end{aligned}\ ] ] where @xmath102 is a small perturbation in the coupled gross - pitaevskii equations @xmath15 and linearize them with respect to small @xmath103 .
this gives us the evolution problem for the perturbation @xmath103 which in the matrix form reads @xmath104 with the matrix operator @xmath105 given by finally , we look for a solution of ( [ linear1 ] ) ( [ linear3 ] ) in the form @xmath109 where @xmath110 is a constant spectral parameter and @xmath111 is @xmath112-column time - independent vector , what gives us the linear spectral problem : @xmath113 the eigenvalue problem ( [ linear4 ] ) was solved numerically to obtain the dependence of the growth rate @xmath110 on the chemical potential @xmath17 . the soliton is unstable if the respective @xmath110 has a real positive part . | we report a diversity of stable gap solitons in a spin - orbit coupled bose - einstein condensate subject to a spatially periodic zeeman field . it is shown that the solitons , can be classified by the main physical symmetries they obey , i.e. symmetries with respect to parity ( p ) , time ( t ) , and internal degree of freedom , i.e. spin , ( c ) inversions . the conventional gap and gap - stripe solitons are obtained in lattices with different parameters .
it is shown that solitons of the same type but obeying different symmetries can exist in the same lattice at different spatial locations .
pt and cpt symmetric solitons have anti - ferromagnetic structure and are characterized respectively by nonzero and zero total magnetizations .
the progressively growing interest in the physics of mixtures of spinor bose - einstein condensates ( becs ) and in particular in spin - orbit ( so ) coupled becs ( so - becs ) @xcite , is motivated on the one hand by their fundamental importance for the atomic physics , and on the other hand by the rich possibilities they offer for emulating synthetic electric and magnetic fields .
the last property allows one using neutral atoms to simulate numerous condensed matter phenomena @xcite in a tunable way , what would be impossible in direct condensed matter experiments .
meanwhile , the physics of becs is characterized by two essential factors , first , a bec is a nonlinear system , with the nonlinearity stemming from the two - body interactions @xcite . moreover , by changing the configuration of the system ( i.e. of the laser beams ) one can create effective nonlinear interactions of different types @xcite . respectively , such nonlinear objects as skyrmions @xcite , solitons @xcite , anti - ferromagnetic structures and symmetry breaking @xcite , have been recently reported for so - becs .
the second feature , is that a ( quasi- ) stationary state of the condensate requires the presence of the external potential .
this , in particular , imposes constraints on the lower bound of the kinetic energy , what is particularly relevant for so - becs ( see e.g. @xcite ) .
what concerns the external potentials , since the very first experiments with cold atoms loaded in optical lattices @xcite it was widely recognized that lattices ( i.e. periodic potentials ) are particularly efficient in manipulating becs @xcite .
nowadays , the nonlinear properties of the atomic becs held in optical lattices , described in the mean - field approximation are very well studied @xcite ( see also @xcite for a brief review on bec mixtures in optical lattices ) .
it is therefore natural that studies of so - becs in periodic potentials were already initiated .
manipulations with the band - gap structure @xcite and vortex lattice states @xcite supported by so coupling and optical lattices have being reported .
on the other hand a few years ago there have been initiated studies of the properties of electronic gas with rashba @xcite and dresselhaus @xcite so coupling in periodic structures .
those studies included the effect of dimensionality and features of the band - gap structures of a gas in a superimposed periodic and arbitrary confining potentials @xcite , ballistic transport of electrons in periodic magnetic field @xcite , switching controlled by so coupling @xcite , and quantum wire superlattices @xcite .
it was shown that so coupling induced by an external periodic field leads to the appearance of new gaps , making the electronic system tunable even in the condensed matter settings .
even more recently , the so coupling was emulated in experiments with light propagating in arrays optical waveguides @xcite , thus opening a root towards new physical systems where the interplay of so coupling and nonlinearity ( if waveguides are of the kerr type ) can result in novel optical phenomena . bearing in mind the described findings it is natural to address the effect of the lattice induced by such a field on the _ nonlinear _ properties of so - becs .
, classified with respect to their parity ( p ) , time ( t ) , and pseudo - charge ( c ) symmetries .
the chemical potentials of the solitons originate from different gaps in the spectrum of the magnetic lattice for both attractive and repulsive interactions .
we address one - dimensional ( 1d ) mixture of a two - component so - bec which is described by a spinor @xmath0col@xmath1 .
in particular , for a tripod - type atomic scheme , like the one considered in @xcite , the components @xmath2 and @xmath3 are the wavefunctions of the two dark states emulating ( pseudo-)spins @xmath4 and @xmath5 . adopting the units where @xmath6 , and taking into account that the difference between spin - independent and spin - dependent two - body interactions can be made as small as necessary ( in @xcite the difference was of order of 1% of the interaction magnitude ) we explore the hamiltonian of the system in the form @xmath7 where the linear hamiltonian @xmath8 accounts for the periodically modulated zeeman field @xmath9 and for so coupling with the strength @xmath10 ( which depends on the angle between optical beams and can be tuned @xcite )
, @xmath11sign@xmath12 characterizes the of two - body interactions with the scattering length @xmath13 , and @xmath14 are the pauli matrices .
the nonlinear dynamics of the condensate is then governed by the coupled gross - pitaevskii equations ( gpes ) @xmath15 we are interested in the stationary solutions having the form @xmath16 , where @xmath17 is a chemical potential and @xmath18 solves the stationary coupled gpes . for the sake of definiteness , we consider a particular form of the zeeman field @xmath19,\end{aligned}\ ] ] where @xmath20 and @xmath21 are the average and the modulation amplitude of the field , respectively , and without loss of generality the period is taken to be @xmath22 .
nevertheless we emphasize that the approach developed below is valid for more generic forms of zeeman fields obeying the symmetries specified below .
first we take into account that @xmath9 is an even function , @xmath23 and obtain that if @xmath18 is a solution of eq .
( [ gpe ] ) then the spinor @xmath24 ( an overbar stands for the complex conjugation ) defined by the parity and time inversions @xcite , is a solutions of the stationary gpes , as well .
at @xmath25 the field ( [ omega ] ) allows for additional symmetry reductions .
the first one is readily seen in the shifted spatial variable @xmath26 .
indeed , since @xmath27 , if @xmath28 is a solution of ( [ gpe ] ) then the function obtained through the reduction @xmath29 , is also a solution ( @xmath30 by analogy with the charge symmetry @xcite indicates the symmetry related to the inversion of the internal degree of freedom , which is the pseudo - spin in our case ) .
finally , due to the property @xmath31 one more symmetry stems from the possibility of flipping spin components in the @xmath32 shifted lattice : @xmath33 is a solution of the stationary gpes .
bearing in mind the symmetries of the gpes it is natural to use them for classifying the solutions , i.e. to identify solutions which would be the eigenstates of the respective @xmath34 , @xmath35 or @xmath36 operators .
it is important , however , that the pt and cpt symmetries imposed on the solutions , rule out the phase invariance which is an attribute of the gpes , because @xmath37 with a real @xmath38 is not commutative with the @xmath39 operator .
this artificial constraint on the solutions of the gpes can be avoided by generalizing the definition of the symmetric modes .
namely , we define a pt [ cpt ] symmetric solution , @xmath40 [ @xmath41 , as solutions for which there exist a constant phase @xmath38 transforming it into eigenfunctions of the respective symmetry operator , i.e. ensuring @xmath42 [ @xmath43 .
now we turn to a possibility of bifurcations of nonlinear modes obeying the above symmetries from the respective linear bloch spinors @xmath44col@xmath45 which solve the spectral problem @xmath46 , where @xmath47 refers to the wave - vector in the reduced brillouin zone ( bz ) , while @xmath48 indicates the number of the branch corresponding to the given @xmath47 with @xmath49 being the lowest one .
the symmetries described above obviously hold also for the eigenfunctions of @xmath50 . in fig .
[ fig1 ] we show three typical band - gap spectra @xmath51 . in the absence of the constant
splitting ( @xmath25 ) and at relatively weak so coupling the lower band - edges are reached only at the boundary of the bz ( fig .
[ fig1]a ) .
increase of the so coupling results in closure of the first finite gap and appearance of the crossing points in the spectrum , as well as in a shift of the band - gap edges to some internal points of the bz ( fig .
[ fig1]b ) , which are symmetric with respect to the center of the bz and will be denoted as @xmath52 . finally , in fig .
[ fig1]c we observe `` re - opening '' of the first gap due to nonzero constant component of the zeeman field @xmath53 .
notice that the lowest branches of the spectra in figs .
[ fig1]b and [ fig1]c resemble the well known spectrum of the homogeneous so - bec . with the periodic potential ( [ omega ] ) for @xmath54 and ( a ) @xmath55 , @xmath25 ( b ) @xmath56 , @xmath25 , ( c ) @xmath56 , @xmath57 .
horizontal gray lines indicate the band - gap edges .
red and blue arrows indicate the points from which gap solitons bifurcate in the case of attractive and repulsive nonlinearities , respectively . ] when the gap edge corresponds to the bz boundary @xmath58 ( fig . [ fig1]a ) due to the pt - symmetry the bloch spinor can be chosen as @xmath59col@xmath60 , where @xmath61 and @xmath62 are real @xmath63periodic functions .
such bloch states give rise to pt - symmetric gap solitons : @xmath64 , whose common property is that each spin component has symmetric distribution of the atomic density .
indeed , introducing @xmath65 of the densities of the spinor components one verifies that @xmath66 .
for the case of attractive interactions such soliton belonging to the semi - infinite band is illustrated in fig .
[ fig2]a .
notice , that while this situation is seemingly similar to the one observed in the homogeneous medium @xcite the effective periodic potential induced by the zeeman field is inverted for the different components resulting in an immiscible anti - ferromagnetic phase ( the atoms of each component are located in the minima of the respective periodic potentials ) .
a counterpart of the described solitons can be also obtained in a finite gap in the case of repulsive inter - atomic interactions , see fig .
[ fig2]b .
although there exists a general qualitative similarity with the soliton in the semi - infinite gap ( c.f .
fig .
[ fig2]a ) now the domains of the atomic concentration ( i.e. the points where density is maximal ) , especially in the second component , do not coincide with the local minima of the effective potential for the respective component . for the attractive nonlinearity ( a ) ; pt - symmetric ( b ) and cpt - symmetric ( c ) solitons with @xmath67 originating from the first finite gap for the repulsive nonlinearity , and soliton with @xmath68 from the first finite gap for attractive nonlinearity ( d ) .
in all cases @xmath54 , @xmath55 , @xmath25 . the dashed line shows the lattice profile for the @xmath69 component . ] turning to @xmath35-symmetric solitons , we observe that their characteristic feature is the equal number of atoms @xmath70 in the both components , @xmath71 , since the densities are now related as : @xmath72 . .
using this definition we conclude that @xmath35-symmetric solitons have zero total magnetization .
an example of such a gap soliton emerging from the first finite gap in a repulsive condensate is presented in fig .
[ fig2]c .
in such modes the maxima of density distributions in the two components are separated by a half of the lattice period and are located in neighboring extrema of the zeeman lattice @xmath9 . finally , in fig .
[ fig2]d we show a pt - symmetric soliton emerging from the first finite gap in the so - bec with attractive interactions .
now the symmetries of the spinor components are `` inverted '' as compared to the case shown in fig .
[ fig2]a , i.e. now @xmath73 and @xmath74 .
the reason for this `` inversion '' is that the soliton bifurcates from the upper gap edge ( see the upper red arrow in fig . [ fig1]a ) , thus resulting in the change of the sign of the effective mass .
the total numbers of atoms and magnetizations as functions of the chemical potential are shown in fig .
[ fig3 ] . for all solitons
@xmath75 demonstrates the typical behavior known for the one - component small - amplitude gap solitons @xcite : near @xmath48-th gap edge @xmath76 where @xmath77 .
this indicates that the envelopes @xmath78 of all bifurcating solitons in the small amplitude limit , @xmath79 , are described by the nonlinear schrdinger ( nls ) equation . in the panels b and c one
observes that the branches of pt and cpt symmetric solitons seemingly merge .
this however occurs only on the plane ( @xmath80 ) and do not represent a bifurcation .
the both types of the solutions bifurcate from the different linear modes ( obeying the respective symmetries ) .
coincidence of the branches close to the gap edge is explained by the @xmath36-symmetry ensuring the equality of the spatially averaged densities of the two components in the linear bloch spinors , from which the respective solitons bifurcate .
cp and cpt solitons in the first finite gap have also different stability @xcite . in the case of repulsive interactions pt solitons are stable ( unstable ) at lower ( larger ) amplitudes .
cpt solitons are unstable , while in the vicinity of the bifurcation point they can be classified as metastable ones as the instability increment is small enough ( @xmath81 ) within the region of width @xmath82 adjacent to the upper gap edge located at @xmath83 . in the condensate with negative scattering length
the cpt - symmetric solitons are always unstable while in the vicinity of the band - edge , i.e. in the small - amplitude limit ( with the width of the metastable domain @xmath84 ) the pt - symmetric solitons can are metastable . for pt - symmetric solitons emerging from the semi - infinite gap in the case of attractive nonlinearity ( a ) , for pt and cpt symmetric solitons emerging from the first finite gap in the case of repulsive ( b ) and attractive ( c ) nonlinearity . in ( c ) magnetization
is multiplied by 10 .
circles correspond to solitons shown in fig .
[ fig2 ] . in all cases
@xmath54 , @xmath55 , @xmath25 . ]
the magnetization shows rather complex behavior . for sufficiently small detuning @xmath85
one observes the behavior approximated by the law @xmath86 for all types of the reported pt - symmetric solutions ( recall that for the cpt - solitons , like the one shown in fig .
[ fig2]c , the total magnetization is zero ) .
as above , this is explained by the symmetry @xmath36 , ensuring that at small amplitudes the states @xmath69 and @xmath87 are equally populated . at large detuning the magnetization may start to decay ( see fig . [ fig2]b ) .
obtained for attractive nonlinearity .
( b ) the number of particles and magnetization _ vs _ @xmath17 in the attractive condensate .
( c ) a gap - stripe soliton from the first finite gap at @xmath88 for repulsive nonlinearity corresponding to the circle in panel ( d ) , which shows @xmath75 _ . vs _
@xmath17 for repulsive and attractive interactions in the first finite gap . in all cases
@xmath54 , @xmath56 , @xmath57 . ] now we turn to the situation where a band - gap is achieved in some internal points @xmath52 corresponding to a double degenerate eigenvalue @xmath89 ( see fig . [ fig1]b and [ fig1]c ) . in the homogeneous case similar situation gives origin to a stripe - phase @xcite in repulsive condensate or to a stripe - soliton @xcite in a condensate with a negative scattering length . in order to construct a pt - symmetric solution we depart from one of the superposition of the linear spinors ,
say @xmath90 , where the relation between the bloch states is given by @xmath91 .
pt - symmetric gap - stripe solitons belonging to the branch bifurcating from @xmath92 are illustrated in fig .
[ fig4 ] for the lattice whose band - gap structure is presented in fig .
[ fig1]c .
gap - stripe soliton in the semi - infinite gap ( upper row of fig .
[ fig4 ] ) are stable in the whole domain of the existence . in the finite gap the solitons are generically unstable for the depth of potential considered here ,
although we have observed ( numerically ) stable time evolutions of the gap - stripe solitons in the case of repulsive interactions which was accompanied by regular oscillations of the amplitude , indicating on a possibility of a breather solution . comparing the gap - stripe solitons with the `` conventional '' gap solitons considered above , we find that while the envelope of stripe - solitons is also described by the nls equation , i.e. @xmath93 close to the gap edge , the total magnetization now displays very different behavior near the bifurcation point : it approaches a nonzero value .
this difference with the cases of the pt - symmetric solitons follows from the structure of the linear bloch spinors , which now do not obey s - symmetry and thus the states @xmath69 and @xmath94 have different populations . in other words
such states have nonzero magnetization even in the linear limit .
to conclude we have found a diversity of stable gap and gap - stripe solitons in spin - orbit coupled bose - einstein condensate subject to a spatially periodic zeeman field .
the obtained solutions can be classified by the major physical symmetries , as being invariant with respect to either pt or cpt reductions .
the respective solitons have either nonzero or zero total magnetization .
the conventional gap and gap - stripe solitons are observed in lattices with different parameters , while solitons of the same type but obeying different symmetries can be observed in the same lattice but at different spatial locations .
all the obtained solutions obey an anti - ferromagnetic structure and thus represent immiscible phases .
vvk acknowledges support of the fct ( portugal ) grant pest - oe / fis / ui0618/2011 .
fka acknowledges support from the fundao amparo e pesquisa de so paulo(fapesp , brasil ) .
99 y. j. lin , k. jimenez - garcia , and i. b. spielman , nature * 471 * , 83 ( 2011 ) .
y .- j .
lin , r. l. compton , a. r. perry , w. d. phillips , j. v. porto , and i. b. spielman , phys .
rev .
lett . *
102 * , 130401 ( 2009 ) ; y .- j .
lin , r. l. compton , k. jimenez - garca , j. v. porto , and i. b. spielman , nature ( london ) * 462 * , 628 ( 2009 ) ; c. h. l. quay , t. l. hughes , j. a. sulpizio , l. n. pfeiffer , k. w. baldwin , k. w. west , d. goldhaber - gordon , and r. de picciotto , nature phys .
* 6 * , 336 ( 2010 ) .
p. hauke , f. m. cucchietti , l. tagliacozzo , i. deutsch , and m. lewenstein , rep .
prog .
phys . * 75 * , 082401 ( 2012 ) .
r. keil , j. m. zeuner , f. dreisow , m. heinrich , a. tnermann , s. nolte , and a. szameit .
nature commun .
* 4 * , 1558 ( 2013 ) . c. pethick and h. smith , _ bose - einstein condensation in dilute gases _
( cambridge university press : cambridge , 2002 ) ; l. p. pitaevskii and s. stringari , _ bose - einstein condensation _
( clarendon press : oxford and new york , 2003 ) . y. zhang , li mao , and c. zhang , phys . rev .
lett . * 108 * , 035302 ( 2012 ) .
t. kawakami , t. mizushima , m. nitta , and k. machida , phys .
rev .
lett . * 109 * , 015301 ( 2012 ) .
v. achilleos , d. j. frantzeskakis , p. g. kevrekidis , and d. e. pelinovsky , phys .
rev .
lett . * 110 * , 264101 ( 2013 ) y.
xu , y. zhang , and b. wu phys .
rev .
a * 87 * , 013614 ( 2013 ) .
d. a. zezyulin , r. driben , v. v. konotop , and b. a. malomed , phys .
rev . a ( 2013 ) ( in press ) b. p. anderson and m. a. kasevich , science * 282 * , 1686 ( 1998 ) .
o. morsch and m. oberthaler , rev .
mod .
phys . * 78 * , 179 ( 2006 ) .
v. a. braznyi and v. v. konotop mod .
phys .
lett .
b * 18 * , 627 ( 2004 ) ; r. carretero - gonzlez , d. j. frantzeskakis , and p. g. kevrekidis nonlinearity , * 21 * r139 ( 2008 ) .
h. a. cruz , v. a. brazhnyi , v. v. konotop , and m. salerno , physica d , * 238 * , 1372 ( 2009 ) and references therein .
y. zhang and c. zhang , phys .
rev .
a * 87 * , 023611 ( 2013 ) h. sakaguchi and ben .
li , arxiv:1301.3565 [ cond-mat.quant-gas ] y. a. bychkov and e. i. rashba , j. phys .
c * 17 * , 6039 ( 1984 ) .
g. dresselhaus , phys . rev . * 100 * , 580 ( 1955 ) .
s. smirnov , d. bercioux , and m. grifoni , epl * 80 * , 27003 ( 2007 ) ; v. ya .
demikhovski . ,
d. v. khomitsky , and a. a. perov , low temp . phys .
* 33 * , 115 ( 2007 ) .
j .- f .
liu , w .- j .
deng , k. xia , c. zhang , and z. ma , phys .
rev .
b * 73 * , 155309 ( 2006 )
. x. b. xiao , x.m .
li , and y.g .
chen , physics b * 404 * , 4159 ( 2009 ) .
s. j. gong and z. q. yang , j. phys : cond . mat . * 19 * , 446209 ( 2007 ) .
g. thorgilsson , j. c. egues , d. loss , and s. i. erlingsson , phys .
rev .
b * 85 * , 045306 ( 2012 ) .
m. j. edmonds , j. otterbach , r. g. unanyan , m. fleischhauer , m. titov , and p. hberg , new j. phys . *
14 * , 073056 ( 2012 ) .
j. dalibard , f. gerbier , g. juzelinas , and p. hberg , rev .
mod .
phys . * 83 * , 1523 ( 2011 ) .
the wigner time inversion operator @xmath95 applied to a stationary mode @xmath96 is reduced to the complex conjugation .
v. b. berestetskii , l. p. pitaevskii , and e. m. lifshitz `` quantum electrodynamics '' ( elsevier , 1982 ) .
see e.g. l. d. landau , e. m. lifshitz and l. p pitaevskii , `` electrodynamics of continuous media '' ( elsevier , 2004 ) .
t. l. ho and s. zhang , phys .
rev .
lett . * 107 * , 150403 ( 2011 ) ; s. sinha , r. nath , and l. santos , phys .
rev .
lett . * 107 * , 270401 ( 2011 ) .
y. li , l. p. pitaevskii , and s. stringari , phys .
rev .
lett . * 108 * , 225301 ( 2012 ) . |
29 fazekas p and anderson p w 1974 _ philos . mag .
_ * 30 * 423 .
chandra p and coleman p 1991 _ phys .
* 66 * 100 .
kawamura h 1998 _ j. phys . : condens . matter _ * 10 * 4707 .
wannier g h 1950 _ phys .
rev . _ * 79 * 357 .
kano k and naya s 1953 _ prog .
phys . _ * 10 * 158 .
stephenson j 1964 _ j. math .
phys . _ * 5 * 1009 . s@xmath51t@xmath52 a 1981 _ z. phys .
b - condensed matter _ * 44 * 121 .
barry j h , khatun m and tanaka t 1988 _ phys .
b _ * 37 * 5193 .
kawamura h and miyashita s 1984 _ j. phys .
_ * 53 * 4138 southern b w and xu h - j 1995 _ phys .
b _ * 52 * r3836 ( and references therein ) . xu h - j and southern b w 1996 _ j. phys . a : math .
* 29 * l133 ( and references therein ) .
chalker j t , holdsworth p c w and shender e f 1992 _ phys . rev .
lett . _ * 68 * 855 .
moessner r 2001 _ can .
_ * 79 * 1 .
miyashita s and kawamura h 1985 _ j. phys .
jpn . _ * 54 * 3385 .
stephan w and southern b w 2000 _ phys .
b _ * 61 * 11514 .
stephan w and southern b w 2001 _ can .
j. phys . _
* 79 * 1459 .
kuroda a and miyashita s 1995 _ j. phys .
_ * 64 * l4509 .
bekhechi s and southern b w 2003 _ phys .
b _ * 67 * 144403 .
bekhechi s and southern b w 2003 _ phys .
b _ * 67 * 2124xx ( in press ) .
olive j a , young a p and sherrington d 1986 _ phys . rev .
b _ * 34 * 6341 .
hinrichsen h 2000 _ adv .
phys . _ * 49 * 815 ( and references therein ) . | the technique of damage spreading is used to study the phase diagram of the easy axis anisotropic heisenberg antiferromagnet on two geometrically frustrated lattices .
the triangular and kagome systems are built up from triangular units that either share edges or corners respectively .
the triangular lattice undergoes two sequential kosterlitz - thouless transitions while the kagome lattice undergoes a glassy transition . in both cases ,
the phase boundaries obtained using damage spreading are in good agreement with those obtained from equilibrium monte carlo simulations .
frustrated antiferromagnets are particularly interesting because they allow novel kinds of low - temperature magnetic states to develop @xcite which are quite different from those observed in conventional magnets @xcite . when the frustration is due entirely to the geometry of the lattice
, exact results for the @xmath0 ising model on the triangular @xcite and kagome @xcite lattices have shown that there is no long range order at any temperature and the system has a macroscopic ground state degeneracy . in the triangular case , the spin - spin correlation function decays with a power law @xcite whereas , in the kagome geometry , it decays exponentially at zero temperature @xcite . for isotropic vector spin models ( xy or heisenberg ) on a triangular lattice @xcite where the elementary triangles share edges , the frustration is partially relieved and a noncollinear planar ground state forms with neighbouring spins making angles of @xmath1 .
finite temperature transitions are evident in both cases ( xy and heisenberg ) but are topological in nature . however , for the kagome lattice , the cornering sharing geometry leads to a disordered ground state for the heisenberg model with the phenomenon of order by disorder occuring in the limit @xmath2 with fluctuations favouring coplanar spin configurations @xcite in order to explore the heisenberg model on these geometrically frustrated lattices , various types of perturbations have been applied and have shown a strong effect on the ground state manifold @xcite .
we consider the following anisotropic hamiltonian @xmath3 where @xmath4 represents a classical three component spin of unit magnitude located at each site @xmath5 of a ( triangular / kagome ) lattice and the exchange interactions are restricted to nearest - neighbour pairs of sites .
the parameter @xmath6 describes the strength of the exchange anisotropy .
we restrict our attention to the case where @xmath7 represents an easy - axis anisotropy and we measure temperature and energy in units of @xmath8 .
the limit @xmath9 corresponds to the isotropic heisenberg model whereas the limit @xmath10 corresponds to an infinite spin ising model .
monte carlo studies have shown for the triangular lattice the existence of two distinct kt types of defect - mediated phase transitions at finite temperature @xcite for @xmath7 . by examining the limit @xmath9
, the transition at the heisenberg point also appears to be purely topological in character @xcite but with exponentially decaying spin correlations in the low temperature phase .
studies of the same model on the kagome lattice @xcite indicate the existence of a finite , but very low , temperature ferromagnetic transition for intermediate values of @xmath6 .
static results for the magnetization , susceptibility and specific heat indicate an ising - like transition but with a nonuniversal critical behaviour @xcite in which the values of the exponents vary with the strength of the anisotropy @xmath6 .
the individual spins do not exhibit any long ranged spatial order and resemble a glassy phase .
an analysis of the time evolution of the double time spin auto - correlation function following a quench of the system from infinite temperature into a non - equilibrium state at low temperature displays the phenomenon of aging which is a characteristic of many glassy systems .
additional evidence for the glassy nature of this low temperature state has also recently been found in a study of relationship of the spin autocorrelation function to the response of the system to an external magnetic field @xcite .
a violation of the fluctuation dissipation theorem is observed with properties similar to both structural glasses and spin glasses . versus temperature for the triangular lattice with @xmath11 and two differents values of the anisotropy ( a ) @xmath12 and ( b ) @xmath13 for different initial damage : @xmath14(plus ) and @xmath15(crosses ) .
the insets show the average distance associated with the transverse components.,title="fig:",width=453,height=347 ] versus temperature for the triangular lattice with @xmath11 and two differents values of the anisotropy ( a ) @xmath12 and ( b ) @xmath13 for different initial damage : @xmath14(plus ) and @xmath15(crosses ) .
the insets show the average distance associated with the transverse components.,title="fig:",width=453,height=347 ] in the present work we present a numerical study of the model described by hamiltonian ( 1 ) for both lattices using a damage spreading algorithm . in the damage spreading method @xcite ,
the damage distance between two different initial spin configurations is measured as they evolve with the same thermal noise .
the distance between two configurations is defined in the following way : if @xmath16 and @xmath17 are the spins of the two different replicas , then the distance between them at time @xmath18 is defined as @xmath19 where @xmath20 is the total number of spins and @xmath21 .
the procedure of the spreading of damage is as follows : for a given value of the anisotropy @xmath6 and temperature @xmath22 we first initialize the system @xmath16 in a random configuration and then let it evolve to a steady state . then we make two additional copies of the system , @xmath17 and @xmath23 .
we introduce damage in both systems @xmath24 and @xmath25 by inverting a fraction of the spins .
the two copies @xmath24 and @xmath25 correspond to cases where ( 1 ) a randomly chosen spin is flipped with @xmath26 and ( 2 ) all the spins are flipped with @xmath27 .
given these different initial conditions we let all the configurations evolve in time according to the same dynamics , i.e , the same rule and the same random number sequence in the monte carlo procedure . after a relaxation time needed for the damaged copies to be thermalized , we monitor the damage in order to calculate the time average @xmath28 of the distance ( 2 ) .
this procedure is repeated for many different samples ( initial configurations @xmath6 and damaged configurations @xmath24 and @xmath25 ) . at high temperatures , the averages @xmath29 and @xmath30 are equal but at low temperatures they approach different values .
we interpret this result as indicating that the free energy landscape has a single valley at high temperature and that a transition to a more complicated multi - valley landscape occurs as the temperature is lowered .
figure 1 shows the geometry of the two lattices for different sizes @xmath31 . in both cases
the number of sites @xmath32 and periodic boundary conditions are applied .
our simulations have been performed for sizes ranging from @xmath33 to @xmath34 , but all of the results are presented with @xmath11 where finite size effects are sufficiently small and a larger number of monte carlo sweeps ( mcs ) are possible .
for each temperature we calculate the time average of the damage using 10000 mcs and typically 200 samples were used to compute the averages . in the case of the triangular lattice , recent monte carlo simulations
@xcite have shown the existence of two finite temperature kt transitions .
the upper transition @xmath35 corresponds to the onset of power law correlations for the spin components parallel to the easy axis @xmath7 and the lower transition @xmath36 corresponds to the onset of power law correlations for the perpendicular components . in both cases there
is a corresponding spin stiffness coefficient which vanishes . in order to identify both transitions
, we also compute the following measure of the damage distance associated with the transverse ( xy ) components of the spins in the two replicas @xmath37 as shown in figure 2(a ) for the isotropic heisenberg case @xmath38 both distances are independent of the initial damage at high temperature but become dependent on it below the same temperature @xmath39 which agrees quite well with previous estimates of the transition temperature using equilibrium methods @xcite .
however , for values of @xmath7 , three different temperature regions are observed as shown in figure 2(b ) for the case @xmath13 : ( i ) @xmath40 , both distances depend on the initial damage , ( ii ) @xmath41 , @xmath42 is independent of the initial damage whereas @xmath43 still depends on the initial damage and ( iii ) @xmath44 , both distances become independent on the initial damage .
the damage spreading approach is able to identify both transitions which , in this case , correspond to defect unbinding transitions and it also predicts a finite temperature transition in the heisenberg limit . our results for the kagome lattice are shown in figure 3 . in this case
there are only two phases : a low temperature phase @xmath45 where the value of @xmath43 is dependent on the initial value of @xmath46 and a high temperature phase @xmath47 where @xmath43 is independent of the initial damage .
the transverse distance @xmath48 does not indicate any transition at all .
the value of @xmath49 depends on the value of the anisotropy @xmath6 and approaches zero in both the @xmath9 and @xmath50 limits . in figure 4
, we show the phase diagrams obtained using this damage spreading approach for both lattices .
both phase diagrams are in remarkably good agreement with those obtained from equilibrium studies @xcite within errorbars . in summary
, we have shown that the technique of damage spreading can be applied successfully to models where analytical solutions are lacking and where the low temperature phase displays novel behaviour associated with geometrical frustration .
we can clearly identify the damage transitions with those obtained previously using conventional monte carlo simulations . both defect - mediated and glass - like transitions are detected by the method .
further work is needed to extract more detailed information about the critical properties using this approach . + * acknowledgments * the authors would like to acknowledge support from the natural sciences and research council of canada and the high performance computing facility at the university of manitoba . |
many chemical and biological systems exhibit excitability . in small ( zero - dimensional ) geometry
they show threshold behavior , i.e. small perturbations immediately decay , whereas sufficiently large perturbations decay only after a large excursion .
one dimensional ( 1d ) excitable media support travelling pulses , or rather , periodic wave trains ranging in wavelength @xmath0 from the localized limit @xmath1 to a minimal value @xmath2 .
pulses and wavetrains are best - known from nerve propagation along axons . in 2d one typically observes spiral waves .
spirals have been observed for example in the auto - catalytic belousov - zhabotinsky reaction @xcite , in the aggregation of the slime mold dictyostelium discoideum @xcite and in cardiac tissue @xcite . for certain system parameters
the propagation of pulses , wave trains or the development of spiral waves may fail ( see for example @xcite ) .
the analytical tools employed to describe these phenomena range from kinematic theory @xcite , asymptotic perturbation theory @xcite to dynamical systems approaches @xcite .
however , no theory exists which describes propagation failure using only equation parameters , and which reproduces the behavior close to the bifurcation point .
for example , asymptotic perturbation theory fails to describe the square - root scaling behavior of the amplitude and the pulse velocity with respect to the bifurcation parameter at the bifurcation point . in kinematic theory results
are not given entirely in terms of the system parameters .
in this paper we develop a non - perturbative method to study propagation failure and compare the results with numerical simulations .
+ most theoretical investigations are based on coupled reaction - diffusion models .
we follow this tradition and investigate a two - component , two - dimensional excitable medium with an activator @xmath3 and a non - diffusive inhibitor @xmath4 described by @xmath5 this is a reparametrized version of a model introduced by barkley @xcite .
we expect our method to be independent of the particular model used .
note that the diffusion constant @xmath6 is not a relevant parameter as it can be scaled out by rescaling the length .
this model incorporates the ingredients of an excitable system in a compact and lucid way .
thus , for @xmath7 the rest state @xmath8 is linearly stable with decay rates @xmath9 along the activator direction and @xmath10 along the inhibitor direction .
perturbing @xmath3 above the threshold @xmath11 ( in 0d ) will lead to growth of @xmath3 . in the absence of @xmath4
the activator would saturate at @xmath12 leading to a bistable system . a positive inhibitor growth factor @xmath13 and @xmath14 forces the activator to decay back to @xmath15 .
finally also the inhibitor with the refractory time constant @xmath16 will decay back to @xmath17 .
for @xmath18 the system is in zero - dimensional systems no longer excitable but instead bistable .
this relaxation behaviour in the 0d system for super - threshold perturbations gives rise to pulse solutions in the 1d ( and 2d ) case .
the relaxation mechanism mediated by the inhibitor forces the pulse solution to decay in its back .
hence we observe pulses in excitable media and not fronts .
the barkley model is a variant of the class of 2-component fitzhugh - nagumo models . in the traditional fitzhugh - nagumo models
@xmath19 in eq.([barkley ] ) is replaced by @xmath20 .
thus the nullclines @xmath21 , which in the traditional fitzhugh - nagumo models are cubic polynomials , are replaced by straight lines in the barkley model .
most of the qualitative behavior in the relevant parameter ranges is unchanged by this . whereas the barkley model is computationally more efficient and also analytically better tractable , the traditional fitzhugh - nagumo models display a feature which makes them more realistic for the discription of excitable media in biology : the activator experiences an undershoot below its equilibrium value and slow decay in the tail region of a pulse .
we expect that for the phenomena discussed in this paper the difference only leads to quantitative changes ( indeed , we have done some tests to verify this assertion ) . in order to study pulse propagation in 1d
it is useful to first consider the case of constant @xmath4 .
the resulting bistable model is exactly solvable @xcite and the pulse velocity is @xmath22 $ ] .
hence , excitability requires that @xmath11 is below the stall value @xmath23 .
the quantity @xmath24 characterizes the strength of excitability and @xmath25 coincides with the solitary pulse velocity for @xmath26 .
clearly , for @xmath27 and not too large @xmath28 , pulse propagation fails for @xmath13 larger than some @xmath29 .
the critical growth factor @xmath29 describes the onset of a saddle - node bifurcation @xcite .
the saddle node can be intuitively understood when we consider the activator pulse as a heat source , not unlike a fire - front in a bushfire . due to the inhibitor
the width of the pulse decreases with increasing @xmath13 .
hence , the heat contained within the pulse decreases . at a critical width , or a critical @xmath13 , the heat contained within
the pulse is too small to ignite / excite the medium in front of the pulse . for periodic wavetrains
the saddle node depends on the wavelength .
the pulses run into the inhibitor field of their respective preceding pulse .
hence , propagation failure for periodic wave trains is controlled by the decay of the inhibitor and propagation is only possible when the inter - pulse distance @xmath0 becomes larger than a critical wavelength @xmath2 .
note that @xmath2 diverges for @xmath30 when the decay rate of the inhibitor @xmath31 vanishes .
+ in previous analytic works on one - dimensional pulses the limit of small @xmath13 was considered @xcite .
then the solution of the activator @xmath3 is well separated in two flat plateau regions with @xmath32 and @xmath15 which are separated by a steep narrow front .
this approach does not capture the square - root behavior of @xmath33 close to the saddle - node bifurcation point .
close to the bifurcation point the solution resembles rather a bell - shaped pulse than a plateau . in this paper
we will be concerned with the behavior near the bifurcation at @xmath29 and make explicit use of the observed bell - shape form of the solution .
this allows us to describe the scaling behavior close to the bifurcation in terms of the equation parameters . in 2 dimensions spiral waves
are observed .
spirals can be created in excitable media from a finger , i.e. a 1d pulse which is extended in the second dimension and has one free end .
fingers may be created due to inhomogeneities in the excitability of the system @xcite . at long times ,
the free end will either sprout or retract depending on the growing velocity @xmath34 being positive or negative . if @xmath35 , the tip of the finger will sprout into the fresh medium , and in particular it sprouts and curves backwards .
this causes a non - vanishing curvature at the tip of the finger . due to the increased curvature
the finger tip is slower than the flat part of the front further away from the tip .
thus , the extending part far from the tip will curl up .
this leads to the formation of a spiral with the free end at its core .
the criterion @xmath35 is therefore a necessary criterion for spiral formation .
the transition to retraction always occurs before the 1d propagation failure .
it is harder to tackle analytically .
general and universal dynamical system approaches exploiting only the euclidean symmetry of excitable media describe the transition as a drift bifurcation .
these model independent theories give an explanation for the divergence of the core radius at the bifurcation point and also explain why a finger at the bifurcation point is translating with finite speed , i.e. that the transition to retracting fingers always occurs before the 1d propagation failure @xcite .
unfortunately , it has the unphysical result that at the bifurcation point @xmath36 a spiral changes its sense of rotation @xcite . whereas these theories treat the spiral as a global solution of the underlying equations and see the transition to spiral waves as a pitchfork or drift bifurcation @xcite
, we take a local approach and describe the transition not as a bifurcation but as a quantitative change in the velocity of the finger , analogous to a maxwell point in a first - oder phase transition .
asymptotic techniques in the limit @xmath37 have also been performed for this problem @xcite and produced an analytical expression for the onset of retraction for small @xmath13 . moreover ,
the authors were able to go one step further , in a detailed and sophisticated asymptotic analysis , and described the onset of meandering . in this paper
we go beyond the restriction of small @xmath13 and propose a more intuitive approach to the problem of the growing velocity which nevertheless gives good agreement with the numerics .
+ in the following we assume given excitability parameters @xmath28 , @xmath11 and take @xmath13 as the bifurcation parameter .
we will be looking for solutions that move with velocity @xmath38 in the @xmath39 direction and may grow ( or retract ) with velocity @xmath34 in the @xmath40 direction .
thus we rewrite eqs .
( [ barkley ] ) in a frame moving with velocity @xmath41 as @xmath42
we first look for pulse and wavetrain solutions that do not depend on @xmath40 .
the reason for the failure to describe the pulse properties at @xmath29 within the framework of asymptotics employing the smallness of @xmath13 is due to the fact that at @xmath29 the pulse shape for @xmath3 can not be separated into a steep narrow front and a flat plateau .
hence , asymptotic techniques such as inner and outer expansions are bound to fail .
instead the pulse has the shape of a rather symmetric bell - shaped function ( see fig .
1 ) . in the following
we make explicit use of the shape of the pulse close to the critical point and parametrize the pulse appropriately ; a method reminiscent of the method of collective coordinates in the studies of solitary waves @xcite .
we choose @xmath3 of the general form @xmath43 where @xmath44 is chosen as a symmetric , bell - shaped function , for example a gaussian , of unit width and height .
hence , we restrict the solutions to a subspace of bell - shaped functions @xmath44 which is parametrized by the amplitude @xmath45 and the inverse pulse width @xmath46 .
the aim of our method is to determine the so far undetermined parameters .
this is done by minimizing the error made by the restriction to the subspace defined by ( [ ansatz ] ) .
we avoid further uncontrolled approximations and solve for the inhibitor field @xmath4 explicitly @xmath47 where @xmath48 @xmath49 is determined via the periodic boundary condition @xmath50 and is @xmath51 we assume that the width @xmath52 is small compared to the distance between two consecutive pulses @xmath0 . in the temporal domain
this means that the time scale for the decay of the inhibitor is much longer than the activator pulse width .
this assures that the activator field of consecutive pulses is well separated and only the inhibitor field overlaps , and that the interaction between pulses is mediated only through the inhibitor .
otherwise we would have to choose periodic functions @xmath44 .
however , in this case we can now replace the limits of integration @xmath53 by @xmath54 . for the isolated pulse ,
i.e. when @xmath55 , @xmath49 vanishes .
we now determine the parameters @xmath45 and @xmath46 by projecting eq .
( [ barkleyt1 ] ) onto the tangent space of the restricted subspace defined by ( [ ansatz ] ) .
the tangent space is spanned by @xmath56 and @xmath57 .
this assures that the error made by restricting the solution space to the test functions is minimized . to achieve this , we multiply eq .
( [ barkleyt1 ] ) with the basis functions of the tangent - space @xmath58 and @xmath59 , integrate over the @xmath60-domain and require the projection to vanish , i.e. @xmath61 where the brackets indicate integration over the whole @xmath60-domain .
the terms proportional to the velocity @xmath38 vanish .
the resulting equations can be combined to give , at fixed @xmath62 and @xmath28 , a quadratic equation for @xmath45 with two solutions @xmath63 which describe the stable and unstable branch , respectively ( see below for a subtle issue at the saddle node ) .
we obtain @xmath64 where @xmath65 the corresponding inverse width parameters @xmath66 for the stable and unstable branch are given by @xmath67 \ ; .\end{aligned}\ ] ] the velocity @xmath38 can now be determined in the standard way by multiplying eq .
( [ barkleyt1 ] ) by @xmath68 and integrating over @xmath39 .
we obtain @xmath69\ ; .\end{aligned}\ ] ] finally , we can determine @xmath13 from @xmath70 . multiplying ( [ c0 ] ) by @xmath46 one sees that @xmath13 can be computed without calculating @xmath46 and @xmath38 .
we will use a gaussian @xmath71 as an ansatz function ( [ ansatz ] ) . note that other symmetric bell - shaped functions such as a _
sech_-function , are possible , too .
then one has @xmath72 and @xmath73 .
the parameters of the test function @xmath45 and @xmath46 , and the front velocity @xmath38 can now be determined numerically using ( [ u0],[w],[c0 ] ) .
simplifications are possible in the useful limit @xmath74 . in this limit
the ( temporal ) inverse pulse width @xmath75 is small compared to the inhibitor decay time @xmath76 ( see definition ( [ theta ] ) .
then in eqs .
( [ coeffs ] ) and ( [ c0 ] ) the terms proportional to @xmath28 can be omitted , and for the calculation of @xmath77 one can omit the exponentials in ( [ vper],[vstar ] ) leading to @xmath78 note that @xmath79 corresponds to @xmath80 ( see ( [ vper ] ) ) for small @xmath28 .
now @xmath81 can be replaced by the constant @xmath82 in @xmath83 ( the rest is an odd function ) leading to @xmath84 and from ( [ c0 ] ) we infer @xmath85 we consider now isolated pulses for which the wavelength @xmath0 is large compared to the decay length of the inhibitor @xmath86 . in figures 1 and 2 we show a comparison of our results eqs .
( [ u0],[w],[c0 ] ) for @xmath87 and @xmath38 , with a direct numerical simulation of eqs .
( [ barkley ] ) .
the pulse shape , the critical bifurcation point @xmath88 and the behavior near the saddle - node bifurcation of the amplitude @xmath89 and of the velocity @xmath33 are very well recovered .
note the square - root behavior near the saddle node .
let us discuss some systematic features of the isolated pulses at criticality depending on the equation parameters @xmath28 and @xmath11 , which can be extracted from our approach .
solutions @xmath45 of ( [ u0 ] ) exist when the discriminant @xmath90 is positive .
the amplitude at the saddle node @xmath91 is determined by the condition @xmath92 .
the corresponding bifurcation parameter @xmath29 can then subsequently be determined using eqs .
( [ w ] ) and ( [ c0 ] ) .
note that the saddle node , which occurs at the maximal bifurcation parameter @xmath29 , is not given by the relation @xmath93 since the condition of maximal @xmath13 is different from that of maximal @xmath62 ( see for example ( [ epsilon - a-0 ] ) ) . in fig .
3 we show @xmath29 and the corresponding amplitude at the saddle node @xmath91 as a function of @xmath11 for @xmath94 ( continuous line ) for the isolated pulse with @xmath95 .
the points are results from a full solution of the 1-dimensional version of the odes ( [ barkleyt1],[barkleyt2 ] ) . in our approximation
the limit of excitability ( that is the maximum of @xmath11 for which @xmath96 ) is @xmath97 , which is close to the exact stall - value @xmath98 .
note that @xmath99 is independent of @xmath28 , as it should .
note also that pulse propagation in the neutrally stable case @xmath100 is possible .
moreover pulses are supported even for negative @xmath11 which we have checked numerically .
we mention that for this calculation it is crucial to take into account the difference between maximizing @xmath13 and maximizing @xmath62 . in fig .
4 we show @xmath29 and @xmath91 as functions of @xmath28 for @xmath101 .
we see that our approximation ( continuous line ) reproduces all features of the odes ( points ) .
fig.4 reveals that around @xmath102 the saddle - node bifurcation ceases to exist .
the velocity @xmath38 approaches zero for these values .
this correlates well with the fact that in the full system pulses become delocalized around @xmath103 , i.e. front and tail of a pulse separate creating a domain with @xmath104 , which represents a locally stable stationary state of the system . as a matter of fact
, the inverse pulse width @xmath46 diverges here for the test function approach .
+ even if a given set of equation parameters allows for propagation of a single solitary pulse , the system may not necessarily support a wave train consisting of several of such pulses . as a matter of fact ,
if the distance @xmath0 between two consecutive pulses of the train becomes too small , the pulses run into the refractory tail of the preceding pulse and consecutively decay .
the critical wavelength @xmath2 is a lower bound for the wavelength for the existence of periodic wave trains . on
can also think of keeping @xmath0 fixed and , as before , vary @xmath13 .
then the saddle node @xmath105 is a monotonically increasing function .
one can calculate @xmath2 ( or @xmath29 ) essentially as before , except for the complication that due to @xmath82 ( [ v - a=0 ] ) , the expressions for @xmath87 and @xmath38 , ( [ f_0]),([w ] ) and ( [ c_0 ] ) , can not be cast in closed form depending only on @xmath62 .
this is true even in the limit of small @xmath28 .
however , given the equation parameters @xmath13,@xmath28 and @xmath11 one may obtain @xmath2 numerically as a consistency relation requiring that at each @xmath0 there exists a @xmath106 so that the value for @xmath38 obtained by solving eqs .
( [ u0],[w ] ) and using the relation @xmath107 , matches the value for @xmath38 obtained by solving eq .
( [ c0 ] ) .
we obtain very good agreement between our test function approach and the numerically obtained values for the critical wavelength @xmath2 . in fig .
5 we show a comparison of the values obtained by integrating the full system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) with the calculation of the test function approach as described above .
the critical wavelength @xmath2 diverges when @xmath13 approaches @xmath29 where the saddle node of the localized pulse ( i.e. @xmath108 ) causes propagation failure of isolated pulses ( see section 2.1 and fig .
2 ) . in the remainder of this section
we will discuss the limit of large values of @xmath0 , i.e. small perturbations to the saddle node @xmath29 of the isolated pulse .
this causes small shifts of the critical @xmath13 , the amplitude @xmath45 and velocity @xmath38 when compared to their respective values in the case of isolated pulses and @xmath108 . for simplicity
, we restrict the calculation to the particular limit of small @xmath28 .
we write @xmath109 and expand in terms of small @xmath110 .
the correction to the isolated pulse @xmath110 is connected to the exponential tail of the inhibitor from the previous pulse .
it thus captures the interaction between pulses mediated by the inhibitor . at leading order the shift with respect to the case @xmath108 of @xmath29 , @xmath45 and @xmath38 at the saddle node
must be proportional to r. for small @xmath28 one finds at leading order @xmath111 to see this note that the parameters @xmath112 are entirely given as a function of @xmath113 via eqs .
( [ u0]),([w ] ) and ( [ c0 ] ) . hence ,
if we multiply eq .
( [ epsilon - a-0 ] ) by @xmath114 , we have four equations to determine the parameters @xmath112 and @xmath115 .
we expand @xmath116 and @xmath29 with respect to @xmath110 around the solutions of the isolated pulse .
in particular , we write @xmath117 .
the first - order correction @xmath118 can be entirely determined using @xmath119 where we deliberately ignored the subscripts to denote that @xmath120 and @xmath121 need to be expanded in @xmath110 .
the saddle - node condition @xmath122 implies that the derivative of the right - hand side of ( [ epsilon - a-0-r ] ) with respect to @xmath120 vanishes at @xmath123 , and we obtain the first - order correction @xmath118 leading to ( [ stoer - eps ] ) . this result
is confirmed by our numerical simulations .
we have determined the shift of @xmath29 due to finite wave length @xmath0 for @xmath30 and @xmath101 numerically from the odes and find @xmath124 .
this shows the accuracy of our test function approach for the saddle node shift . the leading - order approximation for the saddle node behavior is good down to about @xmath125 .
for @xmath126 we find numerically @xmath127 . in sec .
[ summary ] we will touch on some questions of stability of the wave train solutions near the saddle node .
in this section we develop a @xmath128-dimensional extension of the test function approach .
we study isolated finger solutions , i.e. solutions which in the @xmath129 direction go over into an isolated pulse ( moving with velocity @xmath38 in the @xmath39 direction ) and rapidly decay to zero in the @xmath130 direction . in the @xmath40
coordinate they may be regarded as fronts , which grow or retract with velocity @xmath34 ( see fig . 6 for a retracting case ) .
we derive an explicit formula for the growing velocity @xmath34 .
we now investigate the full @xmath128-dimensional system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) in a frame moving with velocity @xmath41 .
we introduce a product ansatz for the activator field @xmath131 with test function @xmath44 .
this approximation neglects possible curvature at the tip .
again , we avoid any further uncontrolled approximation and solve for @xmath132 and @xmath133 in a systematic way . note that @xmath132 replaces the constant @xmath45 in ( [ ansatz ] ) . the solution of eq .
( [ barkleyt2 ] ) with the ansatz ( [ prod_u ] ) can be written explicitly as @xmath134 where @xmath135 .
further calculations with this full expression appear prohibitive . since we are mainly interested in the reversal point of @xmath34 we resort to an ( asymptotic ) expansion in powers of @xmath34 , restricting ourselves here to the first two terms . from ( [ fullsol ] ) one obtains @xmath136 where @xmath137 @xmath138 coincides with @xmath139 of eq .
( [ vper ] ) with @xmath80 and @xmath108 ( or , for small @xmath28 , with eq .
( [ v - a=0 ] ) ) and @xmath140 note that the first - order correction of the inhibitor ( [ v1 ] ) can also be obtained by inserting the ansatz ( [ prod_v ] ) into ( [ barkleyt2 ] ) and solving for successive orders of @xmath34 . repeating the procedure that led to eq .
( [ projectionsu ] ) for @xmath45 , i.e. multiplying eq .
( [ barkleyt1 ] ) with @xmath44 and integrating over @xmath60 , we obtain using ( [ g0 ] ) @xmath141 with @xmath142\rangle /\langle u^2\rangle . \end{aligned}\ ] ] note that for @xmath143 , where @xmath63 are the solutions of the quadratic equation ( [ u0 ] ) , we have @xmath144 .
we first neglect the higher - order correction of the inhibitor field @xmath145 .
then @xmath146 is a quadratic form in @xmath132 with zeros @xmath63 , and eq .
( [ cubic ] ) can be solved exactly with the ansatz @xmath147 which states that far away from the tip @xmath132 is constant and takes values @xmath148 or @xmath45 .
the constant of proportionality @xmath149 can be determined . using ( [ sneyd ] )
we obtain for the growing velocity @xmath150 the point @xmath151 is fixed by the condition @xmath152 .
the value for @xmath13 where @xmath151 which we denote by @xmath36 , matches very well the value obtained by numerically integrating the full @xmath128-dimensional system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) .
however , the behavior for nonzero growing velocities is not captured by ( [ cg0 ] ) .
in fact , the slope @xmath153 close to the reversal point is too small by an order of magnitude for the parameters used in fig .
1,2,5,6 when compared to the values obtained by the full @xmath128-dimensional simulation . to obtain the correct slope we need to take into account the correction @xmath154 of the inhibitor field .
including the first - order correction @xmath154 in eq .
( [ cubic ] ) we solve for @xmath34 analogously to the determination of @xmath38 in section 2 by multiplying eq .
( [ cubic ] ) by @xmath155 and subsequently integrating over @xmath40 .
to @xmath156 we obtain @xmath157 where @xmath158 as expected the higher - order corrections @xmath159 and @xmath160 do not change the value of @xmath36 , but change the slope of @xmath161 . in fig .
7 we show a comparison of the test function approach and of eq .
( [ cg ] ) with numerically obtained data .
the correspondence close to @xmath162 is striking . to obtain better agreement further away from @xmath162
one would have to include higher - order terms in the expression ( [ prod_v ] ) .
we have developed a non - perturbative method to study critical wave propagation of single pulses and periodic wavetrains in @xmath163 and @xmath128 dimensions .
the method is based on the observation that near the bifurcation point the pulse shape is close to a symmetric bell - shaped function . a test function approximation , optimizing the two free parameters of a bell - shaped function , i.e. amplitude and width , allows to calculate the wave speed of a critical and close - to - critical pulse .
we were able to study propagation failure of isolated pulses and wave trains , and moreover the test is capable of capturing more general features such as the transition from excitability to bistability .
we have also performed our test function method with a more general class of nonsymmetric test functions to calculate the pulse parameters and velocities .
it turns out that near the saddle node the asymmetry indeed becomes irrelevant .
we extended our method to two dimensional situations , and used it to study broken fronts .
depending on parameters these fingers may either retract or sprout and start spiraling .
we studied the growing velocity of a critical finger whose growing velocity is close to zero .
our test function method combined with a separation ansatz for the two - dimensional finger tip yields analytical expressions for the growing velocity which depend only on the equation parameters and which are for large parameter ranges in good agreement with the numerically obtained values .
we elaborated on the importance of the inhibitor field for the growing velocity .
let us finally mention an interesting observation for one - dimensional wave trains .
below some ( rather large ) pulse separation ( wavelength ) the wave train looses stability not via a saddle node but via a hopf bifurcation .
this may be seen in numerical simulations of the full system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) ( and also of other excitable media equations such as the fitzhugh - nagumo equation ) and in analytical approximations , either based on a time - dependent generalization of the test function approach or on a systematic reduction scheme valid near the saddle node . at the codimension @xmath128 point where the saddle node of the wave train coincides with the hopf - bifurcation the hopf frequency becomes infinite .
this time - dependent phenomenon can of course not be captured within the current stationary algebraic framework .
this phenomenon and its implications will be published separately . for completeness
we have included a time - dependent extension of the test function approach which allows for hopf bifurcation in an appendix .
this seems to be related to the phenomenon of alternans @xcite which has been studied in a different parameter regime far away from the saddle node @xcite .
another interesting problem we plan to address is the selection of the wavelength ( or pitch ) of spirals . in simulations we found that near @xmath36 , where growth of fingers becomes small , the selected wavelength diverges .
kinematic theory @xcite addresses this problem .
kinematic theory provides , in principle , a relationship between the rotation frequency of a spiral and its core radius .
however , it fails to provide expressions for either of the two which only depend on the equation parameters .
a connection of our theory with kinematic theory is planned for further research to fill this gap .
+ we would like to thank the inln where parts of this work was done for its hospitality . g.g . would like to thank valentin krinsky for introducing him to this research .
we would like to thank dwight barkley , gianne derks , alain karma , jim keener , ian melbourne , sasha panfilov , alain pumir , sebastian reich and mark roberts for valuable discussions .
was partly supported by a european commission grant , contract number hprn - ct-2000 - 00113 , for the research training network _
mechanics and symmetry in europe _ ( masie ) and by an australian academy of science travel grant .
part of the simulations were done with the xdim interactive simulation package developed by p. coullet and m. monticelli , who l. k. thanks for their support .
in this appendix we present the time - dependent calculation for our test function approach . for simplicity we restrict ourselves to small values of @xmath28 .
this means that the width of the activator pulse @xmath3 is small compared to the width of the decaying inhibitor field @xmath4 .
we include now temporal dependency of the pulse variables @xmath164 and @xmath46 .
we study pulse trains and note that the localized pulse can be obtained in the limit @xmath1 .
we hence choose @xmath3 of the form @xmath165 the sum extends over all @xmath166 pulses .
we defined @xmath167 and @xmath44 is chosen as a gaussian @xmath168 as above .
we allow for individual dynamics of the pulses characterized by the amplitude @xmath169 , the inverse width @xmath170 , and the position @xmath171 . for a stationary wavetrain
the @xmath172 and @xmath173 will be constant and all equal , and @xmath174 , where @xmath121 is the velocity .
we will restrict ourselves to the situation where the perturbations around such a state are small and slowly varying .
we insert the ansatz ( [ ansatzt ] ) into the general expression @xmath175 obtained from the second eq .
( [ barkley ] ) .
assuming that the ( temporal ) inverse pulse width @xmath75 is small compared to the inhibitor decay time @xmath76 , we obtain the following expression valid in the vicinity of the @xmath176 pulse , which is assumed to pass the origin at @xmath177 @xmath178 + \epsilon \langle u\rangle \sum_{l=1}^{\infty } e^{-l \epsilon a l / c_l } \frac{f_l}{w_l c_l } + s_0 \int_0^\eta d \eta ' { \eta ' u(\eta ' ) } \end{aligned}\ ] ] where the brackets indicate integration over the whole @xmath60-domain , and @xmath179)/(w_0 f_0 ) .
\end{aligned}\ ] ] in order to determine @xmath180 , and @xmath181 we again project eq .
( [ barkleyt1 ] ) onto @xmath44 , @xmath182 , and @xmath183 . after combining the first two equations appropriately one obtains @xmath184 where @xmath185 the third projection gives an algebraic relation @xmath186 these equations are written for the pulse @xmath176 .
they apply correspondingly to the other pulses .
first consider the case of a stationary , isolated pulse ( @xmath108 ) where @xmath187 and the sum in eq .
( [ v_p ] ) vanishes .
the resulting algebraic relations are easily solved , and correspond in the limit of @xmath188 to eqs .
( [ u0 ] ) , ( [ w ] ) , ( [ c0 ] ) . with reasonable initial conditions , simulation of the odes
( [ f_0 ] ) , ( [ w_0 ] ) , together with ( [ c_0 ] ) leads to the same result as described in section 2 .
the odes ( [ f_0],[w_0 ] ) relax to the stationary values obtained in section 2 .
+ this time - dependent test function approach allows to go beyond the stationary bifurcations discussed in section [ summary ] .
as mentioned in section 2 the saddle - node bifurcation related to propagation failure for well separated pulses transforms into a subcritical hopf bifurcation via a takens - bogdanov point when the pulse separation is reduced below a critical value @xmath2 .
this work will be published elsewhere .
the hopf - bifurcation can be captured within the ode - system ( [ f_0 ] ) , ( [ w_0 ] ) and ( [ c_0 ] ) , and some preliminary simulations have been done . of a steady front with @xmath189 at @xmath190 .
the points depict the solution obtained by numerically integrating eqs .
( [ barkley ] ) .
the continuous line is the theoretical curve obtained with the test function approach .
note that here @xmath13 is not even very close to @xmath191 . ] as a function of @xmath13 .
( b ) : velocity @xmath38 as a function of @xmath13 .
the parameters used were @xmath192 , @xmath126 and @xmath101 .
the crosses depict the numerically obtained values of integrating the full system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) , the lines depict the stable and unstable branch calculated with the test function approach , i.e. by using eqs .
( [ u0],[w],[c0 ] ) . ] .
( a ) : the critical @xmath91 versus @xmath11 .
( b ) : critical amplitude @xmath29 versus @xmath11 .
points correspond to simulations of the full @xmath163-dimensional version of ( [ barkleyt1],[barkleyt2 ] ) .
the continuous line corresponds to the test function approach . ] .
( a ) : the critical @xmath29 versus @xmath28 .
( b ) : critical amplitude @xmath91 versus @xmath28 .
points correspond to simulations of the full @xmath163-dimensional version of ( [ barkleyt1],[barkleyt2 ] ) .
the continuous line corresponds to the test function approach . ] as a function of @xmath13 .
crosses depict the values obtained by direct simulation of the full system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) .
squares depict the values obtained by the test function approach described in section 3 .
parameters are again @xmath192 , @xmath126 and @xmath101 . in ( b ) the same numerical results as in ( a ) are presented but here the data - points corresponding to the test - function approach are shifted along the @xmath13-axis such that the saddle nodes at @xmath193 ( see fig .
2 ) , coincide . ] as a function of @xmath13 .
the crosses depict the values obtained by direct numerical integration of eqs .
( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) .
the continuous line shows the theoretical curve ( [ cg ] ) using the test function approach of section 4.1 .
we have shifted the curve along the @xmath13-axis by the difference of the @xmath13-values @xmath194 for the saddle nodes obtained by the numerical simulations of the full system ( [ barkleyt1 ] ) and ( [ barkleyt2 ] ) , and the test function approach of section 2 . ] | we present a non - perturbative technique to study pulse dynamics in excitable media .
the method is used to study propagation failure in one - dimensional and two - dimensional excitable media . in one - dimensional media
we describe the behaviour of pulses and wave trains near the saddle node bifurcation , where propagation fails .
the generalization of our method to two dimensions captures the point where a broken front ( or finger ) starts to retract .
we obtain approximate expressions for the pulse shape , pulse velocity and scaling behavior .
the results are compared with numerical simulations and show good agreement .
* excitable media are often found in biological and chemical systems .
examples of excitable media include electrical waves in cardiac and nerval tissue @xcite , camp waves in slime mold aggregation @xcite and intracellular calcium waves @xcite .
excitable media support localized pulses and periodic wave trains . in 2 dimensions rotating vortices ( or spirals )
are possible @xcite .
the critical behavior of pulses , wave trains and spirals , i.e.propagation failure , is often associated with clinical situations .
the study of spiral waves is particularly important as they are believed to be responsible for pathological cardiac arrhythmias @xcite .
spiral waves may be created in the heart through inhomogeneities of the properties of the cardiac tissue .
+ we investigate critical behavior relating to these 3 wave types .
we develop a non - perturbative test function method which allows to study the bifurcation behavior of critical waves .
in particular , we study under what conditions a broken front will sprout and develop into a spiral wave or retract .
analytical formulas for the growing velocity of a broken front are given .
for wave trains we provide a time dependent extension which supports a hopf bifurcation which is also observed in numerical simulations of excitable media .
this seems to be related to alternans @xcite , which also are discussed in the context of cardiac electric pulse propagation .
the methods and results are general , and can be applied to other excitable media . * |
to introduce a model which innately and efficiently represents the salient properties of receptive fields of neurons in the v1 to mt motion processing stream . as one traces the neural pathway connecting the input layers of v1 to area mt ,
neurons become increasingly specialized for motion detection . in ascending order ,
attributes such as orientation selectivity , direction selectivity , speed tuning , increasing preferred speeds , component spatial frequency selectivity , and pattern ( plaid ) spatial frequency selectivity are sequentially acquired .
furthermore , the temporal frequency filtering properties of the neuronal population changes in a particular way along this pathway .
specifically , the proportion of bandpass temporal frequency filtering neurons to lowpass temporal frequency filtering neurons increases .
existing receptive field models do not represent this fundamental emergent property .
our specific aim is to introduce and describe a model for the receptive fields of v1 to mt neurons which innately and efficiently represents the aforementioned properties .
the detection of moving objects in our environment is both critical to our experience of the world and to our very survival .
imagine a hunter in rapid pursuit of a small agile animal .
both predator and prey detect abrupt changes in velocity occurring over an interval of less than a tenth of a second .
the precision of our visual motion detection system is undoubtedly striking . the middle temporal ( mt ) area of the mammalian brain plays a key role in the visual processing of motion .
the importance of area mt in motion processing is supported by a growing wealth of electrophysiological evidence which was initiated by dubner and zeki s 1971 report @xcite . for purposes of motion processing , the main input signals into mt originate in specialized cells of the primary visual cortex ( v1 ) .
kuffler s early studies of retinal ganglion cell response properties led to similar studies of simple cortical cells by hubel and wiesel .
together , their work generated great interest in neuronal receptive fields @xcite .
the functional forms of these receptive fields are at once beautifully simple yet enormously complex .
hence mathematical models have been used in step with electrophysiological studies to advance our understanding of their properties .
initially , focus was predominantly on their spatial structure @xcite .
now , however , their temporal structure is increasingly studied in tandem . in particular , for most neurons in the v1 to mt processing stream , it is now appreciated that spatial and temporal features can not be studied separately .
they are spatiotemporally inseparable entities .
in particular , motion is encoded by orientation in the spectral domain , and spatiotemporally - oriented filters are therefore motion detectors @xcite .
therefore at first glance it may seem an easy matter to mathematically model motion detecting neurons .
the challenge , however , is to develop physiologically sound receptive field models which reflect the hierarchical structure of the motion processing stream .
various models do exist which are spatiotemporally - oriented filters , and are therefore motion detectors from a mathematical standpoint @xcite .
however , they fail to represent one of the most salient characterizing attributes of the motion processing stream : the lowpass to bandpass distribution of temporal frequency filtering properties along the v1 to mt specialization hierarchy . this feature is discussed further in the next subsection .
in addition to the above temporal frequency filter property profile , the neurons of the v1 to mt motion processing stream have certain characteristics which set them apart and prescribe bounds on models of their spatiotemporal receptive field structure .
namely , they are tuned to specific orientations , directions , and speeds @xcite ; and they are more likely than other v1 cells to exhibit `` end - stopping '' phenomena @xcite . of the v1 cells ,
the subclass with direct synaptic projections from v1 to mt are the most specialized towards motion and are thought to be the main channels of motion substrates incident into mt neurons .
for instance , churchland et al showed that v1 and mt neurons lose direction selectivity for similar values of spatial disparity , suggesting v1 is mt s source of direction selectivity @xcite . additionally , these v1 to mt neurons cluster motion substrate properties such as strong direction selectivity and tuning to high speeds .
for example , foster et al found that highly direction selective neurons in the macaque v1 and v2 were more likely to be tuned to higher temporal frequencies and lower spatial frequencies , i.e. higher speeds @xcite . in other words , neurons that are highly direction selective are more likely to also be tuned to higher speeds .
such neurons are higher up in the v1 to mt hierarchy .
similarly , mclean and palmer studied direction selectivity in cat striate cortex , and found that space - time inseparable cells were direction selective while space - time separable cells were not direction selective @xcite . again reflecting the hierarchy of the motion processing stream .
the distribution of some distinct anatomical and histological features have been shown to correlate with the motion - specialization hierarchy .
for instance , shipp and zeki s retrograde tracer studies showed that the majority of direct v1 to mt projecting neurons originate within layer 4b @xcite .
based on their results , they suggested that layer 4b contained a functionally subspecialized anatomically segregated group whose members each have direct synaptic connections with mt cells .
the exact properties of the v1 to mt projectors are likely to be intermediate between those of cells in v1 layer 4c@xmath0 and mt neurons .
this is because input into mt appears to be largely constituted of magnocellular predominant streams from 4c@xmath0 , while parvocellular predominant streams may play a much less role @xcite .
progressive specialization along the v1 to mt motion processing stream is also observable in the morphological characteristics of the neurons .
for instance , distinguishing attributes of the v1-mt or v2-mt projectors such as size , arborization patterns , terminal bouton morphology , and distribution have been observed @xcite .
sincich and horton demonstrated that layer 4b neurons projecting to area mt were generally larger than those projecting to layer v2 .
overall , neurons in the v1 to mt motion processing stream are highly and progressively specialized .
hence description of their receptive fields requires adequately specialized models which reflect not only their individual attributes , but also the emergent hierarchical properties of the network .
next we discuss one of the most fundamental of such emergent network properties : the particular distribution of temporal frequency filtering types along the stream @xcite .
hawken et al found that direction - selective cells were mostly bandpass temporal frequency filters , while cells which were not direction - selective were equally distributed into bandpass and lowpass temporal frequency filtering types @xcite .
foster et al found a similar phenomenon in macaque v1 and v2 neurons .
v1 neurons were more likely to be lowpass temporal frequency filters , while their more specialized downstream heirs , v2 neurons , were more likely to be bandpass temporal frequency filters @xcite .
less specialized neurons located anatomically upstream ( lower down in the hierarchy ) are more likely to have lowpass temporal frequency filter characteristics , while more specialized neurons located anatomically downstream ( higher up in the hierarchy ) are more likely to display bandpass temporal frequency filter characteristics .
for example , lgn cells are at best only weakly tuned to direction and orientation ( @xcite , @xcite ) and are hence equally distributed into lowpass and bandpass categories .
layer 4b v1 cells and mt cells on the other hand , are almost all direction selective @xcite and hence are mostly bandpass temporal frequency filters .
we firmly believe this classification is not arbitrary , but instead is a direct manifestation of the particular spatiotemporal structure of the v1 to mt motion processing stream .
consequently , the receptive field model must reflect this increased tendency for bandpass - ness with ascension up the hierarchy .
in other words , the representation scheme must be one that is inherently more likely to deliver bandpass - ness to a more specialized cell ( such as in layer 4b or mt ) and lowpass - ness to a less specialized cell ( such as in layer 4a , 4c@xmath0 , or 4c@xmath1 ) .
however , none of the existing receptive field models reflect this underlying spatiotemporal structure . in contrast , as will be seen in section ( [ sec : gabor - einstein wavelet ] ) below , the gabor - einstein wavelet s wave carrier is a sinc function which directly confers the aforementioned salient property .
individual gabor - einstein basis elements are lowpass temporal frequency filters ; and bandpass temporal frequency filters can only be obtained by combinations of basis elements .
a study by deangelis and colleagues also corroborates the above .
they found that temporally monophasic v1 cells in the cat were almost always low pass temporal frequency filters , while temporally biphasic or multiphasic v1 cells were almost always bandpass temporal frequency filters @xcite .
the explanation for this finding is inherent and explicit in the gabor - einstein wavelet basis , where bandpass temporal character necessarily results from biphasic or multiphasic combination .
all monophasic elements on the other hand , are lowpass temporal frequency filters .
however , we will see that according to the model , the converse is not true : i.e. not all lowpass temporal frequency filters are monophasic , and not all multiphasic combinations yield bandpass temporal frequency filters .
perceptual and electrophysiological studies reveal intriguing similarities between how space and time are mixed in the visual cortex and how they are mixed in the special theory of relativity ( str ) .
for instance , within certain limits , binocular neurons can not distinguish a temporal delay from a spatial difference .
as a result , inducing a monocular time delay by placing a neutral density filter over one eye but not the other , causes a pendulum swinging in a 2d planar space to be perceived as swinging in 3d depth space .
this is known as the pulfrich effect .
also , most neurons in the v1 to mt stream are maximally excited only by stimuli moving at that neuron s preferred speed .
they are _ speed tuned_. in this subsection , we briefly review and summarize the special theory of relativity .
we identify the stroboscopic pulfrich effect and speed tuning as cortical analogues of str s joint encoding of space and time . and finally , we explain how str is used in ( and motivates ) the design of the gabor - einstein wavelet . in 1905 , albert einstein proposed the special theory of relativity @xcite .
he was motivated by one thing : a firm belief that maxwell s equations of electromagnetism are _ laws of nature_. maxwell s equations are a classical description of light @xcite . according to maxwell
, light is the propagation of electric and magnetic waves interweaving in a particular way perpendicular to each other .
the speed of propagation in a vacuum is the constant @xmath2 m / s .
the other governing constraint on str is the _
relativity principle_. it has been around and generally accepted since the time of galileo , and it states that the laws of nature are true and the same for all non - accelerating observers .
einstein believed maxwell s equations to be laws of nature , and therefore to satisfy the relativity principle .
in other words , einstein believed the speed of light must be the same to all non - accelerating observers regardless of their velocity relative to each other . for the speed of light to be fixed , something(s ) had to yield and be unfixed . what had to yield were the components of the definition of speed , i.e. space and time .
space and time had to become functions of relative velocity and of each other , to ensure the speed of light remains constant to all non - accelerating observers .
the transformation rule of space and time coordinates between two observers moving with constant velocity relative to each other is called the lorentz equations .
it specifies how space and time are mixed in str .
more specifically , the lorentz equations are the transformation rule between the ( spacetime ) coordinates of two inertial reference frames . for two inertial reference frames , @xmath3 and @xmath4 , moving exclusively in the @xmath5 direction with constant velocity @xmath6 relative to each other ,
the lorentz transformation is given by , @xmath7 where ( @xmath8 ) are the coordinates of @xmath4 , ( @xmath9 ) are the coordinates of @xmath3 , @xmath10 is the speed of light , and @xmath11 is the lorentz factor and is given by , @xmath12 we demonstrate below that the stroboscopic pulfrich effect is analogous to str in one respect , and speed tuning of neurons is analogous to str in another respect . in the classical pulfrich effect , a monocular temporal delay results in a perception of depth in pendular swing @xcite .
it has a simple geometric explanation . by the time the temporally delayed eye `` sees '' the pendulum at a given retinal position ,
the pendulum has advanced to a further position in its trajectory , hence the visual cortex always receives images that are spatially offset between the eyes . the spatial disparity results in stereoscopic depth perception .
this form of pulfrich phenomena is termed _ classical _ to distinguish it from the stroboscopic pulfrich effect . in the stroboscopic pulfrich effect ,
the pendulum is not a continuously moving bulb , but instead is a strobe light which samples the trajectory of the classical pendulum in time and space @xcite .
the stroboscopic pulfrich effect can not be explained by the simple geometric illustration that explains the classical pulfrich effect .
this is because depth is still perceived in sequences involving a temporal delay in flash between eyes , but particularly lacking an interocular spatial disparity in the flash @xcite .
+ in stereoscopic depth perception , a planar ( x , y ) difference in retinal image position ( @xmath13 ) _ transforms into _ ( is perceived as ) a displacement along the @xmath14 direction . in the stroboscopic pulfrich effect , a temporal difference ( @xmath15 ) between eyes in retino - cortical transmission of image signal ,
results in a planar pendulum trajectory being perceived as ( transforming into ) depth . in str
, the transformation is between two inertial reference frames moving relative to each other . by analogy
, we assert that for the stroboscopic pulfrich effect , there is a transformation between the interocular retinal space ( @xmath16 ) and the perceptual depth space ( @xmath17 ) .
we identify the variables ( @xmath16 ) of the interocular space as the difference between eyes of the corresponding variables in the retino - cortical space i.e. @xmath18 is the difference between the left and right eye retinal positions of an image ; and @xmath19 is the difference between the right and left eyes in retino - cortical image transmission times .
the @xmath20 transformation may take the form @xmath21 , were @xmath22 is proportional to pendulum velocity and @xmath0 is a function of @xmath22 .
note the similarity between @xmath23 and the equation for @xmath24 in the lorentz transformation , @xmath25 .
this analogy suggests that similar mathematical structures underlie str and the joint encoding of space and time in the visual cortex .
the stroboscopic pulfrich effect is not by itself direct evidence of joint encoding of spatial and temporal disparities @xcite , however direct electrophysiological evidence does exist for such linkage in cortical encoding @xcite .
this mathematical similarity influences our design of the gabor - einstein wavelet in a particular way which we describe in section ( [ gabor_einstein_design ] ) below . but first , we discuss another neuronal phenomenon , speed tuning , which is analogous to str in a certain physical sense .
most motion processing stream neurons in v1 or area mt are speed tuned .
they have a preferred speed , which is the speed of moving stimuli to which they maximally respond .
other speeds also elicit a response , but with a lower rate of action potential . when presented moving sine wave grating stimuli , truly speed tuned neurons respond maximally to their preferred speed independent of the spatial or temporal frequency of the stimuli .
the velocity , @xmath26 , of a moving wave is , @xmath27 where @xmath28 is temporal frequency and @xmath29 is spatial frequency . in higher dimensions , @xmath30 where @xmath29 is the spatial frequency in the @xmath5 direction , @xmath6 is the spatial frequency in the @xmath31 direction , @xmath32 is the @xmath5 component of velocity , and @xmath33 is the @xmath31 component of velocity . from both equations above ,
it is clear that for a speed tuned neuron , a change in the spatial frequency of the stimuli would necessary require a complementary change in the temporal frequency .
hence a broad range of stimuli with widely varying spatial and/or temporal frequencies can maximally excite the neuron , so long as the above equations are satisfied for the neuron s preferred speed @xmath34 .
this is analogous to how the special theory of relativity is motivated by fixing the speed of light . in the case of str
the variables which necessarily change in complementary fashion are the space and time coordinates as described in the lorentz equations .
we constrained the gabor - einstein model by requiring the following properties : 1 .
the minimum possible number of parameters 2 .
relativistic - invariance of the wave carrier in addition to orientation in the spatiotemporal frequency domain , localization in the space , time , spatial frequency , and temporal frequency domains are the essential features of the v1 to mt neuron receptive field .
these are the most basic attributes of the receptive field . as such
, they correspond with the minimum number of model parameters needed .
specifically , speed tuning , the spatial and the temporal localization envelopes , the amplitude modulation parameter , and the spatial and temporal frequencies must all be represented .
if one adopts the notion of a gaussian spatial localizer @xcite , the minimum number of parameters add up to eight : four for the spatial envelope and four for the wave carrier .
the four necessary spatial envelope parameters are : amplitude factor , @xmath35 ; variance in the @xmath5 direction , @xmath36 ; variance in the @xmath31 direction , @xmath37 ; and spatial orientation of envelope , @xmath38 .
the four necessary wave carrier parameters are : temporal frequency , @xmath28 ; spatial frequency in the x direction , @xmath29 ; spatial frequency in the @xmath31 direction @xmath6 , and the phase , @xmath39 .
the orientation of the wave carrier @xmath40 is not an independent parameter .
we use it in computer simulations below only for convenience .
the essential feature that is not yet explicitly accounted for in the above parameter tally is the temporal envelope of the receptive field .
this can be implemented either through the gaussian envelope or through the wave carrier , and in either case may conceivably involve an additional parameter .
the next design criteria , _ relativistic - invariance of the wave carrier _ , resolves this ambiguity without adding another parameter .
given the analogies of spatiotemporal mixing in the visual cortex to spatiotemporal mixing in the special theory of relativity , it is reasonable to conjecture that their endpoints are within proximity of each other .
einstein eventually sought a lexicon in which physical laws are expressed the same way in any two inertial reference frames .
that lexicon is called relativistic - invariance ( or lorentz - invariance ) . by analogy
, we ultimately seek a lexicon in which physical laws can be expressed the same way in both the interocular and perceptual spaces .
lorentz - invariance of physical law is the specification ( or endpoint ) arising out of str .
hence , requiring lorentz - invariance of the receptive field s wave carrier is desirable . moreover
, this immediately resolves the above ambiguity regarding implementation of the temporal envelope profile .
the temporal envelope must be implemented via the wave carrier for relativistic invariance to hold . and
since the receptive field amplitude eventually falls , the sinc function emerges as a natural descriptor .
furthermore , the sinc function has the distinct advantage of not introducing any additional parameters .
the way space and time are mixed in the visual cortex bears resemblance to the way they are mixed in the special theory of relativity .
we demonstrated that the stroboscopic pulfrich effect and speed tuning are cortical analogues of the spacetime mixing mechanics of str . in str
, the mixture of space and time is described by the lorentz transform which relates the spacetime coordinates of inertial reference frames moving with constant velocity relative to each other . in the striate and extrastriate motion processing stream , the analogous transformation is between the 3d interocular space ( x , y planar inter - retinal space + inter - retinocortical time ) and the 3d perceptual space ( x , y , z 3d space ) .
though on the surface , str and cortical spacetime encoding appear to be unrelated processes , we conjecture and have partly shown a shared underlying mathematical structure .
this structure can be exploited to deepen our understanding of cortical encoding of motion , depth , and more fundamentally , the joint encoding of space and time in the visual cortex . here , we proceeded by requiring relativistic - invariance of the receptive field s wave carrier .
we simultaneously required that the model be constrained to the minimum possible number of parameters . under these constraints , the sinc function with energy - momentum relation as argument emerged as a natural descriptor of the receptive field s wave carrier .
furthermore , the gabor - einstein wavelet explains a number of salient physiological attributes of the v1 to mt spectrum .
chief amongst these being the distribution of bandpass to lowpass temporal frequency filter distribution profile ; which we postulate is a fundamental manifestation of the way space and time are mixed in the visual cortex .
the remainder of this paper is organized as follows : section ( [ sec : gabor - einstein wavelet ] ) introduces the gabor - einstein wavelet .
section ( [ sec : simulations ] ) presents computer simulations .
section ( [ sec : discussion ] ) is the discussion .
it includes the following subsections : subsection ( [ subsec : model_framework ] ) introduces a simple framework for classifying and naming the components and hierarchical levels of receptive field models .
subsection ( [ subsec : model_framework ] ) also identifies the gabor - einstein s place within this larger framework of receptive field modeling .
subsection ( [ subsec : related_work ] ) briefly discusses related work .
subsection ( [ subsec : higher_order ] ) discusses some phenomena which can not be explained by models such as the native gabor - einstein wavelet in isolation , i.e. models of single neuron receptive fields early in the visual pathway .
specifically , it discusses some higher order phenomena and non - linearities which necessarily arise from neuronal population and network interactions .
section ( [ sec : conclusion ] ) concludes the paper .
in this section , we present the gabor - einstein wavelet .
it has the following properties : * like the gabor function , it is a product of a gaussian envelope and a sinusoidal wave carrier . *
it differs from the gabor function in that its wave carrier is a relativistically - invariant sinc function whose argument is the energy - momentum relation . * its gaussian envelope contains only spatial arguments . i.e. its spatial envelope is not time - dependent . *
it has the minimum possible number of parameters . *
its fourier transform is the product of a mixed frequency gaussian and a temporal frequency step function . * like the gabor function
, it generates a quasi - orthogonal basis .
we define the gabor - einstein wavelet as follows , @xmath41 where the constant multiplier @xmath35 is amplitude ; @xmath36 and @xmath37 are the gaussian variances in the x and y directions ; @xmath42 and @xmath43 are the respective @xmath5 and @xmath31 coordinates of the gaussian center ; @xmath44 , @xmath45 , and @xmath46 are the frequencies of the wave carrier in the @xmath47 , @xmath5 , and @xmath31 directions respectively ; @xmath39 is the sinusoid phase ; and the _ sinc _ function is defined as , @xmath48 we have rotated our coordinates by an angle @xmath49 from a reference state @xmath50 to a state @xmath51 which we denote @xmath52 for notational simplicity . the transformation , @xmath53 is given by , @xmath54 we proceed here with the following instance of the gabor - einstein wavelet , @xmath55 where we have set @xmath44 equal to one .
we can always do so for one reference neuron by simply defining the unit of time as @xmath56 , the duration the neuron s frame cycle . as we will see , this value , @xmath56 , is the shortest frame duration to which the neuron can respond . in the above equation ,
we have also set @xmath57 , and @xmath58 .
the fourier transform is as follows , @xmath59 where @xmath60 is given by , @xmath61 and the _ sign _ function is defined as , @xmath62 the above fourier transform was obtained using mathematica symbolic software .
the fourier transform of the general case will likely be challenging to obtain analytically either by hand or symbolic software . hence , we anticipate numerical methods may have an important role to play .
the maximum magnitude of the above response function , equation ( [ gabor - einstein - fourier ] ) , is attained where the argument of the exponent is zero , i.e. where , @xmath63 the neuron s preferred spatial frequency , @xmath64 , is dependent on the temporal frequency , @xmath28 , of the stimulus , and is given by the above bivariate quadratic equation s solution , @xmath65 the location where the response maximum is attained is a parametrized curve in 3d frequency space .
it is given by , @xmath66 although the location in @xmath67 space where the neuron s maximum response is attained depends on @xmath28 , the magnitude of the maximum response is itself a constant .
it is given by , @xmath68 the half magnitude response is attained at values of @xmath67 satisfying , @xmath69 taking the natural logarithm of the above equation yields , @xmath70 defining @xmath71 and @xmath72 , equation ( [ half_max ] ) becomes , @xmath73 in the above form , one readily recognizes this as the ellipse centered at @xmath74 , whose principal axes radii are @xmath75 and @xmath76 in the @xmath5 and @xmath31 directions respectively .
the long axis in the frequency domain is the short axis in the spatial domain and vice versa . without loss of generality
, we can assume that prior to rotation of the spatial axes by angle @xmath49 , the long axis of the receptive field envelope is parallel to the x axis .
then the orientation of the on - off bands , i.e. planes of the spatial wave , are aligned parallel to the long axis of the envelope when the rotation angle , @xmath49 , is related to the polar angle @xmath77 by the relationship , @xmath78 on the other hand , the on - off bands are perpendicular to the long axis of the envelope when , @xmath79 in the case of equation ( [ parallel ] ) , the half magnitude frequency bandwidth is readily seen to equal @xmath80 , the length of the short axis , while in the case of equation ( [ perpendicular ] ) , the half magnitude frequency bandwidth equals @xmath81 , the length of the short axis .
the cases of skewed alignment take on values between @xmath80 and @xmath81 and are also computable from the geometry .
unlike the preferred spatial frequency , the half - magnitude frequency bandwidth is not dependent on the temporal frequency of the stimulus . in summary ,
the gabor - einstein wavelet spatiotemporal receptive field model predicts that the magnitude of a v1 to mt neuron s preferred spatial frequency is linearly dependent on the temporal frequency of the stimulus as shown in equation ( [ eqn : f_0 ] ) . in the next section
, we present some receptive field simulations using the gabor - einstein wavelet .
+ [ figur_gaussian ]
in this section , we present computer simulations demonstrating the gabor - einstein wavelet s properties .
we start with simulations which show the model s adherence to basic physiological form .
we then present simulations which demonstrate how the gabor - einstein model innately represents the temporal frequency filtering property distribution along the v1 to mt neuronal hierarchy .
the following notations are used in the figure captions : @xmath38 is the angle of rotation of the axes of the gaussian envelope as described in equations ( [ eqn : envelope_axes_rotate_1 ] ) and ( [ eqn : envelope_axes_rotate_2 ] ) .
@xmath82 is the angle of rotation of the axes of the wave carrier .
ctr is the 2-component vector consisting of the @xmath5 and @xmath31 coordinates of the gaussian envelope center respectively .
@xmath83 is the 2-component vector consisting of the gaussian variances of the envelope in the @xmath5 and @xmath31 directions respectively .
@xmath39 is phase of the wave carrier .
@xmath47 is time .
and @xmath84 is the 3-component vector consisting of the temporal , x - spatial , and y - spatial frequencies of the wave carrier respectively .
for succinctness , @xmath85 , for instance , is taken to be equivalent to @xmath86 .
the same short - hand notation applies to the other multi - component vectors .
figure ( [ fig : gaussians ] ) shows receptive field envelopes with varying aspect ratios and centers ; figure ( [ fig : wavecarriers ] ) shows the sinc function wave carriers at varying times @xmath47 and at one different orientation ; figures ( [ fig : gabor - einstein - center ] ) and ( [ fig : gabor - einstein - center-3d ] ) show gabor - einstein wavelets with varying envelope centers ; figure ( [ fig : gabor - einstein - spatial - freq ] ) shows gabor - einstein wavelets with varying spatial frequencies ; and figure ( [ fig : gabor - einstein - spatial - phase ] ) shows gabor - einstein wavelets of varying phase .
+ [ figur_sinc ] + + [ figur_ge_space ] in figure ( [ fig : gabor - einstein - center ] ) , one sees the anisometry that is conferred by the sinc function wave carrier .
the anisometry increases from fig ( [ fig : gabor - einstein - center]a ) through ( [ fig : gabor - einstein - center]d ) , due to the rapid fall off in the sinc function amplitude outside of the central peak over the two half cycles on each side of zero .
in contrast , fig ( [ fig : gabor - einstein - center]e ) and fig ( [ fig : gabor - einstein - center]f ) appear similar to each other in form , and each look more isometric than fig ( [ fig : gabor - einstein - center]d ) . this illustrates the presence of both isometric and anisometric basis elements which differ only in a single parameter ( phase , or time , or envelope center ) .
this expressiveness in the basis should allow for economical representation of a visual scene . of note ,
the amplitudes of the more isometric elements figs ( [ fig : gabor - einstein - center]e ) and ( [ fig : gabor - einstein - center]f ) are much smaller than those of the more anisometric preceding elements .
if desired , this can be readily modulated by the constant amplitude factor parameter .
+ + [ figur_ge3d_space ] figure ( [ fig : gabor - einstein - center-3d ] ) shows 3-dimensional plots illustrating the above discussed isometry modulation feature of the gabor - einstein basis .
figure ( [ fig : gabor - einstein - spatial - freq ] ) shows receptive fields with various spatial frequency tuning .
figure ( [ fig : gabor - einstein - spatial - phase ] ) shows various receptive fields which are phase - shifted relative to each other . for
any given spatial location in the receptive field , the peak amplitude of a cycle is phase dependent .
this is a feature distinct to the gabor - einstein model .
and like the anisometry - modulation feature , it confers expressiveness to the basis family .
this should allow for efficient representation of natural visual scenes . in the model
, time has a similar effect on peak amplitude . and
this behavior is in agreement with space - time inseparable receptive field profiles in cat striate cortex , where a time dependent phase drift has been shown @xcite .
+ [ figur_phi ] + [ figur_omega ] + + [ figur_sinc_fcn ] + + + [ figur_gef ] figure ( [ fig : sinc and fourier ] ) shows the essential property which the gabor - einstein wavelet inherits from the sinc function .
the sinc function s fourier transform is a lowpass filter .
higher frequency sinc functions have wider bandwidth .
bandpass filters are formed by taking the difference between sinc functions of different frequency as illustrated in figure ( [ fig : sinc and fourier ] ) .
the gabor - einstein wavelet s fourier transform has a step function factor inherited from the sinc function .
this allows it explain the particular distribution of lowpass to bandpass temporal frequency filter properties of v1 to mt neurons ( foster et al 1985 ; deangelis et al 1993b ; hawken et al 1996 ) in a manner innately representative of the motion - processing stream s neuronal hierarchy
. for simplicity of illustration , figure ( [ fig : ge_fourier ] ) shows a 2d ( x , t ) gabor - einstein element .
its fourier transform is shown in figs ( [ fig : ge_fourier]a ) and ( [ fig : ge_fourier]b ) , while its ( x , t ) domain representation is shown in fig ( [ fig : ge_fourier]c ) .
the lowpass temporal frequency nature of the fourier representation is apparent , as the support straddles the zero .
figure ( [ fig : ge_band ] ) shows the construction of temporal frequency bandpass gabor - einstein wavelets .
as illustrated , the temporal frequency bandpass gabor - einstein element results from the difference of two temporal frequency lowpass gabor - einstein elements .
fig ( [ fig : ge_band]a ) and fig ( [ fig : ge_band]b ) are plots of equation ( [ gabor - einstein - fourier ] ) , i.e. they are the fourier transform of the gabor - einstein wavelet described by equation ( [ gabor - einstein ] ) .
we label this basis element `` @xmath87 '' , meaning it is the zero - centered lowpass temporal frequency filter whose temporal frequency bandwidth is one . accordingly , fig ( [ fig : ge_band]c ) and fig ( [ fig : ge_band]d ) are plots of the `` @xmath88 '' gabor - einstein basis element , i.e. they are plots of the zero - centered lowpass temporal frequency filter whose temporal frequency bandwidth is three .
fig ( [ fig : ge_band]e ) and fig ( [ fig : ge_band]f ) plot the bandpass temporal frequency filter basis element obtained by taking the difference of the `` @xmath87 '' and `` @xmath88 '' basis elements .
the left - hand column of figure ( [ fig : ge_striatotemporal spectrum ] ) plots the gabor - einstein basis elements of fig ( [ fig : ge_band ] ) , while the right hand column plots the corresponding temporal response probed at the spatial origin of the receptive field .
there is an increasing complexity of the temporal waveform structure as one progresses from lowpass to bandpass basis element .
all pure gabor - einstein basis elements are lowpass temporal frequency filters . on the other hand , the bandpass temporal frequency filter property necessarily results from summation of gabor - einstein basis elements .
hence bandpass temporal elements are necessarily complex ( not pure )
. however , not all lowpass temporal frequency filters are pure ; and not all basis summations yield bandpass temporal frequency filters .
for instance , figure ( [ fig : ge_sum ] ) shows the sum of two pure gabor - einstein basis elements which yield another lowpass temporal frequency filter .
the temporal waveform of the composite element is indeed more complex than that of its two pure constituents , however , it appears less complex than the composite temporal waveform of fig ( [ fig : ge_band ] ) .
+ + [ figur_sts ] +
the neurons in the v1 to mt dorsal stream are highly specialized towards motion detection .
furthermore , they are embedded in a hierarchical order of increasing motion specialization .
the earliest ( least specialized ) members of the chain are the non - speed tuned v1 cells , while the most specialized members are the mt cells .
naturally , lgn cells precede input layer v1 cells both anatomically as well as in level of motion specialization .
similarly , cells located in the medial superior temporal ( mst ) area and the parietal lobe are further along than mt cells both anatomically and in motion specialization .
we have focused our current study on the v1 to mt spectrum .
we set out to obtain a good mathematical representation for the receptive fields of these neurons .
the term _ good _ here implies that the representation should faithfully capture the salient features of the motion specialization spectrum and serve as an effective label or address schema for neurons in the motion processing stream .
for instance , it is known that mt neurons receive input from v1 neurons and most directly from v1-to - mt projectors , therefore in a good basis , the representation for an mt neuron should result from some combination of representations of some v1 neuron(s ) .
furthermore , the resulting mt representation should naturally manifest known attributes of mt neurons .
appropriate comparison and discussion of the current work and related work requires a more precise language than currently available .
hence , we begin this discussion by introducing a simple framework for classifying and naming the components and hierarchical levels of receptive field models . in the literature
, there is currently no clear categorization of the various mathematical descriptions of receptive fields .
ironically , the laudable increase in discovery of receptive field attributes and modulation mechanisms may have led to further conceptual entanglement of the mathematical forms of receptive fields . here , we propose a simple framework to organize and classify receptive field descriptions and briefly review some basic concepts .
this should help place the current work , the gabor - einstein model , in the appropriate context ; and furthermore , should help organize and direct receptive field research .
we will identify a receptive field by its constituents and method of construction .
a receptive field can be constructed using two essential things : 1 .
primary input receptive fields 2 . input combination scheme the primary input receptive field is the response function of the neuron at the lowest level on the hierarchy of study .
for instance , when studying the motion - processing stream between v1 and mst , the primary input receptive field is the response function of a v1 input layer neuron .
an input combination scheme is a model for the network of neurons projecting onto the study neuron .
for instance , a common scheme is linear summation of weighted v1 responses to yield an mt neuron .
the input combination scheme can be classified according to the stage at which input processing , if any , occurs .
three classes arise : 1 .
pre - processing 2 .
intra - processing 3 .
post - processing pre - processing is modification of the input responses prior to combining them . for instance
, v1 elements are squared in the motion - energy model , after which they are summed to yield the mt response @xcite .
intra - processing is modification of the input responses during the combination process .
for example , the soft max scheme @xcite .
post - processing is modification of the response after input combination .
for instance , divisive normalization schemes @xcite .
the divisive normalization scheme , proposed by heeger in 1992 , is a neuron population based model of v1 to mt function . in that model ,
each neuron s behavior is modulated by that of a population module of which it was a member .
specifically , its manifest response is its native response divided by the net sum of the responses of the other neurons in its population module plus a regularization factor .
the model bears resemblance to receptor models in pharmacokinetics , where the regularization factor plays the role of a half - saturation constant .
the third categorization attribute is the hierarchical level from which input originates relative to the study neuron .
the three following classes arise : 1 .
feed - forward 2 .
isostratal 3 .
feedback feed - forward implies the input comes in from a lower hierarchical level .
for instance v1 projecting to mt .
isostratal implies that the input comes in from the same level .
for instance , normalization of mt response by neighborhood neuron pool @xcite .
feedback implies the input comes in from a higher level in the hierarchy . of note
, most receptive fields likely receive contributions from all three types of inputs . in light of the above framework ,
the gabor - einstein model is a primary input receptive field for the v1 to mt motion processing stream .
it is compatible with any of the aforementioned input combination schemes , hierarchical input routing schemes , and input processing schemes .
the work presented in this paper focused mainly on the primary input receptive field , and for simplicity of illustration , assumed a basic ( i.e. linear ) input combination scheme .
certain important phenomena can not be explained at the level of early ( e.g. input layer v1 ) primary input receptive fields alone .
next , we discuss some aspects of such phenomena including end - stopping , a short selection of non - linearity mechanisms , and contrast - modulated speed tuning .
a key distinguishing feature of the gabor - einstein model is how it naturally represents the distribution of lowpass to bandpass temporal frequency filtering properties along the v1 to mt motion specialization hierarchy .
ratio models are a class of models with a different objective .
the ratio models aim to mechanistically model the way speed tuned units arise from combinations of non - speed tuned bandpass and low pass temporal frequency filters .
these models are based on the idea that stimulus temporal frequency is proportional to the ratio of outputs of spatiotemporally separable bandpass to lowpass neuronal outputs @xcite .
various summation - based input combination schemes along the v1 to mt stream have been used .
heeger and colleagues proposed a two stage weighted linear summation model in which mt fields resulted from weighted sums of v1 fields @xcite .
sereno as well as nowlan & sejnowski proposed multilevel neural network reinforcement learning models in which mt neuron response was a function of a linear summation of v1 responses @xcite .
tsui et al proposed a model in which mt response was obtained by a soft - max weighted summation of v1 inputs @xcite .
each of these schemes differs fundamentally in approach from the gabor - einstein model .
namely , we focus on the receptive field of a single neuron along the v1 to mt pathway , implicitly representing the network spatiotemporal structure in the inherent attributes of the basis itself . specifically , the natural representation of temporal frequency filter distribution along the v1 to mt spectrum . in the gabor - einstein basis ,
the basis itself mandates summation as a means of ascending up the specialization hierarchy .
this is a gratifying result , and supports the notion that the sinc function is a natural way to describe the v1 to mt spectrum . here
, we discuss certain higher order phenomena whose description exceeds the scope of single early neuron receptive fields .
end - stopping is a center - surround type phenomenon in which for instance a full length bar stimulus results in a submaximal response , while some shorter bar stimulus generates maximal response .
in other words , the excitatory portion of the receptive field is shorter than the full length of the receptive field
. typically , the periphery of the receptive field , i.e. its _ end _ , elicits suppression ( or `` stopping '' ) .
hence neurons exhibiting such behavior are said to be `` end - stopped '' .
jones and colleagues found that majority ( 94% ) of v1 cells in the macaque were end - stopped to various degrees @xcite .
moreover , the most prominently end - stopped v1 cells have been found in layer 4b which is known to be the dominant projection layer to mt @xcite .
furthermore , tsui et al have argued a role for v1 end - stopping in mt motion integration and solution to the aperture problem @xcite .
it seems therefore that a complete model of v1 to mt neurons must account for end - stopping .
end - stopping phenomena would be difficult to explain at the single input layer v1 neuron stage , since it likely involves isostratal interactions .
tsui and colleagues modeled end - stopping by specifying the normalization pool as surround cells arranged along the orientation axis of a center cell . with this spatially - specific implementation of normalization ,
surround cell activity yielded end - stopping @xcite .
the response was then fit to a saturation kinetics model equivalent to heeger s divisive normalization model @xcite .
regarding center - surround phenomena such as end - stopping , the gabor - einstein wavelet can be considered a _ classical receptive field_. a scheme similar to that of tsui et al can be used to implement end stopping using the gabor - einstein wavelet as both center and surround , or in hybrid with other wavelet transform as either center or surround . of note
, other schemes have also been used to implement end - stopping .
for instance , skottun proposed an orientation - modulated cancellation strategy @xcite .
the gabor - einstein wavelet is compatible with this scheme as well .
the perception of motion is likely a higher order phenomenon that emerges out of the elaborate nonlinear connection network of which mt is only a part .
some of these non - v1 areas from which mt receives input include lgn @xcite , superior colliculus @xcite , and extrastriate regions @xcite .
some combination of these non - v1 inputs may explain the persistence of visual responsiveness and direction selectivity after v1 lesions or cooling @xcite .
indeed several researchers have drawn specific attention to the necessity of non - linear models . moreover , these nonlinearities are not just a feature of late ( specialized ) visual neurons , but have been demonstrated very early in the visual pathway .
for example , schwartz et al pointed out the need for nonlinear retinal ganglion cell receptive field models @xcite . and
rosenberg et al showed that lgn y cells utilize a nonlinear mechanism to represent interference patterns @xcite .
fine mechanistic modeling of the single v1 to mt neuron s receptive field s most salient features is certain to yield valuable physiological insight into the myriad non - v1 connections to mt .
contrast - modulated speed tuning is likely unexplainable at the level of early primary input receptive fields of single neurons .
existing explanations of such phenomena appear to necessarily involve nonlinear isostratal interactions .
for instance , divisive normalization @xcite . in general
, the effects of contrast on the receptive field of v1 to mt neurons can safely be assumed to be highly modulated by neighboring neurons .
significant data exists to support such population encoding .
for instance , deangelis et al showed that end and side stopping were strongest along the preferred excitatory orientation , yet superimposed suppressive stimuli was much more broadly tuned than the classical receptive field , suggesting a modulatory pool of neurons @xcite .
such neuron pools also likely mediate the effects of contrast on speed tuning .
krekelberg et al found that most mt cells in alert macaques preferred lower speeds for lower stimulus contrast @xcite .
priebe et al found that a significant population of v1 cells have contrast - modulated speed tuning curves @xcite .
livingstone and conway found that the speed tuning curves of most v1 neurons in alert macaques were dependent on contrast stimuli .
specifically , for lower contrast stimuli , v1 cells shifted their tuning curves to lower speeds ; while for higher contrast stimuli , the opposite was observed @xcite .
psychophysical studies also show that both high contrast stimuli and high spatial frequency stimuli are associated with higher perceived speeds @xcite .
the increase in preferred speed with contrast is a phenomenon that likely involves nonlinear isostratal interactions . as such , early single neuron receptive field models in isolation are unable to explain such phenomena .
nonetheless the primary input receptive field is the theoretical and physiological foundation of higher order phenomena .
hence it is essential to have a physiologically sound model such as the gabor - einstein wavelet which naturally generates the hierarchical specialization structure of the motion stream .
the gabor - einstein wavelet can be readily plugged - in to divisive normalization schemes which may be necessary to explain contrast - modulated phenomena .
in this paper , we introduced the gabor - einstein wavelet , a new family of functions for modeling the receptive fields of neurons in the v1 to mt motion processing stream .
we showed that the way space and time are mixed in the visual cortex has analogies to the way they are mixed in the special theory of relativity .
we therefore constrained the gabor - einstein model to a relativistically invariant wave carrier and to the minimum possible number of parameters .
these constraints yielded a sinc function wave carrier with energy - momentum relation as argument .
the model innately and efficiently represents the temporal frequency filtering property distribution along the motion processing stream .
specifically , on the v1 end of the stream , the neuron population has an equal proportion of lowpass to bandpass temporal frequency filters ; whereas on the mt end , they have mostly bandpass temporal frequency filters . from our analysis and simulations
, we showed that the distribution of temporal frequency filtering properties along the motion processing stream is a direct effect of the way the brain jointly encodes space and time .
we uncovered this fundamental link by demonstrating an analogous mathematical structure between the special theory of relativity and the joint encoding of space and time in the visual cortex .
the gabor - einstein model and the experiments it motivates will provide new physiological insights into how the brain represents visual information .
the author thanks rudi weikard , marius nkashama , john mayer , ian knowles , xiaobai sun , and peter blair for helpful suggestions on how to approach the fourier transform of the sinc function of multidimensional argument .
he thanks greg schwartz and ari rosenberg for helpful discussion on nonlinearities in visual receptive fields .
he thanks his mentors and role models in ophthalmology at howard university and the washington area for their dedication to training residents : robert a. copeland jr . , leslie s. jones , bilal khan , earl kidwell , janine smith - marshall , david katz , william deegan iii , melissa kern , ali ramadan , frank spellman , reggie sanders , and michael rivers , emily chew , brian brooks , and wai wong .
he thanks claude l. cowan jr whose excellence inspired him to pursue a career in medical retina .
he thanks his co - residents at howard ophthalmology for the wonderful experience which resulted from their enthusiasm for patient care and collegial learning : sir gawain dyer , salman j. yousuf , animesh petkar , ninita brown , mona kaleem , mikelson mompremier , saima qureshi , nikisha richards , neal desai , chris burris , natasha pinto , usiwoma abugo , chinwe okeagu , and katrina del fierro .
he thanks richard mooney , michael platt , pate skene , fan wang , vic nadler , and kafui dzirasa for supporting his membership to the society for neuroscience .
he thanks susan elner , mark w. johnson , john r. heckenlively , paul p. lee and all the amazing faculty at the university of michigan ann arbor ( w.k .
kellogg eye center ) for awarding me the clinical fellowship training position in medical retina .
he thanks stuart fine and the board of directors of the heed ophthalmic foundation ( nicholas j. volpe , stephen mcleod , joan miller , eduardo alfonso , david wilson , julia heller , and froncie gutman ) for inviting him to participate in the 8th annual heed resident s retreat in chicago illinois .
it was a delight to be at that retreat .
he thanks eydie miller - ellis , mildred olivier , and the the rabb venable excellence in research program board for selecting him as a participant and for encouraging a career in academia . 90 [ 1]#1 [ 1]`#1 ` urlstyle [ 1]doi : # 1 e. h. adelson and j. r. bergen .
spatiotemporal energy models for the perception of motion .
a _ , 20 ( 2):0 284299 , 1985 .
e. h. adelson and j. r. bergen .
the extraction of spatiotemporal energy in human and machine vision . in _ proc .
ieee workshop on visual motion , charleston _ ,
pages 151156 , 1986 .
j. c. anderson and k. a. c. martin . .
_ journal of comparative neurology _
, 4430 ( 1):0 5670 , 2002 . j. c. anderson , t. binzegger , k. a. c. martin , and k. s. rockland . .
_ the journal of neuroscience _ , 180 ( 24):0 1052510540 , 1998 .
a. anzai , i. ohzawa , and r. d. freeman .
joint - encoding of motion and depth by visual cortical neurons : neural basis of the pulfrich effect .
_ nature neuroscience _ , 40 ( 5):0 513518 , 2001 .
p. azzopardi , m. fallah , c. g. gross , and h. r. rodman . .
_ neuropsychologia _ , 410 ( 13):0 17381756 , 2003 . r. a. berman and r. h. wurtz . .
_ the journal of neuroscience _ , 300 ( 18):0 63426354 , 2010 . r. a. berman and r. h. wurtz . . _ the journal of neuroscience _ , 310 ( 2):0 373384 , 2011 .
d. boussaoud , l. g. ungerleider , and r. desimone .
pathways for motion analysis : cortical connections of the medial superior temporal and fundus of the superior temporal visual areas in the macaque .
_ journal of comparative neurology _
, 2960 ( 3):0 462495 , 1990 .
k. r. brooks , t. morris , and p. thompson . .
_ journal of vision _ , 110 ( 14 ) , 2011 .
d. c. burr and j. ross .
how does binocular delay give information about depth ?
_ vision research _ , 190 ( 5):0 523532 , 1979 .
t. carney , m. a. paradiso , and r. d. freeman .
a physiological correlate of the pulfrich effect in cortical neurons of the cat .
_ vision research _ , 290 ( 2):0 155165 , 1989 .
m. m. churchland , n. j. priebe , and s. g. lisberger .
comparison of the spatial limits on direction selectivity in visual areas mt and v1 .
_ journal of neurophysiology _ , 930 ( 3):0 12351245 , 2005 .
j. g. daugman .
uncertainty relation for resolution in space , spatial frequency , and orientation optimized by two - dimensional visual cortical filters .
_ optical society of america , journal , a : optics and image science _ , 20 ( 7):0 11601169 , 1985 .
g. c. deangelis , i. ohzawa , and r. d. freeman . .
_ journal of neurophysiology _ , 690 ( 4):0 10911117 , 1993 . g. c. deangelis , i. ohzawa , and r. d. freeman . .
_ journal of neurophysiology _ , 690 ( 4):0 11181135 , 1993 . g. c. deangelis , r. d. freeman , and i. ohzawa . length and width tuning of neurons in the cat s primary visual cortex
. _ journal of neurophysiology _ , 710 ( 1):0 347374 , 1994 .
r. dubner and s. m. zeki .
response properites and receptive fields of cells in an anatomically defined region of the superior temporal sulcus in the monkey .
_ brain research _ , 1971
. a. einstein . on the electrodynamics of moving bodies .
_ annalen der physik _ , 170 ( 891):0 50 , 1905 .
d. j. felleman and j. h. kaas .
receptive - field properties of neurons in middle temporal visual area ( mt ) of owl monkeys .
_ journal of neurophysiology _ , 520 ( 3):0 488513 , 1984 . d. ferster and k. d. miller .
neural mechanisms of orientation selectivity in the visual cortex .
_ annual review of neuroscience _ , 230 ( 1):0 441471 , 2000 .
k. h. foster , j. p. gaska , m. nagler , and d. a. pollen . .
_ the journal of physiology _ , 3650 ( 1):0 331363 , 1985 .
d. gabor . .
_ electrical engineers - part iii : radio and communication engineering , journal of the institution of _ , 930 ( 26):0 429441 , 1946 .
p. girard , p. a. salin , and j. bullier . .
_ journal of neurophysiology _ , 670 ( 6):0 14371446 , 1992 .
m. j. hawken , r. m. shapley , and d. h. grosof .
temporal - frequency selectivity in monkey visual cortex .
_ visual neuroscience _ , 130 ( 03):0 477492 , 1996
. d. j. heeger .
model for the extraction of image flow .
_ josa a _ , 40 ( 8):0 14551471 , 1987 .
d. j. heeger .
normalization of cell responses in cat striate cortex .
_ visual neuroscience _ , 90 ( 02):0 181197 , 1992 .
d. j. heeger .
modeling simple - cell direction selectivity with normalized , half - squared , linear operators .
_ journal of neurophysiology _ , 700 ( 5):0 18851898 , 1993 .
d. h. hubel .
tungsten microelectrode for recording from single units .
_ science _ ,
1250 ( 3247):0 549550 , 1957 . d. h. hubel .
single unit activity in striate cortex of unrestrained cats . _ the journal of physiology _ , 1470 ( 2):0 226238 , 1959 .
d. h. hubel and t. n. wiesel .
receptive fields of single neurones in the cat s striate cortex . _ the journal of physiology _ ,
1480 ( 3):0 574591 , 1959 . h. e. jones , k. l. grieve , w. wang , and a. m. sillito . .
_ journal of neurophysiology _ , 860 ( 4):0 20112028 , 2001 .
b. krekelberg , r. j. a. van wezel , and t. d. albright .
interactions between speed and contrast tuning in the middle temporal area : implications for the neural code for speed . _ the journal of neuroscience _ , 260 ( 35):0 89888998 , 2006 .
s. w. kuffler .
discharge patterns and functional organization of mammalian retina .
_ j. neurophysiol .
_ , 160 ( 1):0 3768 , 1953 . i. lampl , d. ferster , t. poggio , and m. riesenhuber .
intracellular measurements of spatial integration and the max operation in complex cells of the cat primary visual cortex .
_ journal of neurophysiology _ , 920 ( 5):0 27042713 , 2004 . d. n. lee . a stroboscopic stereophenomenon .
_ vision research _ , 100 ( 7):0 587593 , 1970 .
m. livingstone and d. hubel .
segregation of form , color , movement , and depth : anatomy , physiology , and perception .
_ science _ , 2400 ( 4853):0 740749 , 1988 . m. s. livingstone and b. r. conway . .
_ journal of neurophysiology _ , 970 ( 1):0 849857 , 2007 .
s. marelja .
mathematical description of the responses of simple cortical cells .
_ josa _ , 700 ( 11):0 12971300 , 1980 . j. h. maunsell and d. c. van essen .
the connections of the middle temporal visual area ( mt ) and their relationship to a cortical hierarchy in the macaque monkey . _ the journal of neuroscience _ , 30 ( 12):0 25632586 , 1983 .
j. h. maunsell and d. c. van essen . .
_ journal of neurophysiology _ , 490 ( 5):0 11271147 , 1983 .
j. h. maunsell , t. a. nealey , and d. d. depriest . .
_ the journal of neuroscience _ , 100 ( 10):0 33233334 , 1990 .
j. h. r. maunsell and w. t. newsome . visual processing in monkey extrastriate cortex .
_ annual review of neuroscience _ , 100 ( 1):0 363401 , 1987 . j. c. maxwell . .
_ the london , edinburgh , and dublin philosophical magazine and journal of science _ , 210 ( 141):0 338348 , 1861 . j. c. maxwell .
_ the london , edinburgh , and dublin philosophical magazine and journal of science _ , 210 ( 140):0 281291 , 1861 . j. c. maxwell . .
_ the london , edinburgh , and dublin philosophical magazine and journal of science _ , 230 ( 152):0 8595 , 1862 .
j. c. maxwell . .
_ the london , edinburgh , and dublin philosophical magazine and journal of science _ , 230 ( 151):0 1224 , 1862 .
j. mclean and l. a. palmer .
contribution of linear spatiotemporal receptive field structure to velocity selectivity of simple cells in area 17 of cat . _ vision research _ , 290 ( 6):0 675679 , 1989 .
m. j. morgan .
pulfrich effect and the filling in of apparent motion .
_ perception _ , 50 ( 2):0 187195 , 1976
. m. j. morgan .
perception of continuity in stroboscopic motion : a temporal frequency analysis . _ vision research _ , 190 ( 5):0 491500 , 1979 .
j. a. movshon and w. t. newsome . .
_ the journal of neuroscience _ , 160 ( 23):0 77337741 , 1996 .
j. j. nassi , d. c. lyon , and e. m. callaway . .
_ neuron _ , 500 ( 2):0 319327 , 2006 .
s. j. nowlan and t. j. sejnowski . a selection model for motion processing in area mt of primates . _ the journal of neuroscience _ , 150 ( 2):0 11951214 , 1995 .
c. c. pack , r. t. born , and m. s. livingstone .
two - dimensional substructure of stereo and motion interactions in macaque visual cortex .
_ neuron _ , 370 ( 3):0 525535 , 2003 . j. a. perrone . a visual motion sensor based on the properties of v1 and mt neurons .
_ vision research _ , 440 ( 15):0 17331755 , 2004
. j. a. perrone .
economy of scale : a motion sensor with variable speed tuning .
_ journal of vision _ , 50 ( 1 ) , 2005 .
j. a. perrone and a. thiele . a model of speed tuning in mt neurons . _ vision research _ , 420 ( 8):0 10351051 , 2002 . n. j. priebe , s. g. lisberger , and j. a. movshon .
tuning for spatiotemporal frequency and speed in directionally selective neurons of macaque striate cortex . _ the journal of neuroscience _ , 260 ( 11):0 29412950 , 2006 .
carl pulfrich .
die stereoskopie i m dienste der isochromen und heterochromen photometrie .
_ naturwissenschaften _ , 100 ( 35):0 751761 , 1922 .
n. qian . computing stereo disparity and motion with known binocular cell properties . _ neural computation _ , 60 ( 3):0 390404 , 1994 .
n. qian and r. a. andersen . a physiological model for motion - stereo integration and a unified explanation of pulfrich - like phenomena .
_ vision research _ , 370 ( 12):0 16831698 , 1997 . n. qian and r. d. freeman .
pulfrich phenomena are coded effectively by a joint motion - disparity process . _ journal of vision _ , 90 ( 5 ) , 2009 .
n. qian , r. a. andersen , and e. h. adelson . .
_ the journal of neuroscience _ , 140 ( 12):0 73817392 , 1994 . j. c. a. read and b. g. cumming .
. _ journal of neurophysiology _ , 940 ( 2):0 15411553 , 2005 . j. c. a. read and b. g. cumming .
the stroboscopic pulfrich effect is not evidence for the joint encoding of motion and depth .
_ journal of vision _ , 50 ( 5 ) , 2005 .
j. c. a. read and b. g. cumming .
all pulfrich - like illusions can be explained without joint encoding of motion and disparity . _
journal of vision _ , 50 ( 11 ) , 2005 .
m. riesenhuber and t. poggio .
hierarchical models of object recognition in cortex .
_ nature neuroscience _ , 20 ( 11):0 10191025 , 1999 .
k. s. rockland . .
_ , 30 ( 2):0 15570 , 1989 .
k. s. rockland . .
_ journal of comparative neurology _ , 3550 ( 1):0 1526 , 1995 .
h. r. rodman and t. d. albright .
single - unit analysis of pattern - motion selective properties in the middle temporal visual area ( mt ) .
_ experimental brain research _ , 750 ( 1):0 5364 , 1989 .
h. r. rodman , c. g. gross , and t. d. albright . .
_ the journal of neuroscience _ , 100 ( 4):0 11541164 , 1990 .
a. rosenberg , t. r. husson , and n. p. issa .
subcortical representation of non - fourier image features .
_ the journal of neuroscience _ , 300 ( 6):0 19851993 , 2010 .
m. p. sceniak , m. j. hawken , and r. shapley . visual spatial characterization of macaque v1 neurons . _
journal of neurophysiology _ , 850 ( 5):0 18731887 , 2001 .
g. w. schwartz , h. okawa , f. a. dunn , j. l. morgan , d. kerschensteiner , r. o. wong , and f. rieke . the spatial structure of a nonlinear receptive field . _ nature neuroscience _ , 150 ( 11):0 15721580 , 2012
. m. e. sereno .
_ neural computation of pattern motion : modeling stages of mation analysis in the primate visual cortex_. the mit press , 1993 .
s. shipp and s. zeki . .
_ european journal of neuroscience _ , 10 ( 4):0 309332 , 1989 . e. p. simoncelli and d. j. heeger . .
_ vision research _ , 380 ( 5):0 743761 , 1998 . l. c. sincich , k. f. park , m. j. wohlgemuth , and j. c. horton . .
_ nature neuroscience _ , 70 ( 10):0 11231128 , 2004 . b. c. skottun . a model for end - stopping in the visual cortex . _ vision research _ , 380 ( 13):0 20232035 , 1998 .
m. a. smith , n. j. majaj , and j. a. movshon .
dynamics of motion signaling by neurons in macaque area mt . _ nature neuroscience _ , 80 ( 2):0 220228 , 2005 .
d. j. tolhurst , c. r. sharpe , and g. hart .
the analysis of the drift rate of moving sinusoidal gratings .
_ vision research _ , 130 ( 12):0 25452555 , 1973 . j. m. g. tsui , j. n. hunter , r. t. born , and c. c. pack .
the role of v1 surround suppression in mt motion integration . _
journal of neurophysiology _ , 1030 ( 6):0 31233138 , 2010 .
l. g. ungerleider and r. desimone .
cortical connections of visual area mt in the macaque .
_ journal of comparative neurology _ ,
2480 ( 2):0 190222 , 1986 .
d. c. van essen , j. h. r. maunsell , and j. l. bixby .
the middle temporal visual area in the macaque : myeloarchitecture , connections , functional properties and topographic organization .
_ journal of comparative neurology _ , 1990
( 3):0 293326 , 1981 .
a. b. watson and a. ahumada . _ a look at motion in the frequency domain _ ,
volume 84352 . national aeronautics and space administration , ames research center , 1983 .
a. b. watson and a. j. ahumada , jr . model of human visual - motion sensing .
_ journal of the optical society of america . a , optics and image science _ , 20 ( 2):0 322341 , 1985 .
x. xu , j. ichida , y. shostak , a. b. bonds , and v. a. casagrande .
_ visual neuroscience _ , 190 ( 1):0 97108 , 2001 . n. h. yabuta , a. sawatari , and e. m. callaway .
two functional channels from primary visual cortex to dorsal visual cortical areas . _
science _ , 2920 ( 5515):0 297300 , 2001 .
s. m. zeki .
cells responding to changing image size and disparity in the cortex of the rhesus monkey . _ the journal of physiology _ ,
2420 ( 3):0 827841 , 1974 . s. m. zeki .
the response properties of cells in the middle temporal area ( area mt ) of owl monkey visual cortex .
_ proceedings of the royal society of london .
series b. biological sciences _ , 2070 ( 1167):0 239248 , 1980 . | our visual system is astonishingly efficient at detecting moving objects .
this process is mediated by the neurons which connect the primary visual cortex ( v1 ) to the middle temporal ( mt ) area .
interestingly , since kuffler s pioneering experiments on retinal ganglion cells , mathematical models have been vital for advancing our understanding of the receptive fields of visual neurons . however , existing models were not designed to describe the most salient attributes of the highly specialized neurons in the v1 to mt motion processing stream ; and they have not been able to do so . here , we introduce the gabor - einstein wavelet , a new family of functions for representing the receptive fields of v1 to mt neurons .
we show that the way space and time are mixed in the visual cortex is analogous to the way they are mixed in the special theory of relativity ( str ) .
hence we constrained the gabor - einstein model by requiring : ( i ) relativistic - invariance of the wave carrier , and ( ii ) the minimum possible number of parameters .
from these two constraints , the sinc function emerged as a natural descriptor of the wave carrier .
the particular distribution of lowpass to bandpass temporal frequency filtering properties of v1 to mt neurons ( foster et al 1985 ; deangelis et al 1993b ; hawken et al 1996 ) is clearly explained by the gabor - einstein basis .
furthermore , it does so in a manner innately representative of the motion - processing stream s neuronal hierarchy .
our analysis and computer simulations show that the distribution of temporal frequency filtering properties along the motion processing stream is a direct effect of the way the brain jointly encodes space and time .
we uncovered this fundamental link by demonstrating that analogous mathematical structures underlie str and the joint cortical encoding of space and time .
this link will provide new physiological insights into how the brain represents visual information .
( 1,0)300 |
the 21-cm transition of hydrogen ( ) traces the cool component of neutral gas throughout the universe , the raw material fuelling star formation . in emission ,
the low probability of the 21-cm spin - flip transition compounded by the inverse square law , renders this essentially undetectable at @xmath3 . in absorption , however , this sensitivity limit can be avoided since the line strength is independent of distance , depending only upon the column density of the absorbing gas and the flux from the background emission source .
nevertheless , detections remain rare with only 77 21-cm absorption systems known at @xmath4 35 _ associated _ with the quasar host galaxy ( @xmath5 ) , in addition to 42 systems located at some intermediate redshift along the sight - line to the quasar ( @xmath6 ) .
these _ intervening _ absorbers occur in known or candidate , thus not being sufficiently redshifted to be detected in the optical band by ground - based telescopes . ]
damped lyman-@xmath0 absorption systems ( dlas , defined by their large neutral hydrogen column densities , @xmath7 ) .
currently there are 1500 dlas known @xcite , which could contain over 80% of the neutral gas mass density of the universe @xcite .
the velocity integrated optical depth of the 21-cm absorption , @xmath8 , is related to the total neutral hydrogen column density via @xcite @xmath9 where @xmath10 [ k ] is the mean harmonic spin temperature of the gas , @xmath11 is the depth of the line ( or r.m.s .
noise in the case of a non - detection ) and @xmath12 & @xmath13 the flux density and covering factor of the background continuum source , respectively .
this latter parameter accounts for the possibility that the absorbing gas may not fully cover all of the background radio emission pc in diameter , whereas radio sources can exceed @xmath14 pc ( see @xcite ) . ] and so all of the measured flux , @xmath12 , may not be intercepted , decreasing the sensitivity of the observation . in both the associated and intervening cases
the detections occur overwhelmingly at @xmath15 . for the associated systems , @xmath16 is generally not known , although @xcite attribute the lack of absorption at high redshift to the traditional optical selection of targets biasing searches towards the most luminous objects at these large distances , where the ultra - violet luminosity ionises the neutral gas . in the optically thin regime ( @xmath17 ) ,
[ enew ] reduces to @xmath18 , thus giving a degeneracy in the spin temperature / covering factor ratio ( @xmath19 ) , for the dlas where the column density is known . therefore , since there is a mix of detections and non - detections in dlas at @xmath20 , whereas at @xmath21 21-cm tends not to be detected , @xcite suggest an evolution in the spin temperature of the gas , where this is exclusively high at high redshift .
this has direct bearing on the relative proportions of the cold and warm neutral media ( e.g. @xcite ) and thus the reservoir of raw material available for star formation in the early universe .
however , the conclusion that the spin temperature evolves with redshift relies on the assumption of a covering factor for each dla , and @xcite have shown that , if the degeneracy is left intact , @xmath19 is not the order of magnitude larger at @xmath21 , as suggested by the @xmath10 distribution of @xcite .
furthermore , @xcite suggest that geometry effects introduced by an expanding universe could contribute to the distribution , indicating that the covering factor may actually be the dominant parameter .
since then there have been further high redshift searches @xcite .
this takes the number of confirmed dlas and sub - dlas searched for 21-cm to 25 at @xmath22 , the same number as at @xmath23 , thus allowing us to adequately address the spin temperature versus covering factor issue .
although there are now equal numbers of 21-cm searches above and below @xmath24 , the detection rate at @xmath1 remains considerably lower ( 5 out of 25 ) than for the lower redshift sample ( 14 out of 25 ) . since the velocity integrated optical depth of the 21-cm absorption , normalised by the total neutral hydrogen column density , is proportional to @xmath25 ( sect .
[ intro ] ) , this suggests that either @xmath10 is generally higher at @xmath1 or that @xmath13 is generally lower , or a combination of both , elevating the ratio at high redshift , in order to obtain limits to the line strengths for the non - detections we must assign a profile width to each absorber .
this is usually assumed , but here we estimate the full - width half maximum ( fwhm ) of the profile from the 21-cm fwhm
mgii 2796 equivalent width correlation for the detections , specifically fwhm [ ] @xmath26 [ ] @xcite . where the mgii equivalent widths are not available ( generally at @xmath27 where the transition is redshifted out of the optical band ) , we estimate the putative fwhm via the metallicity mgii equivalent width correlation @xcite , specifically @xmath28 } + 4.0 $ ] @xcite
when neither the mgii equivalent width nor metallicity are available , we assign a fwhm of @xmath29 , the mean of the detections . given the fwhm of the putative line , in order to obtain the best signal - to - noise ratio , based upon the detection of a single channel we would smooth the data to a spectral resolution of @xmath30fwhm .
this has the effect of improving / degrading the measured r.m.s noise , and thus the optical depth , by a factor of @xmath31 .
the newly derived integrated optical depth of the line is thus @xmath32_{\rm new}\approx \tau_{\rm peak}\times \sqrt{\frac{\delta v}{\rm fwhm } } \times { \rm fwhm } \nonumber \\
= \tau_{\rm peak}\times \sqrt{\delta v\,{\rm fwhm}},\end{aligned}\ ] ] where @xmath33 is the peak optical depth of the putative line .
this has the overall effect of scaling the previously published integrated optical depths , where these are calculated per channel , by @xmath34_{\rm new}/\left[\int\tau dv \right]_{\rm previous}$ ]
@xmath35 we therefore recalculate the limits from the published r.m.s .
noise levels and spectral resolutions and in fig .
[ comp - spin ] show the effect this has on the @xmath19 limits compared to the previous values . for the method prescribed above ( shown as circles ) ,
the largest change is for the @xmath36 absorber towards pks0118272 , for which the limit climbs from @xmath37 k @xcite to @xmath38 k and , on average , the scaled values are 160 k `` warmer '' than those published .
as seen from the figure , however , the new limits are mostly concentrated close the @xmath34_{\rm new } = \left[\int\tau dv \right]_{\rm previous}$ ] line , with equal numbers of points to either side .
@xcite attribute the correlation between the velocity spreads of the 21-cm and mgii profiles to the neutral and singly ionised gas being spatially coincident in 21-cm detected dlas .
however , given the possibility that much of the ionised magnesium could arise in warm photoionised gas , and is thus not tracing the same sight - line as the 21-cm , in fig .
[ comp - spin ] we also show the @xmath19 limits derived from assuming a single fwhm . for 10
( e.g. @xcite ) , we find the largest difference to be 4300 k , with a mean rise of 600 k and for 25 , the largest difference is 2500 k , with a mean decrease of 120 k. thus , scalings implementing the mean fwhm of the detections have less of an effect than using 10 , although either method is preferable to the inhomogeneously calculated published values , since each is consistent in smoothing the data to the fwhm and thus calculating the limit per smoothed channel . in fig .
[ toverf ] we show the newly derived limits against the absorption redshift .
applying a survival analysis to the @xmath39 lower limits in @xmath19 , via the asurv package @xcite , we obtain mean values of @xmath40 and @xmath41 . that is , the mean value of @xmath19 at @xmath1 is double that at @xmath42 ( 3600 cf .
this is considerably less than the ratio of 8.0 between the two redshift regimes ( 8100 cf .
1020 k ) found by applying the survival analysis to the spin temperatures of @xcite . since our recalculation of the limits has actually slightly increased the mean value of the spin temperature / covering factor degeneracy , we attribute this significant difference to the covering factor being left as a free parameter .
the addition of the new data , together with our treatment of the non - detections , indicates that a factor of only @xmath43 in the spin temperature / covering factor ratio between the two redshift regimes need be accounted for .
however , it remains unclear whether @xmath10 or @xmath13 is the dominant factor .
there have been several attempts to quantify the covering factor through very long baseline array ( vlba ) imaging of the background quasar , in which @xmath13 is estimated as the ratio of the compact unresolved component s flux to the total radio flux ( e.g. @xcite ) . from this , @xcite conclude that for the 21-cm searches , @xmath44 and hence low covering factors can not explain the dearth of detections at high redshift .
however , this method gives no information on the extent of the absorber nor on how effectively it covers the emission .
neither does it give information on the depth of the line when the extended continuum emission is resolved out .
so although an evolution in spin temperature with redshift is implied , this method invokes several crucial assumptions which could even lead to unphysical results ( e.g. the @xmath45 pc obtained for the extent of pks 0405331 at 398 mhz , @xcite ) . on the other hand , due to the effects of an expanding universe , objects at @xmath46 are always close to a single angular diameter distance ( @xmath47 mpc ) , irrespective of redshift ( e.g. @xcite ) . beyond @xmath48 , the angular diameter distance
_ decreases _ with increasing redshift , which will lead to a natural imbalance between the low and high - z dla samples . in the small angle approximation
the covering factor can be defined as @xmath49 where @xmath50 is the angle subtended by the background emission and @xmath51 and @xmath52 the linear sizes of the cool absorbing gas and background 21-cm emission , respectively ( fig .
[ dla_schem ] ) .
since absorbers at @xmath53 are always at a similar angular diameter distance as the background quasar ( @xmath54 ) , these will therefore , for a given system , always have a lower coverage than a counterpart with @xmath55 @xcite , an inequality which can only exist at @xmath15 . in fig .
[ distance - z ] we show the angular diameter distance ratio , @xmath56 , versus the absorption redshift for the 21-cm searches . note that this distribution is free from any assumptions ( including our own with regard to the fwhms of the non - detections ) . from this
we see a mix of @xmath56 at low redshift ( the low values are due to @xmath15 dlas towards much higher redshift quasars ) , whereas values at @xmath53 are exclusively high .
that is , a reproduction of the `` spin temperature '' distribution of @xcite .
at least the pattern if not the magnitude , given that @xcite suggest that the average spin temperature at @xmath1 is eight times that at @xmath2 , which is clearly not reproduced in fig .
[ distance - z ] . referring to sect .
[ new ] , however , we only require @xmath19 at @xmath1 to be about double that at @xmath2 in order to account for the distribution . from equ .
[ f ] , for a given absorber and emitter size , @xmath57 , and we find mean values of @xmath58 and @xmath59 , a ratio of 1.6 .
it is therefore apparent that the distribution in angular diameter distance ratios , and their effect on the covering factor , is close to that required to account for the observed @xmath19 distribution .
updating the statistics with the new results , we find a binomial probability of @xmath60 of obtaining at least 9 out of 11 detections at @xmath61 and 29 out of 39 non - detections at @xmath62 , by chance .
is the cut used by @xcite and , although arbitrary , this is the lowest ratio which gives an appreciable enough sample size in the lower bin .
when a ratio of 0.9 is used there are 38 dlas with @xmath63 and 12 with @xmath64 , which increases the significance of the correlation to @xmath65 . ]
that is , a significance of @xmath66 , lending weight to the argument that large angular diameter distance ratios are correlated with the non - detection of 21-cm absorption .
since there is no physical link between the spin temperature and angular diameter distance ratio , this suggests that , in the absence of any knowledge of the absorption cross - section in relation to the emitter extent , the covering factor plays a much more major rle than the spin temperature in the detection of 21-cm absorption .
with the advent of 21-cm absorption searches in high redshift dlas there has been much debate on whether an evolution in spin temperature or lower covering factors are responsible for the relatively low number of detections at @xmath1 .
new searches of sdss dlas , which occult the sight - lines to radio - loud quasars , take the number searched at these redshifts to the same as at @xmath2 , giving much needed data with which to address this issue . by giving the 21-cm non - detections a more thorough analysis than in other studies , through estimating the fwhms of the non - detected profiles from the observed optical properties and applying a survival analysis to the lower limits in @xmath19 , we find : 1 .
that mean the spin temperature / covering factor ratio at @xmath1 is only double that found at @xmath2 .
this is significantly lower than the factor of eight in spin temperature , determined from previous estimates of the covering factor .
this reduced factor can be accounted for by the effects of an expanding universe , where the angular diameter distance at @xmath67 is always close to , and sometimes exceeds , that of the background radio source .
that is , the high redshift dlas are _ always _ disadvantaged in how effectively they can intercept the background radio flux for a given dla
3 . since the non - detections tend to be associated with high ratios in the dla qso angular diameter distances ( @xmath68 ) , whereas the detections are associated with low ratios ( @xmath69 ) , at _ all _ redshifts , this strongly suggests that the coverage of the background radio source by the absorber is the dominant factor .
this runs contrary to previous studies which estimate the covering factor via vlba imaging , which support an evolution in the spin temperature .
however , these estimates are based upon several pivotal assumptions , whereas the angular diameter distance ratios are not .
we therefore suggest that there is no evidence for the spin temperature to increase with redshift , which implies no evolution in the fraction of the warm neutral gas content of dlas .
@xcite find a @xmath19metallicity anti - correlation .
if the spin temperature does not evolve with redshift , as does the metallicity @xcite , this may imply that the metallicity is proportional to the covering factor , which could be due to the larger absorbing cross - sections of the larger galaxies , as evident through the spirals which have the largest metallicities and the lower values of @xmath19 @xcite .
+ 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 also made use of nasa s astrophysics data system bibliographic service and asurv rev 1.2 @xcite , which implements the methods presented in @xcite .
the centre for all - sky astrophysics is an australian research council centre of excellence , funded by grant ce11e0090 . | recent high redshift surveys for 21-cm absorption in damped lyman-@xmath0 absorption systems ( dlas ) take the number of published searches at @xmath1 to 25 , the same number as at @xmath2 , although the detection rate at high redshift remains significantly lower ( 20% cf .
60% ) . using the known properties of the dlas to estimate the unknown profile widths of the 21-cm non - detections and including the limits via a survival analysis
, we show that the mean spin temperature / covering factor degeneracy at high redshift is , on average , double that of the low redshift sample .
this value is significantly lower than the previous factor of eight for the spin temperatures and is about the same factor as in the angular diameter distance ratios between the low and high redshift samples .
that is , without the need for the several pivotal assumptions , which lead to an evolution in the spin temperature , we show that the observed distribution of 21-cm detections in dlas can be accounted for by the geometry effects of an expanding universe .
that is , as yet there is no evidence of the spin temperature of gas rich galaxies evolving with redshift . |
in @xcite pirashvili makes explicit a definition of higher order hochschild homology of a @xmath5-algebra @xmath0 with coefficients in an @xmath0-module @xmath1 , implicitly defined in @xcite by anderson .
this is done by considering the composition of functors : @xmath6 @xmath7 and @xmath8 where we let @xmath9 denote the category of finite sets and where @xmath10 is a pointed simplicial set .
this composition yields a simplicial vector space , where the homology of the associated chain complex serves as the definition of higher order hochschild homology . one can naturally extend the definition given in @xcite to also define a notion of higher order hochschild cohomology ( see @xcite for a precise definition ) .
a reason for doing so would be to consider deformation theory in this new setting .
this connection between traditional hochschild cohomology and deformation theory has been studied for some time . in @xcite
, gerstenhaber illustrates the connection between deformations of an associative @xmath5-algebra @xmath0 and @xmath11-cocycles in the hochschild cohomology of @xmath0 with trivial coefficients by showing that @xmath12 is the group of isomorphism classes of square - zero deformations of @xmath0 , and in @xcite and @xcite , yau and laudal , respectively , show the connection between deformations of an @xmath0-module @xmath1 with @xmath2-cocycles of the hochschild cochain complex of @xmath0 with coefficients in @xmath3 more precisely , first order deformations of an @xmath0-module @xmath1 are in bijection with @xmath13 upon examination , it can be seen that the current definition for higher order hochschild cohomology allows only symmetric bi - module coefficients , however to study deformations of a module we need to consider the nonsymmetric bi - modules @xmath3
one knows that classical hochschild cohomology takes coefficients in a bi - modules , so we ask the following question . what coefficients can higher order hochschild cohomology take ?
the goal of this paper thus becomes two fold .
we aim to extend the definition of higher order hochschild cohomology to include multi - modules ( not necessarily symmetric ) as coefficients and in particular to determine what type of coefficient modules can be used when choosing a simplicial set @xmath4 to construct a cosimplicial @xmath5-vector space over . with this in mind , we aim to show the following theorem .
[ mainthm ] let @xmath0 be a commutative @xmath5-algebra .
given a pointed simplicial set @xmath4 , there exists a cosimplicial @xmath5-vector space @xmath14 associated to an @xmath0-module @xmath1 given by @xmath15 with coface and codegeneracy maps given by @xmath16 and @xmath17 if the actions @xmath18 on @xmath1 satisfy the following for simplices @xmath19 and @xmath20 : * @xmath21 if @xmath22 and the dimension of @xmath23 is at least 2 and @xmath24 we call this a
_ sweep around_. * @xmath25 if @xmath26 and the dimension of @xmath23 is at least 2 .
we call this a _ sweep out 1_. * @xmath27 if @xmath28 and there exists a @xmath23 of dimension at least 2 where @xmath29 , @xmath30 and @xmath24 we call this a _ sweep across _ .
* @xmath31 if @xmath32 and the dimension of @xmath23 is at least 2 .
we call this a _ sweep out 2_. where @xmath33 represents the @xmath34 action of @xmath35 whenever @xmath36 and @xmath37 and @xmath38 .
we take a product of such actions to represent the composition of the actions , which we assume to be commutative ( i.e. if @xmath1 has two actions , we actually assume that @xmath1 is a bimodule ) . for @xmath15 in theorem [ mainthm ]
we assume that the trivial @xmath5 as a tensor factor in the domain represents the base point .
while the list of axioms in theorem [ mainthm ] may seem random , we give a intuitive visual description of each axiom in section [ vis ] and in section [ examples ] we use the visual descriptions to give results for a variety of simplicial sets .
we can now define the following .
the cohomology of the cochain complex associated to @xmath14 ( by taking alternating sums of coface maps ) is the _ higher order hochschild cohomology _ of @xmath0 with coefficients in @xmath39 which we denote as @xmath40 for the remainder of this paper , we fix the field @xmath5 and assume @xmath0 is a commutative algebra over @xmath41 we also assume the @xmath0-module @xmath1 can have multiple actions as our quest is to determine what actions may be present in the associated higher order hochschild cochain complex .
we would like to see that the definition given in section [ intro ] agrees with both the classical hochschild cohomology definition as well as the higher order hochschild cohomology definition given by pirashvili .
we provide the connections through the following examples .
let @xmath42 be the simplicial set of @xmath43 which contains one 0-simplex and one non - degenerate 1-simplex .
given an @xmath0 bi - module @xmath39 we see that @xmath44 is classical hochschild cohomology .
the fact that we are able to work over bi - modules , which are not necessarily symmetric is illustrated by example [ classical ] .
for a simplicial set @xmath45 and symmetric @xmath0 bi - module @xmath1 , we see that @xmath46 agrees with the definition of higher order hochschild cohomology given implicitly by pirashvili in @xcite .
as stated in section [ intro ] , one hope of defining higher order hochschild cohomology with multi - module coefficients is to discover additional connections to deformation theory . to begin , consider the following definition .
[ alg ] given a @xmath5-algebra @xmath0 and @xmath0-module @xmath47 we define a @xmath48 of @xmath0 by @xmath49 to be a short exact sequence @xmath50 which splits as @xmath5-vector spaces and with the properties that @xmath51 is a ring map and @xmath52 as an ideal of @xmath53 we see that defining a deformation of @xmath0 by @xmath49 amounts to defining a multiplication for @xmath54 given by @xmath55 but such a multiplication gives a map @xmath56 so that @xmath57 now , if we consider the trivial first order deformation of @xmath0 by @xmath0 i.e. @xmath58 we get a short exact sequence @xmath59/x^2 \xrightarrow{\psi } a \rightarrow 0,\ ] ] which induces a map of modules @xmath60/x^2 ) \rightarrow mod(a)\ ] ] which takes an @xmath61/x^2 $ ] module @xmath1 to @xmath62 this brings us to the following definition . given an @xmath0-module @xmath39
a first order deformation of @xmath1 is an @xmath61/x^2$]-module @xmath63 so that @xmath64 as stated in section [ intro ] , we get the following .
@xcite isomorphism classes of first order deformations of an @xmath0-module @xmath1 are in bijection with @xmath65 seeing the connections between classical hochschild cohomology and deformation theory , one can then ask the question .
are there similar connections to higher order hochschild cohomology when considering additional deformations ?
certainly we could consider modifying definition [ alg ] to let @xmath66 where @xmath67 instead of @xmath52 ( cubed zero deformations ) or let @xmath68/x^2,y^2 $ ] ( or any artinian algebra ) .
we could even consider deformations using steenrod relations .
this would , in principle give a means of computing all isomorphism classes of finitely generated modules over an artinian algebra .
current work of the author and salch seek to answer such questions and discover what structures exist on @xmath69
in order for the cosimplicial structure to exist , we simply need the cosimplicial identities to be satisfied , precisely : * @xmath70 for @xmath71 * @xmath72 for @xmath73 * @xmath74for@xmath75id@xmath76for@xmath77or@xmath78for@xmath79 with this in mind , we prove theorem [ mainthm ] : when composing the coface and codegeneracy maps from section [ intro ] it becomes clear what action identifications must be made in order for the cosimplicial structure to exist . for b ) notice when @xmath73 @xmath80 but this is precisely @xmath81 similarly we get @xmath82 since @xmath83 for all @xmath84 we get @xmath85 now for part a ) notice when @xmath71 @xmath86 @xmath87 and @xmath88 @xmath89 now , if there exists a @xmath90 so that @xmath91 then we must have that @xmath92 ( where @xmath35 is the element for the tensor factor associated to @xmath23 ) but this gives us that @xmath93 which implies @xmath94 so there is a set of actions that must be identified in order for a cosimplicial structure to exist .
the identification shown here is actually part i ) of theorem [ mainthm ] . following an analogous argument
it can be seen that parts ii ) , iii ) and iv ) are also consequences that come from ensuring @xmath70 for @xmath24 with that , part a ) of the cosimplicial identities is satisfied as long as these actions are identified since @xmath95 for all @xmath90 lastly , for part c ) we see that @xmath96 @xmath97 and : @xmath98 @xmath99 finally @xmath100 @xmath101 using the formulas above , we get the following identities on actions : * @xmath102 for @xmath103 , @xmath38 , @xmath104 , @xmath105 and the dimension of @xmath106 is at least 1 . * @xmath107 for @xmath108 , @xmath109 , @xmath104 , @xmath105 and the dimension of @xmath106 is at least 1 .
but v ) and vi ) are consequences of ii ) and iv ) from the list in theorem [ mainthm ] since @xmath106 must be of dimension larger than 1 in order to be a degenerate simplex and have @xmath105 .
notice that a @xmath110 action exists any time there is an @xmath111-simplex @xmath23 with the property that @xmath112 and @xmath38 for some @xmath113 we will refer to this as the @xmath114 action of @xmath23 and visually we can think of this action as being `` pointed '' towards the @xmath114 face of @xmath115 we now consider ways to visualize the identifications of the actions , presented in section [ proof ] . for iii )
( sweep across ) if two faces @xmath20 and @xmath106 of an @xmath111-simplex @xmath23 have a common face of @xmath116 then the action of @xmath20 which points towards @xmath117 is the same as the action of @xmath106 which points towards @xmath118 this is illustrated below with @xmath23 , a 2-simplex , @xmath119 the @xmath120 face and @xmath121 the @xmath122 face , while the 0-simplex labeled @xmath2 represents @xmath123 notice the @xmath120 face of @xmath106 is @xmath117 and the @xmath124 face of @xmath20 is @xmath123 this gives us that the associated actions , which point towards @xmath117 are identified .
we can see that in this particular instance , the action of @xmath106 points in the direction of the orientation of @xmath106 and the action of @xmath20 points against the orientation of @xmath125 for this reason , when dealing with simplices of dimension @xmath2 we will refer to the action as either being forward or backward . in this case the forward action of @xmath126 is identified with the backwards action of @xmath125 @xmath127 \matrix ( m ) [ matrix of math nodes , row sep=1em , column sep=1em , text height=1.5ex , text depth=0.25ex ] { 0 & & & & & & & & 1=\ast \\ & & \sigma & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ 2 & & & & & & & & \\ } ; \path[- > , thick , ] ( m-1 - 1 ) edge node[anchor = south ] { $ \omega$ } ( m-1 - 9 ) edge node[anchor = east ] { $ \gamma$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ \mu$ } ( m-6 - 1 ) ; \end{tikzpicture}\ ] ] for ii ) and iv ) ( sweep out 1 and sweep out 2 ) we see that if an @xmath111-simplex @xmath23 has an action which points towards a face @xmath128 then any other face of @xmath23 , which is not equal to @xmath117 has an equal action pointing in the corresponding direction , towards some face of @xmath129 this is illustrated below with @xmath130 a @xmath11-simplex whose first face is @xmath131 which is in fact @xmath118 we also have that @xmath20 is the @xmath120 face of @xmath23 and @xmath106 is the @xmath122 face of @xmath132 notice that the action of @xmath130 which points towards @xmath131 is the same as the backward action of @xmath106 and the forward action of @xmath125 @xmath127 \matrix ( m ) [ matrix of math nodes , row sep=1em , column sep=1em , text height=1.5ex , text depth=0.25ex ] { 0 & & & & & & & & 1 \\ & & \sigma & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ 2 & & & & & & & & \\ } ; \path[- > , thick , ] ( m-1 - 1 ) edge node[anchor = south ] { $ \omega$ } ( m-1 - 9 ) edge node[anchor = east ] { $ \ast = \gamma$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ \mu$ } ( m-6 - 1 ) ; \end{tikzpicture}\ ] ] lastly , for i ) ( sweep around ) a visualization is not completely necessary , as we see that the indication of i ) is that if @xmath23 is a simplex of dimension greater than or equal to 2 , then there is at most one action of @xmath23 . in other words ,
any two actions of @xmath23 are equal .
in this section , we will use the techniques discussed in section [ vis ] to show how one might determine the possible coefficient modules for a given simplicial set . before doing so
, we consider the following : when determining the possible coefficient modules , we can simply consider and identify the actions among non - degenerate simplices , since for a degenerate @xmath133-simplex @xmath134 if @xmath135 has an action @xmath136 then @xmath137 which gives that if @xmath71 we have that @xmath138 so @xmath38 and @xmath23 has action @xmath139 .
furthermore , by v ) we get @xmath140 similarly , if @xmath141 then @xmath23 has action @xmath142 which is equal to @xmath143 which gives us that no additional actions come from degenerate simplices . to see that new identifications do not take place , we notice if @xmath144 then @xmath145 and @xmath146 if @xmath71 this gives that @xmath147 however , any face of the degenerate @xmath133-simplex @xmath135 is degenerate via a face of @xmath23 and it can be checked that the actions identified above would also be identified with part iii ) using the faces of @xmath115 @xmath127 \matrix ( m ) [ matrix of math nodes , row sep=1em , column sep=1em , text height=1.5ex , text depth=0.25ex ] { * & & & & & & & & * \\ & & \sigma & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ & & & & & & \tau & & \\ * & & & & & & & & * \\ } ; \path[- > , thick , ] ( m-1 - 1 ) edge node[anchor = south ] { $ a$ } ( m-1 - 9 ) edge node[anchor = east ] { $ b$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ c$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ b$ } ( m-6 - 9 ) ( m-6 - 1 ) edge node[anchor = north ] { $ a$ } ( m-6 - 9 ) ; \end{tikzpicture}\ ] ] it is immediately evident from the picture above that there is a forward action of @xmath35 , which must agree with the backward action of @xmath148 , since they are both faces of @xmath149 which each have a face of @xmath118 the other implications coming from @xmath23 are that the forward action of @xmath148 is equal to the forward action of @xmath150 and the backward action of @xmath35 is equal to the backward action of @xmath151 similarly the backward action of @xmath148 must agree with the forward action of @xmath150 through @xmath152 and we should get that the forward action of @xmath35 is equal to the forward action of @xmath150 and the forward action of @xmath148 is equal to the backward action of @xmath35 . from the three @xmath2-simplices @xmath153 and @xmath148 it can be seen that we would start with 6 actions ( one for each direction ) , but through our identifications , we get that these all must be the same .
since no other simplices have a face of @xmath117 we get the following proposition .
the figure below is an illustration for the minimal simplicial decomposition of the pinched torus , where the @xmath124 face of @xmath23 and the @xmath124 face of @xmath155 are identified with @xmath117 @xmath127 \matrix ( m ) [ matrix of math nodes , row sep=1em , column sep=1em , text height=1.5ex , text depth=0.25ex ] { & & & & & & & & \\ & & \sigma & & & & & & \\ & & & & & & & & \\ & & & & & & & & \\ & & & & & & \tau & & \\ & & & & & & & & \\ } ; \path[- > , thick , ] ( m-1 - 1 ) edge node[anchor = south ] { $ a$ } ( m-1 - 9 ) edge node[anchor = east ] { $ \ast$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ c$ } ( m-6 - 1 ) ( m-1 - 9 ) edge node[anchor = west ] { $ \ast$ } ( m-6 - 9 ) ( m-6 - 1 ) edge node[anchor = north ] { $ a$ } ( m-6 - 9 ) ; \end{tikzpicture}\ ] ] here we see that the @xmath124 action of @xmath23 , the backward action of @xmath35 and the forward action of @xmath148 are all equal , while the @xmath124 action of @xmath155 , the forward action of @xmath35 and the backward action of @xmath148 are all equal .
no other identifications can be made , so we have the following proposition .
@xmath127 \matrix ( m ) [ matrix of math nodes , row sep=1em , column sep=1em , text height=1.5ex , text depth=0.25ex ] { * & & & & & & & & * \\ } ; \path[- > , thick , ] ( m-1 - 1 ) edge node[anchor = south ] { $ a$ } ( m-1 - 9 ) ; \end{tikzpicture}\ ] ] when @xmath111 is larger than @xmath2 we see that there is exactly one @xmath111-simplex , which has @xmath117 as every face , so by i ) , every arising action is identified , leaving us with the following proposition .
* acknowledgment .
* i would like to thank andrew salch for introducing me to hochschild cohomology and for his many hours of discussing the topics presented in this paper and the referee for his or her suggestions on how to improve the paper .
i would like to thank my wife , kendall for her continued support .
lastly , i would like to thank my newborn son , amos for allowing me to get the mathematics in this paper written down before joining the world . | when studying deformations of an @xmath0-module @xmath1 , laudal and yau showed that one can consider @xmath2-cocycles in the hochschild cohomology of @xmath0 with coefficients in the bi - module @xmath3 with this in mind , the use of higher order hochschild ( co)homology , presented by pirashvili and anderson , to study deformations seems only natural though the current definition allows only symmetric bi - module coefficients . in this paper
we present an extended definition for higher order hochschild cohomology which allows multi - module coefficients ( when the simplicial sets @xmath4 are accommodating ) which agrees with the current definition .
furthermore we determine the types of modules that can be used as coefficients for the hochschild cochain complexes based on the simplicial sets they are associated to . |
this is the second part of the study devoted to hypothesis testing problems in the case when the observations are inhomogeneous poisson processes .
the first part was concerned with the regular ( smooth ) case @xcite , while this second part deals with non regular ( singular ) situations .
we suppose that the intensity function of the observed inhomogeneous poisson process depends on the unknown parameter @xmath0 in a non regular way ( for example , the fisher information is infinite ) .
the basic hypothesis is always simple ( @xmath1 ) and the alternative is one - sided composite ( @xmath2 ) . in the first part we studied the asymptotic behavior of the score function test ( sft ) , of the general likelihood ratio test ( glrt ) , of the wald test ( wt ) and of two bayes tests ( bt1 and bt2 ) .
it was shown that the tests sft , glrt and wt are locally asymptotically uniformly most powerful . in the present work we study the asymptotic behavior of the glrt , wt , bt1 and bt2 in two non regular situations .
more precisely , we study the tests when the intensity functions has a cusp - type singularity or a jump - type singularity . in both cases
the fisher information is infinite .
the local alternatives are obtained by the reparametrization @xmath3 , @xmath4 .
the rate of convergence @xmath5 depends on the type of singularity . in the cusp case @xmath6 , where @xmath7 is the order of the cusp , and in the discontinuous case @xmath8 .
our goal is to describe the choice of the thresholds and the behavior of the power functions as @xmath9 .
the important difference between regular and singular cases is the absence of the criteria of optimality .
this leads to a situation when the comparison of the power functions can be only done numerically .
that is why we present the results of numerical simulations of the limit power functions and the comparison of them with the power functions with small and large volumes of observations ( small and large @xmath10 ) .
recall that @xmath11 is an inhomogeneous poisson process with intensity function @xmath12 , @xmath13 , if @xmath14 and the increments of @xmath15 on disjoint intervals are independent and distributed according to the poisson law @xmath16 we suppose that the intensity function depends on some one - dimensional parameter , that is , @xmath17 .
the basic hypothesis is simple : @xmath1 , while the alternative is one - sided composite : @xmath2 .
the hypothesis testing problems for inhomogeneous poisson processes were studied by many authors ( see , for example , @xcite , @xcite , @xcite and the references therein ) .
we consider the model of @xmath10 independent observations of an inhomogeneous poisson process : @xmath18 , where @xmath19 , @xmath20 , are poisson processes with @xmath21 we use here the same notations as in @xcite . in particular , @xmath0 is one - dimensional parameter and @xmath22 is the mathematical expectation in the case when the true value is @xmath23 .
the intensity function is supposed to be separated from zero on @xmath24 $ ] .
the measures corresponding to poisson processes with different values of @xmath0 are equivalent and the likelihood function is defined by the equality @xmath25{\rm d}t\right\}.\ ] ] in non - regular situations we do not have a laump test , and it is interesting to compare the power functions of different tests with the power function of the neyman - pearson test ( n - pt ) .
let us recall the definition of the n - pt .
suppose that we have two simple hypotheses @xmath26 and @xmath27 and our goal is to construct a test @xmath28 of size @xmath29 , that is , a test with given probability of the error of the first kind @xmath30 .
as usually , the test @xmath28 is the probability to reject the hypothesis @xmath31 and , of course , to accept the hypothesis @xmath32 .
let us denote the likelihood ratio statistic as @xmath33 then , by the neyman - pearson lemma @xcite , the n - pt is @xmath34 where the constants @xmath35 and @xmath36 are solutions of the equation @xmath37 in this work we consider the construction of the tests in the following hypothesis testing problem @xmath38 that is , we have a simple hypothesis against one - sided composite alternative .
the log likelihood ratio function can be written as follows : @xmath39{\rm d}t.\ ] ] the power function of a test @xmath40 is @xmath41 , @xmath2 .
we denote @xmath42 the class of tests @xmath43 of asymptotic size @xmath44 : @xmath45 in this work we study several tests which belong to the class @xmath46 . to compare these tests by their power functions we consider , as usual , the approach of _ close _ or _ contiguous _ alternatives ( since for any fixed alternative the power functions of all tests converge to the same value @xmath47 ) .
we put @xmath48 , where @xmath49 . here
@xmath50 and the rate of convergence depends on the type of singularity of the intensity function .
now the initial problem of hypothesis testing can be rewritten as follows : @xmath51 the considered tests are usually of the form @xmath52 where the constant @xmath53 is defined with the help of the limit random variable @xmath54 ( suppose that @xmath55 under hypothesis @xmath56 ) by @xmath57 if the limit random variable @xmath54 is continuous , and by @xmath58 if @xmath54 has distribution function with jumps
. the corresponding power function will be denoted @xmath59 and the comparison of the tests will be carried in terms of their limit power functions .
we consider two different non regular models . in both of them
, the intensity function @xmath60 is not differentiable and the fisher information is infinite .
more precisely , we study the behavior of the tests in two situations . the first one is when the intensity function has a cusp - type singularity ( it is continuous but not differentiable ) , and the second one is when it has a jump - type singularity ( it is discontinuous ) . in both cases the intensity functions @xmath61 has no derivative at the point @xmath62 .
note that these statistical models were already studied before in the problems of parameter estimation ( see @xcite for the cusp - type singularity and @xcite for discontinuous intensity function ) , so here we concentrate on the properties of the tests .
the main tool is , of course , the limit behavior of the normalized likelihood ratio function , which was already established before in the mentioned works but in a slightly different settings .
the proofs given in this work are mainly based on the results presented in @xcite and @xcite .
recall that in the non regular cases considered in this work we do not have a laump tests , and that is why a special attention is paid to numerical simulations of the limit power functions .
suppose that the intensity function of the observed poison processes is @xmath63 where @xmath64 , @xmath65 , @xmath66 , @xmath67 , and @xmath68 is a known positive bounded function . to study the local alternatives we introduce the normalizing function @xmath69,\ ] ] where @xmath70 is the _ beta - function _ and @xmath71 is the _
hurst parameter_. as usually , the change of variables @xmath72 reduces the initial hypothesis testing problem to the problem .
we introduce the stochastic process @xmath73 where @xmath74 is a _ fractional brownian motion_. further , we define the random variable @xmath75 by the relation @xmath76 and we introduce @xmath77 and @xmath78 as the solutions of the equations @xmath79 respectively .
note that @xmath80 is the likelihood ratio of a similar hypothesis testing problem ( @xmath81 against @xmath82 ) in the case of observations @xmath83 of the following type @xmath84 the uniformly most powerful test in this problem does not exist , and we do not have a laump tests in our problem .
the glrt is defined by the relations @xmath85 where @xmath77 is the solution of the first of the equations , @xmath86 and @xmath87 is the maximum likelihood estimator ( mle ) .
let us introduce the function @xmath88 > \ln
h_\varepsilon -\frac{\left|u_*\right|^{2h}}{2}\right\},\quad u_*\geq 0 .\ ] ] the properties of the glrt are given in the following proposition .
the glrt @xmath89 belongs to @xmath90 and its power function in the case of local alternatives @xmath91 , @xmath92 , has the following limit : @xmath93 introduce the normalized likelihood ratio process @xmath94 and let the function @xmath95 be linearly decreasing to zero on the interval @xmath96 $ ] and equal to @xmath97 for all @xmath98 . now the random function @xmath99 is defined on @xmath100 .
let us fix some @xmath101 and denote @xmath102 the space of continuous functions on @xmath103 with the property @xmath104 . introduce the uniform metric on this space and denote @xmath105 the corresponding borel @xmath106-algebra .
when we study the likelihood ratio process under hypothesis @xmath107 , we take @xmath108 and consider the corresponding measurable space @xmath109 . under the alternative @xmath110 , @xmath92
, we will use this space with @xmath111 .
let @xmath112 be the measure induced on the measurable space @xmath109 by the stochastic processes @xmath80 , and @xmath113 be the measure induced ( under the true value @xmath23 ) on the same space by the processes @xmath114 .
the continuity with probability @xmath47 of the random functions @xmath114 follows from the inequality below and the kolmogorov theorem .
suppose that we already proved the following weak convergence @xmath115 then the distribution of any continuous in the uniform metric functional @xmath116 converge to the distribution of @xmath117 .
in particular , if we take @xmath118 we obtain @xmath119 therefore the test @xmath120 .
let us note , that we do not know an analytic solution of the equation defining the constant @xmath121 , that is why below we turn to numerical simulations ( see section [ ssc ] ) .
note also that @xmath122 and does not depend on @xmath123 . to study the power function we consider the same likelihood ratio process but under the alternative @xmath124 .
we can write @xmath125 with an obvious notation .
the difference between @xmath126 and @xmath127 is that the `` reference value '' in the first case is fixed ( is equal to @xmath128 ) and in the second case it is `` moving '' ( is equal to @xmath129 ) .
the random variable @xmath130 converge in distribution to @xmath131 . for the stochastic process
@xmath132 we have a similar convergence , and so , for any fixed @xmath133 , we have @xmath134 now , let @xmath135 be the measure induced on the measurable space @xmath136 by the stochastic processes @xmath137 , and @xmath138 be the measure induced ( under the true value @xmath23 ) on the same space by the stochastic processes @xmath139 .
suppose that we already proved the weak convergence @xmath140 then for the power function we can write @xmath141 > \ln h_\varepsilon \!\right\}\\ & \quad={\mathop{\mathbf{\kern 0pt p}}\nolimits}\left\{\sup_{u > 0 } \left[w^h\left(u\right)-\frac{\left|u - u _ * \right|^{2h}}{2}\right ] > \ln
h_\varepsilon -\frac{\left|u_*\right|^{2h}}{2}\right\}= \hat\beta \left(u_*\right ) . \ ] ] this limit power function is obtained below with the help of numerical simulations ( see section [ ssc ] ) .
let us also note that the limit ( under the alternative @xmath142 ) of the likelihood ratio process @xmath143 is the process @xmath144 defined by @xmath145 to finish the proof we need to verify the convergence . to do this we follow the proof of the convergence given in @xcite .
we introduce the following relations . 1 .
_ the finite - dimensional distributions of @xmath146 converge to those of @xmath147 .
_ there exists a positive constant @xmath148 such that _ @xmath149 3 .
_ there exists a positive constant @xmath150 such that _ @xmath151
the proofs of these relations are slight modifications of the proofs given in @xcite .
note that the characteristic function of the vector @xmath152 can be written explicitly and the convergence of this characteristic function to the corresponding limit characteristic function can be checked directly ( see lemma 5 of @xcite ) .
the inequalities and follow from the lemma 6 and lemma 7 of @xcite respectively .
these relations allow us to obtain the weak convergence by applying the theorem 1.10.1 of @xcite .
note that the convergence is a particular case of with @xmath153 .
recall that the mle @xmath87 is defined by the equation @xmath154 the wald test ( wt ) has the following form : @xmath155 where @xmath156 is the solution of the second of the equations .
introduce as well the random variable @xmath157 as solution of the equation @xmath158 the wt @xmath159 belongs to @xmath90 and its power function in the case of local alternatives @xmath160 , @xmath92 , has the following limit : @xmath161 the mle ( under hypothesis @xmath107 ) converges in distribution @xmath162 hence @xmath163 . for the proof
see @xcite .
recall that this convergence is a consequence of the weak convergence .
let us study this estimator under the alternative @xmath164 , @xmath92 .
we have @xmath165 here , as before , @xmath166 now , the limit of the power function of the wt is deduced from this convergence : @xmath167 which concludes the proof .
let us note , that we can also give another representation of the limit power function using the process
@xmath144 : @xmath168 where @xmath169 is solution of the equation @xmath170 the threshold @xmath78 and the power function @xmath171 are obtained below by numerical simulations ( see section [ ssc ] ) .
suppose that the parameter @xmath0 is a random variable with _ a priori _
density @xmath172 , @xmath173 .
this function is supposed to be continuous and positive .
we consider two bayes tests .
the first one is based on the bayes estimator , while the second one is based on the averaged likelihood ratio .
the first test , which we call bt1 , is similar to wt , but is based on the bayes estimator ( be ) rather than on the mle .
suppose that the loss function is quadratic .
then the be @xmath174 is given by the following conditional expectation : @xmath175 we introduce the test bt1 as @xmath176 where the constant @xmath177 is solution of the equation @xmath178 introduce as well the function @xmath179 the bt1 @xmath180 belongs to @xmath90 and its power function in the case of local alternatives @xmath160 , @xmath92 , has the following limit : @xmath181 the bayes estimator @xmath182 is consistent and has the following limit distribution ( under hypothesis @xmath107 ) @xmath183 ( for the proof see @xcite ) .
hence @xmath184 .
for the power function we have @xmath185 let us study the normalized difference @xmath186 .
we can write ( using the change of variables @xmath187 ) @xmath188 hence @xmath189 ( since @xmath190 and @xmath191 ) .
the detailed proof is based on the properties 13 of the likelihood ratio ( see @xcite or ( * ? ? ? * theorem 1.10.2 ) ) .
let us note , that we can also give another representation of the limit power function using the process @xmath144 : @xmath192 where @xmath193 . the second test , which we call bt2 , is given by @xmath194 here @xmath195 and @xmath196 is solution of the equation @xmath197 introduce as well the function @xmath198 the bt2 @xmath199 belongs to @xmath46 and its power function in the case of local alternatives @xmath200 , @xmath92 , has the following limit : @xmath201 let us first recall how this test was obtained . introduce the mean error of the second kind @xmath202 under alternative @xmath32 of an arbitrary test @xmath203 as @xmath204 where @xmath205 is the double mathematical expectation , that is , the expectation with respect to the measure @xmath206 if we consider the problem of the minimization of this mean error , we reduce the initial hypothesis testing problem to the problem of testing of two simple hypotheses @xmath207 then , by the neyman - pearson lemma , the most powerful test in the class @xmath90 @xmath208which minimizes the mean error @xmath209 is @xmath210 where the averaged likelihood ratio @xmath211 and @xmath212 is chosen from the condition @xmath213
now , it is clear that the bt2 @xmath214 coincides with the test @xmath215 if we put @xmath216 . in the proof of the convergence in distribution of the bayes estimator it is shown ( see ( * ? ? ?
* theorem 1.10.2 ) and @xcite ) that @xmath217 therefore ( under hypothesis @xmath107 ) , @xmath218 and the test @xmath219 belongs to the class @xmath90 . using a similar argument
, we can verify the convergence @xmath220 under the alternative @xmath221 , which concludes the proof .
let us consider the following example .
we observe @xmath10 independent realizations @xmath18 , where @xmath222\right)$ ] , @xmath223 , of an inhomogeneous poisson process .
the intensity function of this processes is @xmath224 where the parameter @xmath225 .
we take @xmath226 as the value of the basic hypothesis @xmath107 .
of course it is sufficient to have simulations for the values @xmath227 , but we consider a wider interval to show the behavior of the likelihood ratio on the both sides of the true value .
the hurst parameter is @xmath228 and the constant @xmath229 \approx 1.027 $ ] .
a realization of the normalized likelihood ratio @xmath95 , @xmath230 $ ] , and its zoom @xmath95 , @xmath231 $ ] , under the hypothesis @xmath56 are given in figure [ z_n_u_cusp ] .
to find the thresholds of the glrt and of the wt , we need to find the point of maximum and the maximal value of this function . in the case of the chosen intensity function ,
the maximum is attained at one of the cusps of the likelihood ratio ( that is , on one of the events of one of the observed poisson processes ) .
it is interesting to note that if the intensity function has the same singularity but with a different sign : @xmath232 , then it is much more difficult to find the maximum ( see figure [ z_n_u_cusp_invers ] ) .
the thresholds of the glrt , of the wt and of the bt1 are presented in table [ thr_cusp ] .
.[thr_cusp]thresholds of glrt , wt and bt1 [ cols="^,^,^,^,^,^,^,^",options="header " , ] it is interesting to compare the studied tests with the neyman - pearson test ( n - pt ) corresponding to a fixed value @xmath233 of @xmath234 .
of course , it is impossible to use this n - pt in our initial problem , since @xmath233 ( the value of @xmath234 under alternative ) is unknown . nevertheless , its power ( as function of @xmath233 ) shows an upper bound for power functions of all the tests , and the distances between it and the power functions of studied tests provide an important information .
let us fix some value @xmath92 and introduce the n - pt @xmath235 where @xmath236 and @xmath237 are solutions of the equation @xmath238 denoting @xmath239 , we can rewrite this equation as @xmath240 here @xmath241 is a poisson random variable with parameter @xmath233 , and so the quantities @xmath242 and @xmath36 can be computed numerically .
a similar calculation yields the limit power of the n - pt : @xmath243 where @xmath244 is a poisson random variable with parameter @xmath245 .
the results of simulations are presented in figure [ pf_dis_rho_3 ] for two cases : @xmath246 and @xmath247 . in both cases
the limit power function of the glrt is the closest one to the limit power of the n - pt , and the limit power function of the bt1 arrives faster to @xmath47 than the others .
this study was partially supported by russian science foundation ( research project no .
14 - 49 - 00079 ) .
the authors thank the referee for helpful comments . | we consider the problem of hypothesis testing in the situation where the first hypothesis is simple and the second one is local one - sided composite .
we describe the choice of the thresholds and the power functions of different tests when the intensity function of the observed inhomogeneous poisson process has two different types of singularity : cusp and discontinuity .
the asymptotic results are illustrated by numerical simulations .
msc 2010 classification : 62m02 , 62f03 , 62f05 .
_ key words : _ hypothesis testing , inhomogeneous poisson processes , asymptotic theory , composite alternatives , singular situations . |
the dwarf spheroidal ( dsph ) satellites of the galaxy provide an important test of dark matter dynamics .
simulations of heirarchical merging indicate that dwarf galaxies were the first structures to form in the early universe , and the merging and accretion of these objects created the galaxies and clusters of galaxies which exist today .
it is the collapse of dark matter which has driven the development of large - scale structure .
observations indicate that the local dsph galaxies are potent repositories of dark matter ( see @xcite and references therein ) where the luminous material resides at the centre of a large dark halo . however , these measurements of mass - to - light ratio ( @xmath5 ) assume the system is in virial equilibrium .
@xcite proposed that the interaction of a satellite within the tidal field of its host galaxy may result in an artificially inflated estimate of the virial mass .
however , the simulations of @xcite imply that the central mass - to - light ratio is not substantially affected by tidal forces .
in contrast , @xcite and @xcite have created models of satellite systems _ without _ dark matter yet with apparently similar characteristics to those observed in the galactic satellites .
thus , tidal disruption by the milky way may be producing the appearance of dark matter in dsph galaxies .
the subsequent analysis of the draco dsph by @xcite did not detect the predicted line - of - sight depth in this system , however a quantitative measurement of the gravitational influence of the galaxy on its satellites is yet to be achieved .
consequently , there have been attempts to detect evidence of interactions between large galaxies and their satellites in the nearby universe .
perhaps the best example of this process is the sagittarius dwarf galaxy ( @xcite ; @xcite and references therein ) , a satellite currently being torn apart by the gravitational potential of the galaxy .
the simulations of @xcite have reproduced the general appearance of this disrupted system , and they find it is on a highly eccentric polar orbit ( @xmath6 ) .
however , the structure of a satellite during the initial interaction with its host is still not understood .
numerical simulations by @xcite , @xcite and @xcite predict that during the early phase of interaction with the galaxy , energy is injected into the outer regions of the satellite , resulting in a slightly inflated structure .
also , depending on the level of interaction , the internal structure may be distorted .
an example of this initial stage may be the ursa minor dsph , which has a structure that displays significant asymmetries and extra - tidal stars @xcite .
the sculptor dsph ( scl ) lies at a galactocentric distance of 80 kpc @xcite .
previous studies have attempted to detect the initial stage of galaxy - satellite interaction by examining the radial profile of the sculptor system .
for example , wide - field photometric surveys in a single colour @xcite have identified possible substructure and found evidence for extra - tidal stars in this dsph . indeed ,
@xcite claim the possible detection of tidal tails aligned with the scl major axis .
however , the effect of foreground and background sources becomes important in the outer regions of this system due to the relatively small number of scl stars compared to foreground galactic stars .
as such , the existing radial profiles are highly uncertain in the outer regions , and it has therefore been difficult to accurately measure any possible inflation of the dsph structure . without a clear understanding of the level of milky way - scl interaction , and hence whether this system is in virial equilibrium , it is not possible to accurately constrain the amount of dark matter in scl using traditional velocity dispersion measurements . radial velocity studies by @xcite and @xcite of k giants found a central mass - to - light ratio of @xmath7 .
if the system is undergoing substantial distortion due to the galactic tidal field , then this mass - to - light ratio may be overestimated .
we have searched the sculptor dsph for structural and kinematic distortion .
firstly , we collected wide - field photometry in two colours covering a 10 deg@xmath3 area centred on scl , well beyond the nominal tidal radius of this galaxy .
the colour - magnitude diagram ( cmd ) obtained from the inner parts of the system was used to select candidate scl stars over the dataset area , thereby increasing their contrast relative to field sources in the outer parts . for examples of this process ,
see the analyses by @xcite , @xcite , @xcite and @xcite . in the second stage , we have obtained spectra for @xmath8700 of the brightest candidate members .
this allowed a further refinement of the dataset by selecting stars based on radial velocity and metal abundances .
this paper describes our search for substructure in sculptor based on the combined photometric / spectroscopic dataset .
a companion paper ( @xcite ; hereafter c05 ) presents the results for a similar wide - field photometric survey of fornax .
the observational strategy for this system followed that of the fornax dsph ( c05 ) .
images of the sculptor region were obtained at the siding spring observatory 1 metre telescope equipped with the wide field imager ( wfi ) .
sixteen fields were observed in two filters ( @xmath9 and @xmath10 ) , thereby covering a @xmath11 area on the sky using a @xmath12 mosaic of wfi fields .
similar to the fornax observations , accurate photometry was required to magnitudes of @xmath0 and @xmath1 , encompassing the range of the sculptor red giant branch ( rgb ) .
a full list of observations , including the atmospheric seeing , is given in table [ sculptorobservations ] .
the data were reduced using the iraf program , and the stellar magnitudes were measured using the daophot program within iraf .
an astrometric calibration of this dataset was made using the ucac1 catalogue @xcite .
a full description of the data reduction and photometric calibration techniques can be found in c05 .
multiply - detected stars in the @xmath9 and @xmath10 maps were removed by matching sources in the overlap regions with a search radius of @xmath13 .
the same search algorithm was used to combine the @xmath9 and @xmath10 datasets , thus producing a @xmath14 colour for all stars in common between the two maps .
[ sclcmd_cand ] shows the resulting cmd from the region within @xmath15 of the centre of sculptor . to measure the photometric completeness of the survey , a luminosity function was generated for each of the sixteen sculptor fields in both @xmath9 and @xmath10 bands
we then assumed the completeness limit of each field to be 0.2 mag brighter than the turnover in the luminosity function . to ensure this method
provided an accurate measurement of the completeness limit , we created a dataset of @xmath85000 artificial stars with a distribution resembling the scl rgb .
these stars were placed in each field , and then recovered using the photometry routines described above .
it was found that @xmath16 to @xmath17 of the artificial stars were recovered from each field to their respective photometry limits .
thus , the luminosity function allowed an adequate measurement of the photometric limit for each field .
the overall magnitude limit of the survey is then defined by the field with the least depth : field scl5 which is complete to @xmath1 and @xmath0 .
the 2df instrument at the prime focus of the anglo - australian telescope ( aat ) allows the simultaneous spectroscopy of up to 400 targets over a field of view @xmath18 in diameter .
a full description of 2df is given by @xcite .
our aim was to collect spectra at 850 nm for candidate scl stars , and to use the ca ii triplet spectral features to the separate scl members ( the ` signal ' ) from foreground / background sources ( the ` noise ' ) .
we chose 1408 candidate scl stars from the photometric survey ( described in the previous section ) using the cmd selection range displayed in fig .
[ sclcmd_cand ] .
spectra were obtained for 893 of these candidates , however only 764 yielded spectra with the required signal - to - noise . the majority of the stars for which spectra were not obtained lie close to , or within , three core radii , where the limit on fibre - to - fibre spacing of 2df restricts the object selection .
for the 527 stars in the sample beyond the tidal radius , however , @xmath19 have acceptable spectra .
the faint limit of this spectroscopic survey was set to be @xmath20 ( approximately @xmath21 at the colour of the rgb , well above the photometric completeness limit given in the previous section ) thus sampling the upper 2.0 mag of the scl rgb .
the spatial distribution of these stars is shown in fig .
[ sclxy_cand ] , where the five dashed circles outline the observed 2df fields .
note that we have observed potential member stars well beyond the scl nominal tidal radius .
approximately 100 stars located in the overlap regions between 2df fields were observed more than once to ensure we obtained consistent velocities and ca ii line strengths .
the selected stars were observed with the 1200r gratings , and the resulting spectra were centred near the ca ii ir triplet with a wavelength range of @xmath22 and a scale of 1.07 per pixel .
although the red spectral region is dominated by sky emission lines , we reduced such contamination using the slow - beam switching method , which intersperses object exposures with offset sky frames containing the same fibre configuration ; that is , object and sky were observed through the same fibre .
in addition , each configuration included 42 fibres assigned to sky positions .
hence , the observation sequence for each scl field consisted of @xmath23min exposures and three offset sky images .
each sequence also included three wavelength calibration frames ( arcs ) comprising four cu - ar and two cu - he lamps with an exposure time of 30s .
an arc frame at the start , middle and end of the sequence accounted for possible instrumental wavelength shifts . approximately @xmath24 hours were required to completely observe each field .
a list of observations is given in table [ 2dfobs ] .
we also obtained spectra for 16 and 13 red giants in the globular clusters ngc 1904 and ngc 2298 respectively to be used as radial velocity standard stars .
additionally , the metallicities of these clusters ( ngc 1904 , [ fe / h]@xmath25 ; ngc 2298 , [ fe / h]@xmath26 ; from @xcite ) are approximately equal to that of scl ( @xmath27 [ fe / h ] @xmath28 ; @xcite ) . for old , metal poor stars ( such as those in scl and the globular clusters )
the individual equivalent widths of the calcium lines can be related to the metal abundance , and we can use the globular cluster observations to calibrate this relation .
thus , additional to selecting stars based on radial velocity , the ca ii line strengths provided a second constraint to select members of the scl system .
the spectra were reduced using the 2dfdr program in an interactive mode . in the 2df setup
, 400 spectra are divided between two ccds , including 21 sky fibres per ccd .
as a first step in the data reduction process , a tram - line map was generated for both ccds .
this map tracked the output spectra on each ccd , and was obtained from the fibre flat field images due to their high s / n .
the tram line map was then shifted and rotated to fit each individual data frame and the signal for each fibre extracted .
the arc frames then provide a wavelength calibration for each fibre by matching the pixel position of the observed lines with the spectral line list .
an accurate sky subtraction is dependent on a reliable fibre throughput calibration , since the relative throughputs of individual fibres can vary by @xmath8@xmath29 .
the offset sky exposures were used to measure the relative fibre throughput .
these were taken with the same configuration as the object field , and multiple offset sky exposures at dithered positions ensured each fibre was accurately sampling the sky .
the offset sky frames were median combined to ensure the sky spectrum was not accidentally contaminated by sources or cosmic rays .
the scl frames were then reduced and fit - extracted , and the combined sky spectrum ( from the 21 sky fibres ) , scaled by the individual fibre throughputs , was subtracted to remove the night - sky features .
the six exposures for each field were then median - combined to produce the final set of spectra . due to
less than favourable observing conditions ( seeing @xmath30 ) , only 14 of the 29 globular cluster ( gc ) stars in the observed configurations had spectra with sufficient signal - to - noise ratio to be used as radial velocity standard stars .
six of these were recorded on ccd 1 and eight on ccd 2 . to ensure these spectra were adequate templates
, we used them to independently measure the mean systematic velocity of the two clusters , which were then compared to the known values .
the stellar radial velocities were measured using the _ rvidlines _ routine in iraf .
each velocity was calculated as the mean value determined from the three ca ii lines .
these were corrected for the motion of the earth and we removed the ccd - dependence of these velocities using the @xmath31 values determined below .
thus , we calculated a mean radial velocity of @xmath32 km s@xmath33 for ngc 1904 ( from six stars ) and @xmath34 km s@xmath33 for ngc 2298 ( from seven stars , after removing a star whose radial velocity of 71 km s@xmath33 indicated it is not a cluster member ) where the errors are the standard deviation of the mean . these compare well to the values listed in the online globular cluster catalogue ( @xcite ; @xmath35 and @xmath36 km s@xmath33 respectively ) . in order to use these spectra as radial velocity standard spectra , we first needed to measure the velocity offset between each gc star .
for each ccd , the star with the highest s / n spectrum was defined to be the master star , and the remaining gc spectra were cross - correlated with the master spectrum to measure velocity offsets .
this resulted in six ( eight ) template spectra on ccd 1 ( ccd 2 ) with a known velocity offset from the master star .
radial velocities for the candidate scl stars were then measured by cross - correlating the spectrum with all templates from the appropriate ccd , and shifting the resulting velocity by the appropriate offset such that they were measured relative to the master star for that ccd .
the spectra were cross - correlated in the wavelength range @xmath37 , which encompasses the ca ii triplet lines ( @xmath388498 , @xmath388542 and @xmath388662 ) and reduces the number of strong sky features which may not have been completely removed from the stellar spectra ( see fig . 1 in @xcite ) .
this was accomplished using the _ fxcor _ routine in the iraf radial velocity package .
consequently , each star whose spectrum was recorded on ccd 1 ( ccd 2 ) was cross - correlated with six ( eight ) gc spectra , resulting in multiple radial velocity measurements .
since the velocities of the candidate scl stars were measured relative to the two master stars , an adjustment was required to transform these to heliocentric velocities .
the heliocentric velocities of both master stars were previously measured using the _ rvidlines _ routine as described above . also , we calculated the heliocentric conversion for the scl frames .
the difference between these values was added to the velocities of the candidate sculptor stars , resulting in a set of radial velocity measurements relative to the sun .
we combined these values to give a mean radial velocity for each potetial scl member , and the velocity uncertainty was calculated as the standard deviation from the mean .
a final velocity zeropoint shift was achieved by comparing our measured radial velocities with those in a previous catalogue , which includes the observations of @xcite .
high resolution spectra for 38 stars were recorded over three seasons ( @xmath39 ) at the ctio 4 metre telescope , where the stars were selected from the @xcite and @xcite photographic surveys of scl .
@xcite described the data reduction and velocity measurement techniques and estimated the radial velocities had a mean error of @xmath84 km s@xmath33 .
the current survey included 27 stars in common with this catalogue , where 14 spectra were recorded on ccd 1 and 13 on ccd 2 .
from these , the mean velocity difference between this program and the @xcite catalogue was found to be , @xmath40 where the errors are the standard deviation of the mean .
these @xmath31 values were added to the measured velocities to produce a final radial velocity for 764 stars . to determine the mean uncertainty of the scl radial velocities
, we compared the measurements for multiply observed stars . during the 2df observing run
, multiple observations were obtained of 120 stars located in the overlap regions between 2df fields .
of these stars , 92 yielded multiple radial velocities with an uncertainty less than 20 km s@xmath33 , and these cover the full magnitude range of the survey .
approximately 60 of these were observed on both ccds , and over half were obtained with significantly different fibre numbers .
this allowed cross - ccd comparisons between spectra recorded at different positions in each ccd .
the ccd velocity corrections listed above were applied to these values , and a comparison of these multiple observations yielded a mean velocity difference of @xmath41 km s@xmath33 , with a standard deviation of @xmath42 km s@xmath33 .
thus , dividing this value by @xmath43 , the scl velocities have a mean uncertainty of approximately 10 km s@xmath33 .
some of the measured velocities contained a large uncertainty , hence those with a velocity error greater than 20 km s@xmath33 were removed .
this criterion removed 41 stars , resulting in a final catalogue containing 723 candidate scl stars .
the majority of the rejected stars were located towards the faint end of the 2df survey limit , where the mean velocity uncertainty was found to be @xmath812 km s@xmath33 , as compared with @xmath82 km s@xmath33 at the bright end of the survey .
[ sclxy_cand ] shows the spatial distribution of the final catalogue 723 stars .
the lines comprising the ca ii triplet are the strongest features in the far red component of a stellar spectrum , and are also situated close to the peak of the spectral energy distribution displayed by red giant stars .
therefore , the pseudo - equivalent widths , or line strengths , of these three lines ( @xmath44 , @xmath45 and @xmath46 ) in rgb stars can be easily measured by medium - resolution spectroscopy . a detailed description of the calcium triplet lines and their relation to red giant metallicities is given by @xcite .
@xcite first noted that if @xmath47 is plotted against @xmath48 ( where @xmath49 is the magnitude of the horizontal branch ) for several stars in numerous globular clusters , then the resulting slopes are independent of cluster abundance .
this relation has been further investigated by @xcite , @xcite , @xcite , @xcite and @xcite .
we refer the reader to fig . 2 in @xcite
, whose study of several galactic globular clusters demonstrated that the best fitting line for the @xmath50 relation has a slope of @xmath51 / mag .
consequently , this linear relationship can be used to remove the dependence of line strength on surface gravity by defining the reduced equivalent width to be @xmath52 .
an important feature of @xmath53 is that it can be directly related to the iron abundance @xcite .
@xcite analysed several galactic globular clusters with a large abundance range and demonstrated that the @xmath53-[fe / h ] relation can be split into two linear regimes .
this relation was further investigated by @xcite , who found a similar result ( albeit with a slightly different definition of the reduced equivalent width ) using data for 52 galactic globular clusters .
therefore , the strengths of the ca ii lines provide an independent measure of metallicity for old , metal - poor stars . in this survey , equivalent widths of the ca ii lines at @xmath388452 and @xmath388662 were measured using the method described by @xcite .
two windows to either side of each feature were used to trace the continuum at the feature wavelength .
the equivalent widths were then measured using a gaussian fitting technique .
the same rest wavelength windows given in table 2 of @xcite were used to fit the gaussian functions and continuum line , however for the blue continuum bandpass of the ca ii @xmath54 line we used the wavelength range @xmath55 .
the weakest of the ca ii triplet lines , centred at a rest wavelength of @xmath388498 , generally does not have sufficient s / n in our spectra , especially at low metallicities , for accurate measurement of the equivalent width .
hence this line was not included in the following line strength analysis .
[ ewgc](a ) shows the sum of the two strongest ca ii line strengths as a function of magnitude difference from the horizontal branch for the two globular clusters ngc 1904 ( closed circles ) and ngc 2298 ( open circles ) .
we have discounted one of the stars in the ngc 2298 dataset , since its velocity and equivalent width data indicate it is not a member of the cluster .
also , two of the stars have been removed from the set of ngc 1904 points , as their @xmath56 values contained large uncertainties and differed significantly from the remaining four stars .
the dashed lines in fig .
[ ewgc](a ) trace the best linear fit to the data with a slope of @xmath51 / mag .
this slope provided an adequate linear fit to the data points for both clusters .
to ensure these results are consistent with previous observations , we calculated the mean reduced equivalent width ( @xmath57 ) for both clusters .
@xcite list values for the two clusters which can be easily modified to the current system . from the transformation derived by @xcite
, we find their @xmath57 values for the clusters ngc 1904 and 2298 to be @xmath58 and @xmath59 respectively .
we calculated @xmath53 for each star using the data shown in the top panel of fig .
[ ewgc ] , and found the following mean reduced equivalent width for each cluster , @xmath60 where the uncertainties are the standard deviation of the mean .
these are consistent with the @xcite values .
a second verification of these reduced equivalent widths was to ensure they satisfy the empirical @xmath57-[fe / h ] relation derived by @xcite using the globular cluster metallicities provided by @xcite .
the dashed line in fig .
[ ewgc](b ) illustrates this relationship for metal - poor clusters , and the two data points represent the @xmath57 values listed above .
the [ fe / h ] values are from @xcite , to ensure consistency with the @xcite study .
the datapoints are adequately described by the abundance relation , hence the mean reduced equivalent widths derived for these two globular clusters are consistent with previous observations .
the overall uncertainty of the line strength measurements was obtained by comparing 71 multiply observed stars in the overlap regions between 2df fields that have the same approximate luminosity distribution as the total set of scl candidates . for the @xmath45 and @xmath46 lines
, we measured the standard deviation from the mean to be 0.31 and 0.38 respectively .
the total uncertainty for @xmath56 was measured to be @xmath61 .
this uncertainty is dominated by instrumental and sky noise , and is only weakly dependent on magnitude .
the early photographic studies of @xcite established many of the fundamental structural parameters of scl , and did not detect any significant deviations from a smooth stellar density profile .
however , later analyses @xcite found that the region within the core radius is almost circular , yet the ellipticity increases with radius . these analyses also detected possible extra - tidal stars .
more recently , @xcite conducted a structural analysis of scl based on a ` blue ' ( iiiaj ) photographic plate , and found that its structure closely resembles the numerical simulations of tidally distorted galaxies by @xcite .
irwin & hatzidimitriou also found some evidence for extra - tidal stars .
the ccd survey of @xcite detected possible tidal tails extending along the major axis of scl in the east - west direction and concluded that the outer structure of scl is strongly influenced by the galactic tidal field .
consequently , to further address these claims , we have investigated the structure of scl using a variety of methods applied to the photometric dataset .
the cmd provides a filter to select candidate scl stars , thereby decreasing the number of field stars . in a similar photometric survey of the fornax dwarf spheroidal galaxy
, we described a simple colour - magnitude selection technique which decreased contamination due to field stars ( c05 ) .
for the scl survey we applied a more complex cmd - selection process based on that of @xcite , which has since been adapted by @xcite .
this method required the cmd to be divided into an array of cells , and provided a ` signal ' value of scl stars throughout colour - magnitude space . those cells with a signal greater than the limiting value
were then used to select candidate scl stars , maximising the ratio of scl to field stars in the outer regions of the dwarf galaxy .
a full description can be found in @xcite , however we include a summary of the process below , detailing how it was applied to our data . to determine the signal of scl stars over the cmd
, we divided the photometric data into two spatial regions ; the ` core ' region ( within three core radii of the scl centre ) , which is dominated by scl stars ; and the field star region ( located outside the tidal radius and within the survey limit ) .
a grid was overlaid on the cmd , subdividing it into a series of cells with indices @xmath62 .
each cell had a width of 0.05 mag in colour and a height of 0.10 mag . to minimise the effects of photometry errors
, we only considered the magnitude range @xmath63 , where the mean colour uncertainty is less than 0.03 mag . for each colour - magnitude cell @xmath62 , we counted the number of stars located in the core region @xmath64 , and the number in the field region @xmath65 .
thus , the signal - to - noise ratio of core to field stars for each cmd cell is , @xmath66 where @xmath67 is the ratio of the core area to the field star area . in effect , the array @xmath68 listed which regions of the cmd contained a high signal - to - noise of sculptor stars . following the method of @xcite , we defined a limiting signal value @xmath69 to optimise the cumulative signal - to - noise ratio of core stars to field stars .
stars were then selected in those cmd cells with a signal greater than a limiting signal value : @xmath70 . a contour plot showing the signal @xmath68 over colour - magnitude space is given in fig . [ sclcmdlines ] .
the outer contour outlines those cells with a signal greater than @xmath69 .
this region encompasses the scl rgb down to the photometric limit , and hence was used to select candidate rgb stars .
thus , we selected 4355 candidate scl rgb stars , and the spatial distribution is displayed in fig .
[ sclrgbxy ] , where the ellipses represent the core and tidal radii of the rgb population ( see table [ kingpar ] ) . at first glance
, there is no obvious substructure in the distribution of scl rgb stars .
the radial profile of this stellar population is displayed in fig .
[ sclradial ] . the best - fitting @xcite model to all datapoints
is represented as a dashed line , from which we derived a nominal tidal radius of @xmath71 arcmin , which is comparable to the @xcite value of @xmath72 arcmin . however , there is some deviation from this curve beyond @xmath73 .
consequently , we fitted a king profile to the inner @xmath74 , which is represented by a dotted line .
this profile had a limiting radius of @xmath75 arcmin , and the data shown in fig .
[ sclradial ] implies significant ` extra - tidal ' structure .
this is consistent with the results of @xcite who found a tidal radius of @xmath76 and claimed a large population of stars beyond this radius .
we have found that an over - subtraction of the field population in our data can reproduce the results of @xcite , giving a detection of extra - tidal structure .
it is clear that neither model provides an adequate description of the sculptor system , however the king parameters for both models are listed in table [ kingpar ] , where the entries for ` rgb - a ' refer to the profile for the inner @xmath74 of scl . in this context , it is worth considering the results of @xcite , who found that the red giants of sculptor apparently consist of two populations which are kinematically and chemically distinct .
these two populations are probably related to the fact that the red and blue hb stars in sculptor have different radial distributions with the red hb stars being more centrally concentrated @xcite .
section 4.3 contains a full analysis of the blue and red hb star distributions , from which we have derived their spatial extents ; the king parameters for both populations are given in table [ kingpar ] .
we have used these distributions to examine the overall density profile of the rgb stars , finding that a two - component fit gives an excellent representation of the data .
this is shown as the dashed line in fig .
[ tworadial ] , where the dotted lines represent the two components .
these two components have the core and limiting radii listed for the blue and red hb populations respectively .
also , to reduce the number of free parameters , the ratio of the central densities of these two components is the same as the blue hb / red hb ratio , taken from table [ kingpar ] .
the resulting model provides a significantly better fit to the data than a single component model . in the companion paper ( c05 )
we described two methods to detect structure beyond the nominal tidal radius of a stellar system .
the first of these , a density probability function , was created by measuring the stellar density in a large number of randomly placed circles over the survey area . in this way
, we were able to find the probability of measuring a given density of stars _ outside _ the nominal tidal radius .
a uniform distribution of stars ( for example , the field star population ) is expected to display a gaussian density function , while substructure will produce a non - gaussian function .
this function was measured for the distribution of rgb - selected stars outside the nominal tidal radius of scl , where the stellar density was measured within @xmath77 circles of radius @xmath78 .
the circle size was chosen to be large enough to include a reasonable number of stars , yet not too large to miss the detection of medium - scale ( @xmath820 arcmin ) structures .
[ mc_xy ] shows the density function , where the dashed line represents the best - fitting gaussian to the data .
this function indicates the mean background stellar density is 237 stars / deg@xmath3 with a standard deviation of 65 . unlike the corresponding result for fornax ( c05 )
, the extra - tidal region of scl does not display any significant substructure ; the function appears to be consistent with a uniform distribution of sources , which implies a population comprising only field stars .
if the stars beyond the tidal radius are not part of the scl system , then they must be foreground and background sources .
the background source population is made up of distant galaxies , however the majority of these would not be interpreted as stellar sources by the daophot program due to their non - circular morphology .
although a complete star - galaxy separation can not be made given an average seeing of @xmath79 for the survey , we estimate that the total number of background galaxies to be less than @xmath80 of the number of objects detected . also , the cmd selection range shown in fig .
[ sclcmdlines ] further precludes this population .
hence , assuming the extra - tidal population of scl is negligible , the field population is almost entirely assembled from foreground galactic stars .
@xcite estimated the density of the field star population towards sculptor using the galactic models of @xcite .
in the colour - magnitude range @xmath81 ,
@xmath82 , they estimated 252 stars / deg@xmath3 in the scl field , with an approximate error of @xmath80 .
although these colour and magnitude limits do not exactly correspond with our cmd selection range ( fig .
[ sclcmdlines ] ) , the @xcite estimate agrees ( within the uncertainties ) with our mean field star density measurement of 237 stars / deg@xmath3 represented in fig .
[ mc_xy ] .
this supports our inference that the candidate - rgb stars beyond the nominal tidal radius correspond to the field star population , and there is no substantive evidence for scl rgb stars in this region . to determine how effectively the density probability function can detect structures in an otherwise random distribution of stars
, we simulated a series of overdensities in the extra - tidal region of scl .
the monte carlo algorithm was applied to a dataset consisting of the candidate scl rgb stars and two synthetic tidal tails extending in the north and south directions from the centre of the system .
we created six artifical datasets , where the mean stellar density of the synthetic tidal tails was @xmath83 above that of the field population .
the tails uniformly occupied the region @xmath84 over the total declination range of the survey , both inside and outside the nominal tidal radius .
for example , a spatial plot of the rgb dataset with artificial tidal tails of density @xmath85 can be seen in fig .
[ sclrgbtails ] , which ( from the aggregate stellar luminosity over this sky area ) have a surface brightness of 30.4 mag / arcsec@xmath3 above that of the field population .
[ mc_xy_tot ] shows the density functions measured from the six tidal tail simulations . for each panel
, the dashed line traces the density function measured from the true extra - tidal population .
the @xmath86 overdensity population is barely visible in the density function , however a second peak becomes more prominent as the density of the synthetic tidal tails increases . for tidal tails of this type
, we find evidence for substructure when the density is a factor of @xmath8@xmath87 above the background population .
this corresponds to @xmath8100 extra - tidal rgb stars per square degree , or approximately @xmath29 of the light from all sculptor rgb stars .
the simulations were repeated using tidal tails of various widths ( greater than @xmath8@xmath88 ) aligned at several position angles relative to sculptor , and we found that the density probability function was able to detect features with a surface brightness greater than @xmath89 mag / arcsec@xmath3 above the background population , assuming they have the same luminosity function as scl . a second method to search for extra - tidal structure is to measure the angular correlation function for the stellar population beyond the nominal tidal radius .
the angular correlation function is generally used in cosmology to examine if a sample of galaxies are clustered , or alternatively , follow a poissonian distribution @xcite . for a given sample of data points covering a finite area , this function
is measured by placing a large number of random points in a two - dimensional area with the same shape as the dataset .
the number of data - data pairs ( @xmath90 ) , data - random pairs ( @xmath91 ) and random - random pairs ( @xmath92 ) is then determined at a given scale length @xmath93 . the observed angular correlation function
is then calculated as , @xmath94 the parameters @xmath95 and @xmath96 are determined from the number of data and random points ( @xmath97 , @xmath98 ) using the relations , @xmath99 this provides a solution for a survey of infinite extent , however for a finite survey the function must be adjusted by a constant called the integral constraint , defined to be the product @xmath100 @xcite .
the parameter @xmath101 is determined numerically from the number of random - random pairs , and the values of @xmath102 and @xmath103 ( defined below ) are determined as those which provide the best - fitting function to the observed datapoints . after correcting for the integral constraint , the angular correlation function takes the form , @xmath104 where @xmath102 is the function amplitude , and @xmath103 defines the slope .
a purely random distribution of points will produce a flat angular correlation function ( @xmath102 and @xmath103 are both equal to zero , hence @xmath105 ) , while a sloped function ( @xmath102 and @xmath103 are greater than zero ) reflects a clustered sample .
accordingly , we measured the angular correlation function for all stars beyond the nominal tidal radius of scl .
we placed @xmath106 points randomly throughout the survey region and measured the distance between all data - data , data - random and random - random pairs .
these were placed in logarithmic bins of width @xmath107 , where @xmath93 is in arcmin .
[ corrfn ] shows the data points for the observed angular correlation function , calculated using eqn .
[ angcorobs ] .
the errorbars were calculated from the poisson noise of the number of data points in each bin of @xmath93 .
the dashed line defines @xmath108 , which is the expected function for a randomly distributed set of stars .
this line accurately traces the data points , especially at large scales where a large number of data - data pairs in each bin results in small uncertainties .
hence , this test indicates that there is no significant structure in the distribution of rgb - candidate stars beyond the nominal tidal radius of scl . to test the effectiveness of this function
, we repeated the process for the six synthetic tidal tail datasets described in the previous section .
[ acftest ] displays the resulting angular correlation functions , where the parameter at the top right corner of each panel is the density of the tidal tails in terms of the standard deviation of the field population ( @xmath109 ) .
the errorbars represent poisson noise and the dashed line traces @xmath108 , the expected function for a normal distribution of points .
similar to the density function above , we have found an overdensity of @xmath8@xmath87 is easily discernible , while structures with a density less than this value are difficult to distinguish from the background noise .
hence , the angular correlation function was able to detect extra - tidal structures with a surface brightness greater than @xmath11031 mag / arcsec@xmath3 .
that is , in order for an extra - tidal structure to have been detected around scl , it required a surface brightness approximately @xmath111 greater than the background .
both tests indicated there was no substructure in the extra - tidal distribution of stars shown in fig .
[ sclrgbxy ] .
to further investigate the possibility of extra - tidal structure , we created a contour map of scl ( fig .
[ sclcontour ] ) from the rgb - selected stars .
each star was convolved with a gaussian of radius @xmath112 , and the contour smoothing length is @xmath113 .
the outer contour traces all structure with a density @xmath114 above the field population and the second contour level represents a @xmath115 density , where @xmath116 stars / deg@xmath3 is the standard deviation of the field population .
the first contour is highly sensitive to statistical variations in the field star density , hence we define the @xmath115 level as the significance limit of this survey .
the red ellipse in fig .
[ sclcontour ] represents the scl tidal radius .
there are some signs of structure in the outer regions of sculptor .
for instance , the two small @xmath115 density peaks located in the south - west quadrant . @xcite
measured the absolute proper motion of scl to be directed towards the ne , thus it is possible that the south - west ` blobs ' represent tidally stripped stars trailing the sculptor system .
it is also possible , however , that these structures are simply density spikes in the field star population , and are therefore not associated with sculptor .
either way , they represent only a small fraction of the stars in this system .
a second feature is the small ` hook ' structure centred at approximately @xmath117 , @xmath118 , within the tidal radius .
there is some evidence that this structure also exists in the @xcite study , however only at the @xmath119 level in their diagram .
it points towards the south - west quadrant , and possibly to the small @xmath115 density peaks mentioned above .
this structure may represent tidal distortion , or it may also be a random fluctuation in the field star density .
we investigated whether the structure of scl rgb stars is dependent on radius by fitting a series of ellipses to the stellar density , following the method described in c05 .
the best - fitting centre , position angle and ellipticity were calculated at each major axis radius value of @xmath120 arcmin . beyond @xmath121 , the number of stars was too low to accurately fit an ellipse to the data .
[ lscontour ] shows the mean values in radial bins of @xmath122 , where the uncertainties represent the standard deviation from the mean and the dashed lines mark the central coordinates , ellipticity and position angle listed by @xcite .
the upper two plots demonstrate that the centre of scl remains constant out to a radius of approximately @xmath74 , and then moves north for all subsequent radii .
although the position angle of these ellipses effectively remains constant at all radii , the ellipticity plot in fig .
[ lscontour](d ) indicates that the structure of the inner regions is almost circular , and then becomes more elliptical as the radius increases ( see fig .
[ sclcontour ] ) .
this reinforces the results of @xcite , @xcite and @xcite .
hence , there is some evidence for scl distortion . a visual examination of the contour diagram in fig .
[ sclcontour ] does indicate that the outer region has a significantly higher ellipticity compared to the inner region .
@xcite reported the possibility of tidal extensions along the east - west axis ( see their fig .
although @xcite state a limiting magnitude for their survey to be @xmath123 , they also point out that the claim of tidal extensions is subject to non - uniform limiting magnitudes throughout the survey , and is also dependent on the level of star / galaxy separation .
we confirm that the structure of scl appears stretched along this axis in the outer regions , however there is little evidence to support the existence of well - defined tidal tails .
@xcite proposed that the interaction between a satellite system and the host galaxy may artificially increase the velocity dispersion , thereby increasing the measured mass - to - light ratio .
however , in order to increase the radial velocity dispersion the satellite galaxy must be positioned such that its long axis is aligned along the line of sight .
@xcite stated that the resulting distance distribution of the stars should produce a horizontal branch width of @xmath81 mag .
@xcite previously searched for this line - of - sight depth in the draco dsph , without success . in scl
we find the horizontal branch vertical dispersion to be approximately 0.15 mag which does not support a line - of - sight inflation of this system .
for the remainder of this section , we investigate the morphology of the scl horizontal branch in relation to the second parameter effect .
the horizontal branch ( hb ) constitutes core he - burning stars with ages greater than @xmath810 gyr .
studies of globular clusters indicate that metallicity is the first parameter governing the hb morphology ; metal - poor populations tend towards a blue horizontal branch ( bhb ) , while metal - rich populations often form the red horizontal branch ( rhb ) . however , some systems exhibit a second parameter ( 2p ) effect , where the hb morphology is not wholly dependent on metallicity
. the early photometric analyses of scl revealed a surprisingly red hb given its mean abundance @xcite , implying scl is a 2p object .
more recently , deeper photometry over a wide field has revealed that the hb morphology in scl is radially dependent @xcite , such that the rhb stars are more centrally concentrated than the bhb stars .
this has been verified by @xcite .
the results of @xcite and @xcite indicate that the stellar population of scl is bimodal ; that is , the inner region displays the 2p effect and contains a mixture of rhb and bhb stars , while the outer region is dominated by bhb stars .
the spectroscopic study by @xcite showed that sculptor contains two distinct stellar populations represented by either side of the hb , and that these populations are offset from each other in metallicity and kinematics .
the population gradient in scl may indicate that residual gas from the first star formation burst was concentrated at the centre of the system , resulting in younger and/or more metal rich central stellar population ( for a full discussion , see @xcite ) . to investigate the radial dependence of the hb morphology , we selected stars using the colour - magnitude ranges outlined in fig .
[ sclcmd_cand ] . the selection range for both components of the horizontal branch
was chosen to minimise contamination effects due to stars in the rgb and instability strip .
first we examine the red hb stars .
the dashed line represents the completeness limit ( @xmath124 and @xmath125 ) of the four inner fields .
hence , the corresponding spatial distribution of rhb stars displayed in fig . [ sclrhbxy ] is essentially complete for all values of @xmath126 and @xmath127 between @xmath128 and @xmath129 deg .
a radial profile of the rhb candidate stars was assembled by selecting stars within a set of annuli of constant centre , ellipticity and position angle ( listed by @xcite ) .
the background stellar density was calculated as the mean value beyond @xmath130 ( table [ kingpar ] ) .
this value was subtracted from the stellar density measurements , and the resulting radial profile is shown in fig .
[ radialhb](a ) , where the dashed line represents the best - fitting king profile . the resulting core and limiting radii
are represented by the solid ellipses in the rhb - selected spatial map ( fig .
[ sclrhbxy ] ) , while the dashed line represents the limiting radius of the scl rgb stars ( see table [ kingpar ] ) .
now we turn to the blue hb stars . the bhb selection box shown in fig .
[ sclcmd_cand ] lies beyond the photometric limit of the four inner fields .
hence , to determine which fields contain a near - complete sample at this luminosity , we placed @xmath85000 artificial bhb stars in each field and attempted to recover these using the same photometry routines described in [ photsurvey ] above .
the artifical stars were clustered around @xmath131 , @xmath132 with a dispersion of @xmath133 magnitudes in both @xmath9 and colour , resembling the appearance of the bhb as seen in fig .
[ sclcmd_cand ] .
the recovery statistics for each field are listed in table [ bhbstats ] . for each artificial star , we measured the difference between the input and output magnitudes .
the values @xmath134 and @xmath135 are the standard deviations from the mean difference , and hence indicate the level of photometric accuracy achieved by each field at the depth of the bhb .
an examination of table [ bhbstats ] reveals that seven of the sixteen fields recovered significantly less than @xmath136 of the artificial bhb stars .
therefore , bhb stars were selected only in the remaining nine fields , and their spatial distribution is shown in fig .
[ sclbhbxy ] .
the dotted line outlines the seven incomplete fields , which are also labelled .
we constrained the best - fitting bhb radial profile from the nine complete fields using the same technique described for the rhb population above .
[ radialhb](b ) shows the bhb profile , where the errors are the poission noise for each annulus and the dashed line represents the best - fitting king model .
we subtracted a background density from the profile , measured to be @xmath137 stars / arcmin@xmath3 in the area beyond a major axis radius of @xmath138 .
the resulting king profile parameters are listed in table [ kingpar ] , while the core and limiting radii are displayed in the bhb - selected spatial map ( fig .
[ sclbhbxy ] ) .
a comparison of the radial profiles for the red and blue horizontal branch stars is shown in fig .
[ radialhb](c ) . we have thus confirmed the hb gradient in scl , and measured the spatial extent of the two populations described by @xcite .
the scl spectroscopic dataset contained 723 candidate rgb stars with radial velocity errors less than 20 km s@xmath33 .
the velocity distribution of these stars is shown in fig .
[ sclspecv ] .
the population is clearly bimodal in velocity , with the scl population clustered around @xmath139 km s@xmath33 and the field population around @xmath80 km s@xmath33 . to determine the mean scl radial velocity , we selected 188 stars in the radial velocity range 76.5 km s@xmath33 @xmath140 km s@xmath33 , and calculated the mean velocity using the relation from @xcite , @xmath141 where the weight ( @xmath142 ) was defined to be the inverse square of the individual velocity uncertainty for each velocity @xmath143 .
the mean heliocentric velocity of scl was then measured to be , @xmath144 where the error is the weighted standard deviation of the mean .
this value is marked as the dashed line in fig .
[ sclspecv ] , and agrees well with the previous measurements by @xcite and @xcite . in this section
we describe several methods used to eliminate foreground galactic stars from the scl member list .
the primary method was provided by a radial velocity selection range .
we measured the velocity dispersion around the scl mean to be @xmath145 km s@xmath33 . assuming the average global velocity dispersion of the scl system is approximately 10 km s@xmath33 @xcite
, then this @xmath146 value agrees with the overall velocity uncertainty of @xmath810 km s@xmath33 .
this allows us to define the velocity selection range for the scl members . the dotted lines in fig .
[ sclspecv ] illustrate the velocity range for the selection of scl stars .
we have chosen a width of @xmath147 to either side of the mean scl velocity ( @xmath148 km s@xmath33 ) .
this velocity range gave 157 likely members of the scl system .
an important concern in this selection criterion is residual background noise from galactic stars ; fig .
[ sclspecv ] shows the distribution in velocity of these foreground stars included in the scl observations , and clearly some portion of the galaxy s population extends into the scl velocity range . the colour selection range for these stars illustrated in fig .
[ sclcmd_cand ] allows two possible populations of stars to contribute to the field population ; halo rgb stars or galactic ( thin and/or thick ) disk main sequence stars . assuming the field population comprises giant stars
, then the apparent magnitude selection range implies a distance of at least 45 kpc ( from the yonsei - yale isochrones ; @xcite ) .
the stellar halo contains @xmath8@xmath149 of the stellar mass of the galaxy , and has a power density distribution @xmath150 where @xmath151 depending on the stellar population @xcite , thus we expect the contribution of halo stars to the field population to be negligible . in contrast , if we assume the field population comprises dwarf stars occupying the galactic disk , then the colour - magnitude selection range imply these stars are located at a distance of @xmath152 pc .
now , the thin disk displays a scale height of @xmath8300 pc ( for example , @xcite ) , whereas the scale height of the thick disk @xcite is approximately 1000 pc ( @xcite ; see @xcite for a recent review ) .
therefore , we expect a negligible contribution from the thin disk , and hence the field population composed almost entirely of galactic thick disk stars . the scl dsph is located near to the south galactic pole ( sgp ) , hence the width in velocity space of the field stars is determined almost exclusively by the vertical velocity dispersion of the galactic thick disk .
recent measurements @xcite have found this value to be @xmath153 km s@xmath33 .
we calculated the _ observed _ velocity dispersion of the field stars towards scl to be 34.1 km s@xmath33 .
assuming an instrumental uncertainty of 10 km s@xmath33 gave the true velocity dispersion of these stars to be 32.6 km s@xmath33 which agrees with previously determined values .
there were 364 stars in the velocity range @xmath154 km s@xmath33 , hence we deduce the dataset includes approximately 430 galactic stars overall . assuming the velocity distribution follows the gaussian function illustrated in fig .
[ sclspecv](b ) , then the @xmath147 star selection criterion is expected to include approximately eight high - velocity galactic thick disk stars .
a second test for the membership status of the candidate scl stars was provided by abundance measurements . from the psuedo - equivalent widths we were able to calculate @xmath56 for the 157 candidate scl stars , which can be directly related to [ fe / h ] for red giants .
given that any residual field objects would be foreground dwarf stars , we expect their surface gravities may provide a ca triplet measurement markedly different from the range seen in sculptor .
the resulting @xmath56 values are plotted against magnitude in fig .
[ eqwidthobj ] . based on ca ii triplet observations of approximately 300 rgb stars in scl
, @xcite observed an abundance range in scl of @xmath155 [ fe / h ] @xmath156 .
the upper abundance limit is marked as the dashed line in fig .
[ eqwidthobj ] .
however , the line strength index has an uncertainty of @xmath61 , therefore we extended the @xmath53 selection limit by a factor of @xmath157 , and this limit is marked by the solid line in fig .
[ eqwidthobj ] . in order for a star to be included as a scl member
, its line strength must lie below this value .
the open circles represent the two stars which do not fulfill this criterion and have been discarded as scl members .
@xcite revealed that scl contains a metallicity gradient , such that the mean [ fe / h ] value of the inner region is higher than that of the outer region .
we found a similar result in our data .
[ radmet ] shows the metallicity of these stars as a function of ( major axis ) radius , under the assumption that scl has an ellipticity of @xmath158 @xcite .
the two dashed lines mark the core and nominal tidal radii .
we subdivided these data at the core radius ( @xmath159 ) , and found the mean metallicity for both datasets to be , @xmath160 } \rangle = \left \ { \begin{array}{lll } -1.56 \pm 0.04 & \mbox{~if ~ } r \le r_c ; & n=8 \\
-1.92 \pm 0.03 & \mbox{~if ~ } r_c < r < r_t ; & n=135 \\
\end{array } \right.\ ] ] where the errors are the standard deviation of the mean .
that is , we have confirmed the metallicity gradient noted by @xcite .
scl is located towards the sgp , hence we expect many of the field stars to display a high proper motion . proper motions of all stars in the refined scl member list were extracted from the supercosmos science archive , which makes use of observations at up to four different epochs . the astrometry from this catalogue
is described by @xcite .
the proper motions of the potential members are shown in fig .
[ propmot ] .
we measured the standard deviations from the mean motions in ra and dec to be @xmath161 milliarcsec / year and @xmath162 milliarcsec / year respectively .
thus , the proper motion limits for scl stars were set to be @xmath163 and @xmath164 , and this is marked as the dashed ellipse in fig .
[ propmot ] .
the four stars with a proper motion ( within their uncertainties ) beyond this value were discarded as scl members and are represented as open circles .
the sigma - clipping processes above removed six contaminating objects , two based on calcium triplet measurements and four on proper motion .
this reduced the number of possible scl members to 151 stars , of which ten are located outside the nominal tidal radius .
three of these extra - tidal candidates have relatively large velocity errors , which makes it difficult to be certain of membership , and colours that place them @xmath80.05 mag to the red of the the definite scl rgb members in the cmd ( see fig . [ sclspeccmd ] ) . for these reasons
they were discarded from the scl members sample .
consequently , we have seven candidate extra - tidal members . the characteristics of these stars are listed in tables [ xtidalcandsphot ] and [ xtidalcandsspec ] .
[ sclspecxy ] shows the spatial distribution of the 148 scl members , where the closed triangles represent the seven extra - tidal stars .
note that those stars within the tidal radius uncertainty of this limit are not considered to be extra - tidal . to aid the reader in matching the data in tables [ xtidalcandsphot ] and [ xtidalcandsspec ] with the data points in fig .
[ sclspecxy ] , we have also listed their @xmath126 and @xmath127 coordinates , and their distance from the centre of scl , @xmath165 .
the colour - magnitude distribution of the final member star list is shown in fig .
[ sclspeccmd ] , where the asterisks represent the three stars rejected based on colours .
the parameters we have measured for all sculptor members are given in table [ sclmembers ] .
@xcite found that the ellipticity of scl increased with radius , which has since been confirmed in subsequent photographic plate analyses @xcite .
we also noted this effect . in our analysis ,
the position angle of these ellipses did not significantly change with radius ( @xmath166 ) . a possible indication of tidal distortion was the fluctuation of the central coordinates of scl .
that is , as the ellipse radius increased , the centre shifted by approximately @xmath167 north . as such , the internal structure of scl displays what may be the hallmarks of tidal distortion .
however , it is worth noting that a triaxial structure can also produce these chracteristics .
@xcite described a feature extending south from the eastern lobe of scl ( see their fig . 2 ) .
this feature was located near their tidal limit of the system ( @xmath168 ) and they stated that it may indicate a tidal extension .
we find no evidence of this feature in the current study . also , @xcite measured the tidal radius of sculptor to be @xmath76 ( significantly less than previous surveys ) , with clear signs of tidal structure .
in contrast , we have fitted a two - component king model to sculptor with a limiting radius of @xmath169 arcmin , and found that it adequately describes the system with little sign of tidal inflation .
this reinforces the @xcite result , who found that sculptor contains two separate stellar populations .
tidal distortion may also produce kinematic substructure ( such as apparent rotation ) in a satellite system .
we have identified 148 scl members through a spectroscopic analysis , and the spatial distribution of these is shown in fig .
[ sclspecxy ] .
an examination of the velocities of these stars as a function of position angle revealed no significant evidence for rotation , confirming the original result by @xcite .
@xcite conducted a radial velocity analysis of the umi dsph and discovered a kinematically cold population of stars towards the centre of the system , which they hypothesise may represent an underlying clump of dark matter .
to investigate the presence of kinematic substructure within scl , we measured the offset in velocity of each star from the mean velocity of scl , and examined this offset as a function of position .
there were no obvious kinematic substructures , however the small number of stars in the outer regions of scl for which we have radial velocities precludes a strong result , as do the relatively large velocity errors .
our analysis of the photometric survey set the upper limit for extra - tidal material to be approximately @xmath29 of the sculptor mass . with the spectroscopic survey , we were able to strengthen this result . by selecting stars
based on radial velocity , metallicity and proper motion , we found seven extra - tidal candidates out of a total member list of 148 objects .
these seven stars may include some residual galactic thick disk stars , thus they correspond to an upper limit for the total mass outside the scl tidal radius .
however , the spectroscopic survey was not complete : we obtained radial velocities for 723 of the 1408 cmd - selected stars .
thus , by examining the number of candidates surveyed beyond the tidal radius , we find that the seven extra - tidal candidates correspond to @xmath170 stars in a complete survey of the rgb candidates .
similarly , the density of the field stars in this cmd selection range is @xmath8@xmath171 stars / deg@xmath3 , from which we infer that approximately 600 of the 1408 stars within the tidal radius are sculptor members .
therefore , assuming the extra - tidal candidates represent a stellar population with the same luminosity function as scl , then we have measured the upper limit of extra - tidal luminous mass to be @xmath4 of the scl total , where the uncertainty represents poisson noise .
this result can be compared to the previous study by @xcite , who analysed 602 variable stars ( mostly rr lyrae type ) in the scl region using the catalogue of @xcite , described to be @xmath172 complete . from an analysis of the stellar positions ( see their fig .
2 ) we find @xmath8@xmath173 of these are located beyond the nominal tidal radius of the rgb population determined in the current study , @xmath174 .
that is , only @xmath175 of these stars are extra - tidal .
this result agrees with our upper mass limit for extra - tidal structure stated above .
although there is little evidence of significant extra - tidal structure in our spectroscopic sample , it is worth refering to the work of @xcite , who measured the absolute proper motion of scl to be towards the north - east .
four of the seven extra - tidal stars we have identified are located in this quadrant .
if these stars are members of sculptor , then they would represent leading tidal material .
it is difficult , however , to make any definite statement in this regard based on such a small dataset .
also , there is no correlating north - east population of extra - tidal stars in the @xcite analysis .
we have conducted a combined photometric and spectroscopic survey of the scl dsph to determine if its structure and kinematics have been influenced by the tidal field of the galaxy . in the first phase of the survey
, we obtained photometry in two colours ( @xmath9 and @xmath10 ) over a 10 deg@xmath3 area centred on scl .
the dataset was complete to a magnitude of @xmath0 , encompassing the scl red giant branch .
stars were selected in the rgb colour - magnitude range , thereby increasing the contrast of scl stars to the foreground galactic population .
the sculptor structure appears to follow a two - component model , where each component is represented by the blue and red horizontal branch stars respectively .
although there was no obvious extra - tidal structure , we determined that up to @xmath29 of the light from scl could be outside the tidal radius and still evade photometric detection .
hence , to further probe the scl extra - tidal region , we obtained far red spectra for over 700 stars located in the upper - rgb region of the scl cmd .
stars were selected as scl members based on three criteria : radial velocity , ca ii triplet strength , and proper motion .
the final scl member list contained 148 stars , with seven located outside the tidal radius .
thus , if scl does possess extra - tidal structure , then it makes up less than @xmath4 of the luminous mass in this system .
therefore , we have found little sign of tidal interaction between sculptor and the galaxy .
this lack of tidal interaction makes it highly likely that velocity dispersion studies do provide a true measure of the dark matter mass of this dsph .
the authors would like to thank brian schmidt for providing the astrometric calibration program and laura dunn for contributing stellar population synthesis models .
mc also thanks paul allen for his helpful discussions on the angular correlation function , and john norris for advice concerning the properties of the galactic thick disk .
the authors acknowledge the 2df observing assistance provided by terry bridges , and the help of jeremy bailey in reducing these data at the anglo - australian observatory .
mc acknowledges the financial support provided by an australian postgraduate award .
the authors also thank the anonymous referee for a detailed report which led to an improved paper .
this research has been supported in part by funding from the australian research council under discovery project grant dp034350 .
this research has also made use of data obtained from the supercosmos science archive , prepared and hosted by the wide field astronomy unit , institute for astronomy , university of edinburgh , which is funded by the uk particle physics and astronomy research council .
armandroff , t. e. & da costa , g. s. 1986 , , 92 , 777 armandroff , t. e. & da costa , g. s. 1991 , , 101 , 1329 bahcall , j. n. & soneira , r. m. 1980 , , 44 , 73 bahcall , j. n. & soneira , r. m. 1984 , , 55 , 67 chen , b. , et al .
2001 , , 553 , 184 chiba , m. & beers , t. c. 2000 , , 119 , 2843 cole , a. a. , smecker - hane , t. a. , tolstoy , e. , bosler , t. l. , & gallagher , j. s. 2004 , , 347 , 367 coleman , m. g. , da costa , g. s. , bland - hawthorn , j. , & freeman , k. c. 2005 , , 129 , 1443 ( c05 ) da costa , g. s. 1984 , , 285 , 483 da costa , g. s. 1988 , iau symp . 126 : the harlow - shapley symposium on globular cluster systems in galaxies , 126 , 217 da costa , g. s. & armandroff , t. e. 1995 , , 109 , 2533 da costa , g. s. , armandroff , t. e. , caldwell , n. , & seitzer , p. 1996
, , 112 , 2576 demers , s. , krautter , a. , & kunkel , w. e. 1980 , , 85 , 1587 eskridge , p. b. 1988a , , 95 , 1706 eskridge , p. b. 1988b , , 96 , 1336 gilmore , g. 1984 , , 207 , 223 gilmore , g. & reid , n. 1983 , , 202 , 1025 grillmair , c. j. , freeman , k. c. , irwin , m. , & quinn , p. j. 1995 , , 109 , 2553 hambly , n. c. , davenhall , a. c. , irwin , m. j. , & macgillivray , h. t. 2001 , , 326 , 1315 harbeck , d. , et al .
2001 , , 122 , 3092 harris , w. e. 1996 , , 112 , 1487 helmi , a. & white , s. d. m. 2001 , mnras , 323 , 529 hodge , p. w. 1961 , , 66 , 384 hodge , p. w. 1965 , , 142 , 1390 hodge , p. w. 1966 , , 71 , 204 hurley - keller , d. , mateo , m. , & grebel , e. k. 1999 , , 523 , l25 ibata , r. a. , gilmore , g. , & irwin , m. j. 1994 , nature , 370 , 194 innanen , k. a. & papp , k. a. 1979 , , 84 , 601 irwin , m. & hatzidimitriou , d. 1995 , , 277 , 1354 johnston , k. v. , choi , p. i. , & guhathakurta , p. 2002
, , 124 , 127 kaluzny , j. , kubiak , m. , szymanski , m. , udalski , a. , krzeminski , w. , & mateo , m. 1995 , , 112 , 407 kim , y. , demarque , p. , yi , s. k. , & alexander , d. r. 2002 , , 143 , 499 king , i. r. 1966 , , 71 , 64 klessen , r. s. , grebel , e. k. , & harbeck , d. 2003 , , 589 , 798 klessen , r. s. & kroupa , p. 1998
, , 498 , 143 kleyna , j. t. , wilkinson , m. i. , gilmore , g. , & evans , n. w. 2003 , apj , 588 , l21 koch , a. , grebel , e. k. , odenkirchen , m. , martnez - delgado , d. , & caldwell , j. a. r. 2004 , , 128 , 2274 kroupa , p. 1997
, new astronomy , 2 , 139 kuhn , j. r. & miller , r. h. 1989 , , 341 , l41 kunkel , w. e. & demers , s. 1977 , , 214 , 21 law , d. r. , johnston , k. v. , & majewski , s. r. 2004 , apj , submitted ( astro - ph/0407566 ) layden , a. c. , hanson , r. b. , hawley , s. l. , klemola , a. r. , & hanley , c. j. 1996 , , 112 , 2110 lewis , i. j. , et al . 2002 , , 333 , 279 light , r. m. 1988 , ph.d .
thesis , yale university majewski , s. r. , siegel , m. h. , patterson , r. j. , & rood , r. t. 1999 , , 520 , l33 majewski , s. r. , et al .
2002 , asp conf .
ser . 285 : modes of star formation and the origin of field populations , 199 majewski , s. r. , skrutskie , m. f. , weinberg , m. d. , & ostheimer , j. c. 2003 , , 599 , 1082 martnez - delgado , d. , alonso - garca , j. , aparicio , a. , & gmez - flechoso , m. a. 2001 , , 549 , l63 mateo , m. 1998 , , 36 , 435 mayer , l. , governato , f. , colpi , m. , moore , b. , quinn , t. , wadsley , j. , stadel , j. , & lake , g. 2001 , apj , 559 , 754 norris , j. 1987 , , 314 , l39 norris , j. e. 1999 , , 265 , 213 odenkirchen , m. , et al .
2001 , , 548 , l165 palma , c. , majewski , s. r. , siegel , m. h. , patterson , r. j. , ostheimer , j. c. , & link , r. 2003 , aj , 125 , 1352 peebles , p. j. e. 1980 , research supported by the national science foundation .
princeton , n.j . , princeton university press , 1980 .
435 p. , piatek , s. & pryor , c. 1995 , , 109 , 1071 piatek , s. , pryor , c. , armandroff , t. e. , & olszewski , e.
w. 2001 , , 121 , 841 queloz , d. , dubath , p. , & pasquini , l. 1995 , , 300 , 31 ratnatunga , k. u. & bahcall , j. n. 1985 , , 59 , 63 roche , n. & eales , s. a. 1999 , , 307 , 703 rutledge , g. a. , hesser , j. e. , stetson , p. b. , mateo , m. , simard , l. , bolte , m. , friel , e. d. , & copin , y. 1997 , , 109 , 883 rutledge , g. a. , hesser , j. e. , & stetson , p. b. 1997 , , 109 , 907 schweitzer , a. e. , cudworth , k. m. , majewski , s. r. , & suntzeff , n. b. 1995 , , 110 , 2747 suntzeff , n. 1993 , asp conf . ser . 48 : the globular cluster - galaxy connection , 167 tolstoy , e. , et al . 2004 , arxiv astrophysics e - prints , astro - ph/0411029 van agt , s. l. t. j. 1978 , publications of the david dunlap observatory , 3 , 205 vivas , a. k. , et al . 2001 , , 554 , l33 walcher , c. j. , fried , j. w. , burkert , a. , & klessen , r. s. 2003 , , 406 , 847 yi , s. , demarque , p. , kim , y. , lee , y. , ree , c. h. , lejeune , t. , & barnes , s. 2001 , , 136 , 417 zacharias , n. , et al .
2000 , , 120 , 2131 zinn , r. , & west , m. j. 1984 , , 55 , 45 table [ sclmembers ] is presented in its entirety in the electronic edition of the astronomical journal .
a portion is shown here for guidance regarding its form and content .
the uncertainties for @xmath9 and @xmath14 are relative only , and do not include those from the photometric calibration . | the results of a wide - field survey of the sculptor dwarf spheroidal galaxy are presented .
our aims were to obtain an accurate map of the outer structure of sculptor , and to determine the level of interaction between this system and the galaxy .
photometry was obtained in two colours down to the magnitude limits of @xmath0 and @xmath1 , covering a @xmath2 deg@xmath3 area centred on sculptor .
the resulting colour - magnitude data were used as a mask to select candidate horizontal branch and red giant branch stars for this system .
previous work has shown that the red horizontal branch ( hb ) stars are more concentrated than the blue hb stars .
we have determined the radial distributions of these two populations and show that the overall sculptor density profile is well described by a two component model based on a combination of these radial distributions .
additionally , spectra of the ca ii triplet region were obtained for over 700 candidate red giant stars over the 10 deg@xmath3 region using the 2df instrument on the anglo - australian telescope .
these spectra were used to remove foreground galactic stars based on radial velocity and ca ii triplet strength .
the final list of sculptor members contained 148 stars , seven of which are located beyond the nominal tidal radius .
both the photometric and spectroscopic datasets indicate no significant extra - tidal structure .
these results support at most a mild level of interaction between this system and the galaxy , and we have measured an upper mass limit for extra - tidal material to be @xmath4 of the sculptor luminous mass .
this lack of tidal interaction indicates that previous velocity dispersion measurements ( and hence the amount of dark matter detected ) in this system are not strongly influenced by the galactic tidal field . |
the identification of the progenitor stars of supernova ( sn ) explosions is one of the central problems of stellar astrophysics . in the case of core collapse supernovae ( cc sne : types ii , ib , ic , and derived subtypes ) the progenitors
are known to be massive ( @xmath5 ) stars whose inner cores collapse to a neutron star or a black hole . in a few cases , it has been possible to constrain the properties of the progenitor star using pre - explosion images or the turn - off masses of compact clusters @xcite , but there are still many open issues regarding which stars lead to specific subtypes of cc sne ( for an extended discussion and a complete set of references , see * ? ? ?
* ; * ? ? ?
in the case of thermonuclear ( type ia ) sne , the situation is much more complex .
although a co white dwarf ( wd ) in some kind of binary is almost certainly the exploding star , the exact nature of the progenitor system has never been firmly established , either theoretically or observationally ( see * ? ? ?
* and references therein ) .
when direct identifications are not possible , the properties of the progenitors can be constrained using the stellar populations around the exploding stars .
a number of studies have done this for sne in nearby galaxies ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , but this approach has important limitations .
first , the available information , be it photometric ( e.g. * ? ? ? * ) or spectral ( e.g. * ? ? ?
* ) , is usually integrated over the entire host galaxy , although local measurements at the sn sites have been made for a small number of objects ( e.g. * ? ? ?
this effectively ignores the metallicity and stellar age gradients that must be present in the host .
second , even in surveys that work with complete host spectra , the stellar populations are not resolved . among other things , this means that the stellar light is weighted by luminosity , which can conceal many important properties of the stellar populations . in practice , the information that can be obtained from this kind of observations is restricted to average metallicities and ages , unless sophisticated fitting techniques are used to extract the star formation history ( sfh ) of the host ( see * ? ? ?
ideally , one would want to study _ resolved _ stellar populations associated with sn progenitors .
the information that can be obtained in this way is much more detailed and reliable , but it requires focusing on very nearby sne .
the present work is the first in a series of papers aimed at constraining the fundamental properties of cc and ia sn progenitors in the magellanic clouds by examining the stellar populations at the locations of the supernova remnants ( snrs ) left behind by the explosions . to do this
, we take advantage of the large amount of observational data accumulated on the stellar populations of the clouds , in particular the star formation history ( sfh ) maps published by @xcite for the smc and @xcite ( henceforth , hz09 ) for the lmc . to identify the sites of recent sne
, we rely on the extensively observed population of snrs in the mcs @xcite .
much information about the sn explosions can be extracted from the observations of snrs of both ia and cc origin @xcite , and in some cases this information can be complemented by light echoes from the sne themselves @xcite . in this first installment , we focus on the eight youngest snrs in the lmc : sn1987a , n158a , n63a , and n49 ( cc snrs ) ; [email protected] , [email protected] , n103b and dem l71 ( ia snrs ) .
this paper is organized as follows . in
[ sec : targets ] we describe the criteria that have led to the selection of our eight target snrs . in
[ sec : sntosnr ] we review their types and the characteristics of the parent sne that can be inferred from their observational properties . in [ sec : map ] we review the fundamental details of the sfh map of the lmc presented in hz09 . in
[ sec : relevance ] we discuss the relevance that the local sfh has for the properties of the sn progenitors , given our knowledge about the global sfh and the stellar dynamics of the lmc . in
[ sec : sfh ] we examine the local sfhs for the target snrs , with specific comments relating each sfh to the sne that originated the snrs . in
[ sec : disc ] we discuss the impact that our findings have in the context of our current understanding of cc and ia sn progenitors .
finally , in
[ sec : conc ] we present our conclusions and we outline some avenues for future research .
we will focus on young snrs because they are usually ejecta - dominated and still contain a great deal of information about their parent sne - in particular , the risk of mistyping young cc and ia snrs is minimal ( see [ sec : sntosnr ] ) . identifying the youngest snrs in a given set , however , is not trivial . among the snrs in the lmc ,
only one has a known age ( sn 1987a ) , and only three ( [email protected] , [email protected] , and n103b ) have more or less accurate age estimates from light echoes @xcite . in the absence of consistent age estimates for all objects ,
size is the best criterion to select the youngest ones .
much information about the snr population in the lmc can be found in the _ rosat _ atlas by @xcite , but the snr sizes in particular are not reliable and must be revised .
sizes of snrs with sharp outer boundaries are overestimated due to the large _ rosat _ psf ( e.g. [email protected] , * ? ? ?
* ) , while sizes of diffuse snrs are underestimated due to the low _ rosat _ effective area ( e.g. n23 , * ? ? ?
we have searched the literature for more recent _ chandra _ observations to constrain the lmc snr sizes , and we have selected the eight smallest objects ( sizes @xmath6 arcmin , see table [ tab-1 ] ) .
the age estimates listed in table [ tab-1 ] merit a few comments .
for snrs without sn or light echo information , ages are calculated from the snr size assuming a specific model for the snr dynamics , which can introduce large uncertainties .
in particular , the standard dynamical models for young snrs ( e.g. * ? ? ?
* ) ignore the effect of cosmic ray acceleration at the forward shock .
it is now widely accepted that energy losses due to cosmic ray acceleration can affect the size of young snrs in a noticeable way @xcite , which implies that calculations based on unmodified snr dynamics can overestimate the age by as much as 20@xmath7 ( see 5.2 in * ? ? ? * for a discussion ) . in figure [ fig-1 ]
, we illustrate the location of our eight target snrs within the large scale structure of the lmc using the data from field 13 of the southern h - alpha sky survey atlas ( shassa , * ? ? ?
two snrs , sn 1987a and n158a , are located in the 30 dor region , the most prominent active star forming region in the lmc .
two more , n49 and n63a , are in the northern part of the disk , embedded in the north blue arm discussed in hz09 and @xcite .
snrs [email protected] and n103b are in the outer parts of the lmc bar .
the last two objects , [email protected] and dem l71 , are in rather inconspicuous parts of the lmc disk , in the area called the northwest void by hz09 .
more specific discussions about the location of each snr will be given in
[ sec : sfh ] .
typing snrs as cc or type ia can be an uncertain and treacherous business . both cc and ia sne deposit a similar amount of kinetic energy ( @xmath8 erg ) in the ambient medium ( am ) , which often makes it impossible to distinguish mature cc from mature ia snrs based on their size or morphology alone .
a much more reliable way to type snrs is to examine the evidence left behind by the explosion itself : x - ray spectrum from the sn ejecta and am , snr dynamics , and properties of the compact object or pulsar wind nebula ( pwn ) , if present . in general , this can only be done for relatively young objects ( but see * ? ? ?
* ; * ? ? ?
* ) . by using methods along these lines ,
we have been able to determine the type of all the objects in our list with a high degree of confidence , and in some cases even the sn subtypes within the broader cc and ia categories . in this section
we will discuss the classification and sn subtypes of our target snrs , but before going into the details of each object , it is important to mention the work of @xcite .
these authors attempted to type _ all _ the lmc snrs known at the time by noting the distance from each object to hii regions and ob associations .
although this is a very crude method , their conclusions regarding the cc or type ia nature of our target snrs coincide with ours , except in the case of snr n103b , which will be discussed in detail in
[ sec : sntosnr : iasnrs ] and [ sec : sfh : ia ] .
our theoretical understanding of core collapse sne is still incomplete @xcite . in particular , the mapping between progenitor mass and cc sn subtype is uncertain , because key processes like stellar mass loss and binary interactions are not well understood @xcite . to set the stage for further discussions ,
the stellar evolution models of @xcite for single stars of lmc metallicity ( @xmath9 ) predict that stars between @xmath10 and @xmath3 will explode as red supergiants , retaining most of their h envelope and becoming type iip sne , stars between @xmath11 and @xmath12 will lose a large part of their envelopes and explode as type iil or type iib sne , and stars above @xmath12 will lose all their envelopes and become naked cc sne of types ib and ic . within naked cc sne
, there is some evidence that type ic sne , which are linked to long duration gamma - ray bursts @xcite come from more massive stars than type ib sne @xcite . stars that retain a massive h envelope but explode as blue supergiants instead of red supergiants form a separate class , often referred to as sn 1987-like events .
the lifetimes associated with these stellar masses range between 41 myr for an isolated @xmath13 star and 5.4 myr for an isolated @xmath12 star ( always taking the @xmath9 models from * ? ? ?
* ) . in principle
, mass loss will be facilitated by binary interactions , leading to fewer red supergiants and more type ib / c sne in binary systems , but stellar evolution calculations that include these effects are subject to an entirely different set of uncertainties @xcite .
from the point of view of the snrs , the complex and turbulent structure of most young cc snrs makes a quantitative interpretation of the x - ray spectrum in terms of specific explosion models and progenitor scenarios very challenging ( e.g. see * ? ? ? * ; * ? ? ? * ; * ? ? ? * ) .
many times , it is hard to infer the sn subtype from the observational properties of the snr , but the large intrinsic diversity of cc sne as a class can often be used to some advantage in snr studies .
@xcite argues that several aspects of snr evolution are expected to be very different depending on the subtype of the parent sn : mixing in the ejecta , fallback onto the central compact object , expansion of the pwn , interaction with the csm , and photoionization of the am by shock breakout radiation .
using arguments along these lines , @xcite inferred from the positions of the fluid discontinuities , the presence of high velocity h , and the extent of the clumpy photoionized pre - sn wind in the cas a snr that its progenitor must have been a type iin or type iib event .
this ` prediction of the past ' was later confirmed by the spectroscopy of the light echo of the cas a sn @xcite , which is very similar to the spectrum of the type iib sn1993j .
although this agreement is certainly encouraging , we must insist that studies based on snrs are still a long way from providing a robust method of subtyping cc sne - as an example , the type iin / iib classification of cas a by @xcite was challenged by @xcite , who argued for a type ib progenitor . in the following paragraphs
, we examine each of the four target cc snrs in more detail . for a summary , see table [ tab-2 ] .
[ [ snr - n49 ] ] snr n49 + + + + + + + this snr harbors one of only two magnetars known outside the milky way : sgr 0526@xmath066 . in principle , the presence of a compact object should immediately classify this object as a cc snr , but the association between this magnetar and the snr has been controversial @xcite . even disregarding the compact object , the ejecta emission shows significantly enhanced abundances from o and si , but a comparatively small amount of fe @xcite , and the snr is located within the ob association lh53 @xcite .
taken together , these arguments lend strong support to a cc origin .
the complex filamentary structure of the shocked am suggests that dense material surrounded the sn at the time of the explosion @xcite , which favors a progenitor with a slow wind , maybe a type iip sn .
unfortunately , this is just an educated guess , because the complex multi - phase x - ray emission of the snr and the poorly known age make the interpretation of the observations very challenging .
[ [ snr - n63a ] ] snr n63a + + + + + + + + this snr has no detected compact object , although the upper limits do not exclude the presence of a low - activity pwn @xcite .
it is embedded in the large hii region n63 , and it also appears to be located within an ob association ( ngc 2030 , * ? ? ?
* ) , making a cc type very likely .
the size , morphology , and x - ray spectrum show evident signs of a complex interaction with a highly structured am , and they seem to indicate that it is expanding into a large cavity @xcite , which suggests a massive progenitor with a fast wind , maybe a type ib / c sn . however , there is no additional evidence to support this conclusion because the x - ray emission , which is dominated by the shocked am , reveals very little about the properties of the sn ejecta @xcite .
[ [ sn-1987a ] ] sn 1987a + + + + + + + + the classification and subtype of sn 1987a are obvious from the sn spectroscopy . the vast amount of information available on this object
is summarized in @xcite and other publications in the same volume . for our purposes , it suffices to mention that the progenitor of this sn is known to have been a blue supergiant star , sk @xmath14 , whose initial mass has been estimated at @xmath15 @xcite , and might have been part of a close binary system @xcite .
[ [ snr - n158a ] ] snr n158a + + + + + + + + + this object harbors a well - observed pwn that types it as a cc snr and constrains its age to be @xmath16800 yr @xcite . the sn subtype classification has been rather controversial .
the snr dynamics indicate that the shock wave is moving into dense , clumpy csm similar to what can be found around a massive wolf - rayet star , and the presence of strong o and s lines in the x - ray spectrum of the innermost ejecta reveal that at least some of the heavy elements in the ejecta have not fallen back onto the central neutron star @xcite .
this would favor a massive type ib / c progenitor , but both the detection of h in the pwn filaments @xcite and a recent re - analysis of the ejecta emission @xcite seem to indicate that the progenitor might have been in the @xmath17 range , implying a type iip explosion ( for more detailed discussions , see * ? ? ?
* ; * ? ? ?
our current understanding of type ia sn progenitors is still extremely sketchy @xcite , but several interesting trends have been inferred from the bulk properties the host galaxies .
any theoretical model for type ia progenitors must account for the fact that type ia sne explode in elliptical galaxies with very little star formation ( sf ) , but at the same time the rate of ia events in star forming galaxies appears to scale with the specific star formation rate @xcite .
moreover , type ia sne exploding in elliptical galaxies are on average dimmer that those exploding in star forming galaxies @xcite .
@xcite and @xcite used these observational facts to postulate _ two _ populations of type ia sn progenitors : a ` prompt ' population with short delay times ( of the order of a few hundred myr ) , associated with recent sf and leading to somewhat brighter sn ia , and a ` delayed ' population with longer delay times ( of the order of gyr ) , not associated with recent sf and leading to somewhat dimmer sn ia . in this two - component model , the specific type ia sn rate in a given galaxy is expressed as @xmath18 , with @xmath19 being the total stellar mass in the galaxy and @xmath20 the specific star formation rate .
it is important to stress that the observed rates do not _ require _ the existence of two components - in some theoretical scenarios , type ia sne from a single progenitor channel can explode with both very short and very long delay times @xcite .
nevertheless , there have been several attempts to associate the prompt and delayed progenitor populations to the two leading theoretical scenarios for type ia sne : single degenerate ( sd ) systems , in which the wd accretes material from a non - degenerate companion and double degenerate ( dd ) systems , in which the wd accretes material from another wd .
so far , none of these attempts has succeeded ( see f.i . * ? ? ? * ; * ? ? ? * ) , and the identity of type ia sn progenitors remains a mystery . despite all the uncertainties regarding their progenitors , type ia sn explosions as a class are much more homogeneous and have less intrinsic dispersion than cc sne .
in particular , there is a simple relationship between the structure of the ejecta and the peak brightness of the sn that is well reproduced by one - dimensional delayed detonation ( ddt ) explosion models @xcite .
this allows to map the vast majority of sn ia onto a sequence of bright to dim events based on the amount of @xmath21ni that they synthesize . generally speaking , type ia snrs
are also much less turbulent than cc snrs , and most of them seem to be interacting with a relatively unmodified am @xcite , although there are exceptions like the kepler snr @xcite and snr n103b ( * ? ? ?
* , see also the discussion in [ sec : conc - ia ] ) .
thanks to this set of circumstances , it is generally easier to interpret the x - ray emission of type ia snrs quantitatively in terms of specific explosion models , provided that the dynamic evolution of the snr and the nonequilibrium processes in the shocked plasma are properly taken into account @xcite .
it is also possible to estimate the brightness of the parent event from the mass of @xmath21ni synthesized by the preferred ddt explosion model , as shown by @xcite in the case of the tycho snr .
these four objects were classified as type ia snrs by @xcite based on their lack of compact object or pwn and the general properties of their x - ray emission , which is dominated by fe lines and has only weak or absent lines from o. in order to confirm these classifications and derive the sn subtype , it is necessary to perform an in - depth analysis of the ejecta emission , as done by @xcite for snr [email protected] .
the x - ray emission of the other three objects will be the subject of a forthcoming publication ( * ? ? ? * henceforth bh09 ) , but the main results of that analysis are presented in the following paragraphs , and summarized in table [ tab-3 ] .
[ [ snr - dem - l71 ] ] snr dem l71 + + + + + + + + + + + this is the oldest object in our type ia snr list , and the only one without a light echo age estimate . @xcite
determine an age of @xmath22 from the snr dynamics , and yet the x - ray spectrum appears dominated by shocked fe from the sn ejecta , specially in the center @xcite .
the old age of this snr makes the analysis of the ejecta emission somewhat challenging , but bh09 find that it can be reproduced by ddt models for normal type ia sne .
[ [ snr - n103b ] ] snr n103b + + + + + + + + + this object was initially classified as a cc snr by @xcite based on its location at the edge of the hii region dem 84 and 40 pc away from the ob association ngc 1850 . however , the x - ray spectrum is strongly suggestive of a type ia origin @xcite .
the snr is also remarkable in that it shows a strong east - west asymmetry @xcite , which has been interpreted as a sign of some kind of csm interaction , mainly by analogy to the kepler snr .
this asymmetry also makes the ejecta analysis challenging , but bh09 find a relatively good match to the spectrum using ddt models for moderately bright type ia sne with a snr age close to the 800 yr estimated from the light echo by @xcite . [
[ snr-0509 - 67.5 ] ] snr [email protected] + + + + + + + + + + + + + + + + + + + + + + + + + + + next to sn1987a , this snr has the most secure subtype classification in the lmc . in 2008 , two teams analyzed independently the optical spectrum of the sn light echo @xcite and the x - ray emission and dynamics of the snr @xcite , and came to the same conclusion : snr [email protected] was originated @xmath23 yr ago by an exceptionally bright type ia sn that synthesized @xmath24 of @xmath21ni .
this agreement is a very important validation of the modeling techniques introduced in @xcite that bh09 apply to the other three target type ia snrs , and in particular of the capability of the models to recover the sn subtype .
[ [ snr-0519 - 69.0 ] ] snr [email protected] + + + + + + + + + + + + + + + + + + + + + + + + + + + the final object in our type ia snr list has an estimated age of @xmath25 yr from its light echo @xcite .
its x - ray emission is well reproduced by a moderately bright type ia sn model that synthesizes @xmath26 of @xmath21ni ( bh09 ) .
the sfh map that we use in the present work is described in full detail in hz09 .
the map was elaborated using four band ( _ u _ , _ b _ , _ v _ , and _ i _ ) photometry from the magellanic clouds photometric survey @xcite , which has a limiting magnitude between @xmath27 and @xmath28 in _ v _ , depending on the local degree of crowding in the images .
data from more than 20 million stars was assembled to produce color - magnitude diagrams in 500 @xmath29 cells encompassing the central @xmath30 of the lmc ( see figure 4 in hz09 ) , and then the starfish code @xcite was applied to derive the local sfh for each cell .
cells with enough stars in them , like the eight cells that contain our target snrs , were further subdivided into four @xmath31 subcells .
the sfh of each cell is given at thirteen lookback times between @xmath32 ( 6.3 myr ) and @xmath33 ( 17.8 gyr ) , and it is broken into four metallicity bins : @xmath34 and @xmath35 . for reference in further discussions , we reproduce the sfh of the entire lmc from hz09 in figure [ fig-2 ] .
the error bars on the total sfh represented with the gray shaded area are dominated by crowding effects ( see 3.3 in hz09 ) .
although no metallicities were fitted for ages below 50 myr , the plots display the canonical lmc metallicity ( @xmath9 ) in the most recent age bins , which is reasonable in view of the high degree of homogeneity in the metallicity of the ism and the young stars in the lmc @xcite .
a detailed discussion of the sfh of the lmc and its interpretation in the context of the lmc s past history can be found in 5 of hz09 .
for our purposes , it suffices to note that , after an initial episode of sf in the distant past , the lmc went into a quiescent period that lasted until 5 gyr ago , and since then it has been forming stars at an average rate of @xmath36 , with episodes of enhanced sf at 2 gyr , 500 myr , 100 myr , and 12 myr . from figure [ fig-2 ] , it is obvious that the vast majority of the stars in the lmc have ages above 1 gyr .
most of these old stars have metallicities of one tenth solar or lower .
during the lifetime of a galaxy , several processes naturally mix the stellar populations . these include both internal processes like the ` churning ' of the disk by spiral arms @xcite and external processes like tidal interactions and mergers @xcite . in this context , the properties of the stellar population ( and hence the sfh ) in the neighborhood of a young snr will only be representative of the sn progenitor up to a certain lookback time , @xmath37 . in principle
, @xmath37 can be calculated for each location within a galactic disk provided there is a viable dynamic model that includes all the relevant processes .
unfortunately , no such model exists for the lmc , despite the wealth of observational information available .
the lmc disk is warped @xcite and might also be flared @xcite , and it has a rich history of tidal interactions with the smc and ( maybe ) the milky way , which may have important effects on the stellar dynamics ( see * ? ? ?
* ; * ? ? ?
the vestigial arms seen in hi @xcite are probably originated by these tidal interactions , but the details of this process are not well understood @xcite .
even the nature of the most prominent feature in the disk - the lmc bar - and its role in the dynamics of the galaxy are unclear @xcite . without a reliable way to calculate @xmath37 for each of the subcells that contain our target snrs ,
all we can do is estimate the relevant timescales for a number of different processes .
the physical size of the subcells in the hz09 map is @xmath38 pc ( assuming @xmath39 kpc , * ? ? ?
* ) , and the velocity dispersions for the young disk and old disk populations determined by @xcite are @xmath10 and @xmath40 , respectively .
thus , the length of time that it would take an average star of the young ( old ) disk to drift from one subcell to the next in the absence of restoring forces , @xmath41 , is 43 ( 16 ) myr .
these timescales are not relevant for the progenitors of cc sne , which should belong to the young disk , but they will be very important for sn ia progenitors , which could be quite older . in any case , @xmath37 should be much larger than @xmath41 , because ( a ) some regions of the lmc are more homogeneous than others , which means that stars have to drift over larger distances in order to find substantially different stellar populations , and ( b ) there are restoring forces like gravity that maintain the structural integrity of the disk and act to limit stellar drift .
we have quantified the spatial homogeneity of the stellar populations around our target snrs in figure [ fig-3 ] .
we plot the absolute value of the relative differences in the stellar populations as a function of the distance from the center of each of the eight subcells that contain the cc and ia snrs in our list . to calculate the relative differences , we have integrated the sfh in each neighboring subcell , taking all the time bins were the differences between the neighbor and the snr subcell were statistically significant ( i.e. , the error bars did not overlap ) up to a lookback time of 1.1 gyr , and then divided by the total number of stars formed in the snr subcell . at each distance
, the relative difference is the mean of the relative differences between the snr subcell and all the subcells at that distance . figure [ fig-3 ] shows that some of our target snrs are in remarkably homogeneous regions of the lmc disk .
these include the cc snrs n49 and n63a in the blue arm and the type ia snr [email protected] in the northwest void , with average relative differences in the stellar populations below @xmath42 within 1400 pc of the central subcell .
this distance translates into @xmath41 values of 215 and 80 myr for young and old disk stars in the absence of restoring forces .
the effect of restoring forces on the value of @xmath37 is more difficult to estimate .
if no chaotic processes intervene , neighboring stars will tend to move together through the disk , which explains why some lmc structures like the bar and the northwest arm show up in the hz09 map with lookback times as large as 1 gyr ( see their figure 8) .
this long survival time is not restricted to large structures - in the solar neighborhood , there is evidence that several groups of old stars ( 2 to 8 gyr ) are moving together through the disk of the milky way @xcite .
but smaller structures like young stellar clusters seem to disappear on timescales of the order of @xmath43 myr in the lmc @xcite , which is roughly equivalent to the dynamic crossing time of the lmc disk
. we will adopt this value as a figure of merit for @xmath37 in a single subcell of the sfh maps .
since the result of @xcite applies to young stars , this implies that restoring forces increase the value of @xmath41 by at least a factor @xmath44 , but we stress that this is just a very rough estimate .
we conclude that the relevance of the local sfhs for type ia sn progenitors will depend on both the homogeneity of the stellar populations around each subcell ( figure [ fig-3 ] ) and the age of the progenitors .
if the lifetime of type ia progenitors in the prompt channel is as short as the 180 myr claimed by @xcite , it should be possible to use the local sfhs of type ia snrs in the lmc to explore their properties . for progenitors with longer lifetimes , the stellar context of each snr
should be taken into account .
objects like snr [email protected] might allow exploration of timescales up to several hundred myr , but snrs like [email protected] probably will not .
the local sfhs in the subcells containing the eight snrs in our list are plotted in figures [ fig-4 ] and [ fig-6 ] .
the lifetime of an isolated @xmath13 star with @xmath9 from @xcite has been indicated by a dashed vertical line in all the plots for illustrative purposes . for simplicity ,
we have collapsed all the sfh bins at ages above 2 gyr into a single bin at 10 gyr .
several interesting average quantities can be calculated from the local sfhs .
we have listed two such quantities in tables [ tab-2 ] and [ tab-3 ] : the average metallicity of all the stars formed in the subcell , @xmath45 and their average age @xmath46 .
these averages are always dominated by the large population of old stars in each subcell , and therefore they are irrelevant for the properties of cc sn progenitors - the values in table [ tab-2 ] are merely provided for comparison with the values in table [ tab-3 ] ( see discussion in [ sec : disc ] ) .
the average values for the entire lmc are @xmath47 and @xmath48 gyr .
the most salient feature of the sfhs around the four cc snrs ( figure [ fig-4 ] ) is that they are strongly dominated by intense bursts of star formation in the recent past ( @xmath49 myr ) .
this is of course expected , and in the cases where the snrs have been typed for their close proximity to young stellar clusters ( notably , snr n63a ) , it does not reveal any new information .
however , the timing of these bursts and their intensity will determine the properties of the population of massive stars that can be found at each location , and hence the likelihood of each cc sn subtype . to highlight these aspects
, we display the most recent bins of the sfhs associated with the cc snrs in greater detail in figure [ fig-5 ] , alongside the lifetimes of isolated massive stars with @xmath9 from @xcite .
we have convolved the three most recent sfh bins with a standard salpeter imf to calculate the fraction of massive stars that are exploding now as cc sne ( @xmath50 ) from progenitors in three mass intervals : @xmath10 to @xmath51 , @xmath52 to @xmath53 and above @xmath2 .
we list these fractions in table [ tab-2 ] for each of the cc snrs .
the interval cuts are the stellar masses whose lifetimes correspond to the upper edges of the first and second age bins in the sfhs : 9.4 and 18.9 myr .
we remind the reader that these values of @xmath50 are calculated using isolated star models that do not take into account the potentially large effects of binarity on stellar evolution .
an entirely different but also potentially serious problem comes from the fact that massive stars are notoriously difficult to study using photometry alone @xcite . because starfish uses _
all _ the stars ( not just the massive ones ) in each subcell to calculate the sfr at each age , the values of @xmath50 might not be severely affected by this , but to this date there has been no systematic calibration of the starfish results for young stellar populations . to reflect these and other caveats , we do not list error bars on the @xmath50 values , which should only be regarded as approximate . [ [ snr - n49 - 1 ] ] snr n49 + + + + + + + the integrated sfh for the north blue arm region that contains snr n49 and snr n63a is dominated by a coherent episode of low - metallicity star formation 100 myr ago ( see 5.1.3 and figure 16 in hz09 ) , which is apparent in the corresponding panels of figure [ fig-4 ] .
this is the reason why the values of @xmath45 and @xmath46 for snrs n49 and n63a are lower than those of the other snrs .
many parts of the north blue arm have also had noticeable star formation activity in the last 40 myr , although not as intense as in the more prominent star forming regions of the lmc , 30 dor and constellation iii .
snr n49 is in one such region , which had a moderately intense sf burst 12 myr ago , but very little sf activity in the most recent bin centered at 6.3 myr ( see figure [ fig-5 ] ) . from these properties of the sfh
, the expectation is that the majority of the cc sn progenitors in this subcell should be stars between @xmath52 and @xmath53 ( see table [ tab-2 ] ) . even taking the upper limit of the sfr in the most recent bin and the lower limit on the bin at 12 myr
, the fraction of cc sn progenitors with masses above @xmath53 remains below @xmath54 , with all the caveats associated to the calculated values of @xmath50 .
this is in good agreement with the properties of the snr discussed in
[ sec : sntosnr : ccsnrs ] , and it has interesting implications for the association of snr n49 with sgr 0526@xmath066 .
magnetars are thought to be originated by very massive ( @xmath55 ) stars @xcite , but such stars appear to be very scarce around snr n49 .
one possibility is that the magnetar was formed elsewhere and the association is coincidental .
@xcite examined this issue in detail , and came to the conclusion that the link between snr n49 and sgr 0526@xmath066 is considerably less convincing than those of other magnetars in snrs .
more recently , @xcite performed a nir survey of the area around snr n49 and identified a young stellar cluster ( sl 463 ) at a projected distance of @xmath56 pc northeast of sgr 0526@xmath066 that could have been the birthplace of the magnetar .
this cluster does fall partially on a neighboring subcell of the hz09 map with intense sf at 6.3 myr , consistent with the 5 to 20 myr age estimates for sl 463 by @xcite , and much more promising as a birthplace of massive cc progenitors ( @xmath57 above @xmath53 ) .
as pointed out by @xcite , if this hypothesis is true the magnetar must have been ejected from its birthplace with a certain velocity , and should have a measurable proper motion ( see their 3.2 ) .
another possibility is that the association between snr n49 and sgr 0526@xmath066 is indeed real , but not all magnetars have stellar progenitors more massive than @xmath3 .
[ [ sn-1987a-1 ] ] sn 1987a + + + + + + + + the sfh associated with this snr is of particular interest because , together with the known mass of the progenitor , @xmath1 , it can provide some test of the robustness of our sfh approach to cc sn progenitors .
@xcite conducted a detailed study of the immediate neighborhood of sn1987a within the 30 dor region using _ hst _ data , and found a loose young cluster with an age of @xmath58 myr , which they identified as the birthplace of sn1987a s progenitor .
the local sfh drawn from the hz09 map is indeed dominated by an intense sf episode 12 myr ago , in good agreement with the results of @xcite . from the sfh
, we expect @xmath59 of the cc sn progenitors in this region to be stars between @xmath52 and @xmath53 ( see figure [ fig-5 ] ) .
this agreement is encouraging , and indicates that some information about the progenitor mass of cc sne can be recovered from the sfh maps of hz09 .
[ [ snr - n63a-1 ] ] snr n63a + + + + + + + + like snr n49 , snr n63a is located in a part of the blue arm with sf activity in the recent past .
the sfh in this subcell , however , is different from that around snr n49 in that it is dominated by the most recent bin centered at 6.3 myr . as a result , @xmath60 of the cc sn progenitors in this subcell
are expected to be stars more massive than @xmath53 . with a salpeter imf , roughly @xmath61 of these stars
will be in turn more massive than @xmath12 , which makes a type ib / c origin for snr n63 plausible , as suggested by the properties of the snr .
if this were true , snr 63a would be one of the youngest nearby type ib / c snrs , and a closer examination of this object to locate the elusive compact object and study the ejecta emission in greater detail would be of the highest interest .
[ [ snr - n158a-1 ] ] snr n158a + + + + + + + + + this object is also located in 30 dor , but in a region with even more vigorous recent sf than the neighborhood of sn 1987a .
according to our estimates , @xmath62 of the cc sn progenitors in this subcell should be stars more massive than @xmath53 .
again , this is compatible with the relatively massive progenitor suggested by the snr properties ( see [ sec : sntosnr : cc ] ) , but unfortunately the temporal resolution of the sfh does not allow us to discriminate between a type iip progenitor with @xmath63 @xcite and a type ib / c progenitor with @xmath64 @xcite .
from the point of view of the stellar population around snr 158a , both hypotheses are equally plausible .
the local sfhs around the four type ia snrs are displayed in figure [ fig-6 ] . since the age of the progenitors is not known _ a priori _ , we also display the integrated sfhs in figure [ fig-7 ] to provide a more intuitive picture of the makeup of the stellar populations in these locations and how they have evolved with time . with one exception ( snr n103b ,
see below ) , these sfhs are very different from those associated with the cc snrs .
the sfhs of snrs dem l71 , [email protected] , and [email protected] show very little ( but nonzero ) activity in the last 200 myr , resulting in old and metal poor stellar populations , typical of the regions of the lmc without ongoing star formation .
the sfhs of these three snrs are also punctuated by bursts of sf at 600 myr and 2 gyr whose prominence varies from object to object , but this is probably more related to the global properties of the lmc ( see figure [ fig-2 ] ) than to the type ia sn progenitors .
given the coincidence of the lmc disk crossing time and the upper limit for the delay time of short - lived sn ia progenitors ( 180 myr according to * ? ? ?
* see [ sec : relevance ] ) , it is possible to use the a+b model introduced in [ sec : sntosnr : ia ] and the local sfhs to predict what fraction of type ia sne will come from prompt and delayed progenitors in each subcell ( @xmath65 ) . for the prompt component ,
we have calculated an average star formation rate by integrating the sfhs between 64 myr ( the minimum time necessary to produce a co wd at the metallicity of the lmc , * ? ? ? * ) and 180 myr , and we have multiplied it by the value of @xmath66 from @xcite , @xmath67 . for the delayed component , we have simply multiplied the number of stars in each subcell by the value of @xmath68 from @xcite , @xmath69 .
the values of @xmath65 for each type ia snr are listed in table [ tab-3 ] , and can be compared with the values obtained for the whole lmc using the integrated sfh from figure [ fig-2 ] : @xmath70 and @xmath71 . even more so than in the case of cc sne , we stress that these numbers should be considered with caution , because they make strong assumptions about the properties of type ia sn progenitors .
in particular , increasing the upper limit to the age of the prompt component from 180 to 300 myr results in an increase of the fraction of _ delayed _ type ia progenitors between 8 and 15 percentual points , depending on the snr . at a first glance , it is somewhat surprising that the sn ia rates from prompt and delayed progenitors are always comparable both in the whole lmc and at the locations of the individual type ia snrs .
this can be easily understood by examining figure 6 in @xcite : for low values of the sfr , the _ specific _ rate of type ia sne from the prompt and delayed components is very similar , and low - intensity sf has been widespread across the lmc during the last 5 gyr ( see [ sec : map ] and hz09 ) .
[ [ snr - dem - l71 - 1 ] ] snr dem l71 + + + + + + + + + + + the exceptionally low rate of sf between 64 and 180 myr in the subcell containing this snr makes it the most likely object ( @xmath72 probability ) to be associated with a delayed type ia sn progenitor .
unfortunately , dem l71 is also the oldest type ia snr , and detailed information about its parent sn is hard to extract from the observations . in particular ,
the ejecta emission is not remarkable in any way , and seems consistent with normal type ia sn explosion models ( bh09 ) .
the forward shock is running into material with a factor @xmath73 density range @xcite , but this can be easily explained by inhomogeneities in the ism .
the snr dynamics do not indicate any substantial modification of the am by the progenitor @xcite .
[ [ snr - n103b-1 ] ] snr n103b + + + + + + + + + of all the sfhs associated with type ia snrs that we discuss here , that of snr n103b is without doubt the most remarkable .
this snr is in a region of the lmc bar that has seen vigorous sf activity in the recent past , with a prominent extended peak between 100 and 50 myr ( probably associated to the nearby cluster ngc 1850 , * ? ? ?
* ) and a more recent one 12 myr ago .
the intensity of this last peak is even larger than those associated with sn 1987a and snr 158a in the 30 dor region .
it is not surprising that @xcite mistyped snr n103b as a cc snr before a good quality x - ray spectrum was available , based on this evident association with recent sf .
this does not necessarily imply that the progenitor of snr 103b was a young star , although the predicted fraction of prompt ia progenitors ( @xmath74 ) is higher for this snr than for any of the others . however
, if the properties of the snr itself are taken into account , the association with recent sf becomes more intriguing .
@xcite noted a number of similarities between the strong lateral asymmetry of snr n103b and that of kepler s snr . in the case of snr n103b , it is unclear whether this asymmetry is due to an interaction with an ism structure ( e.g. , the nearby hii region dem 84 ) or to some kind of csm modified by the sn progenitor . in the kepler snr , however , it has been shown that the asymmetry is indeed due to interaction with a csm modified by the type ia sn progenitor ( see * ? ? ? * and references therein ) , suggesting that either the progenitor or its companion must have been relatively massive . from the shape of the sfh , a relatively young ( less than 150 myr ) and metal - rich ( @xmath9 ) progenitor for snr n103b seems a likely possibility .
[ [ snr-0509 - 67.5 - 1 ] ] snr [email protected] + + + + + + + + + + + + + + + + + + + + + + + + + + + this snr is in a region that hz09 called the northwest void due to its conspicuous lack of recent star formation .
in fact , hz09 argue that the old and metal - poor stars in this part of the lmc are representative of the primordial stellar population of the lmc . moreover , the snr is in a very homogeneous region ( see figure [ fig-3 ] ) , with neighboring subcells having very similar stellar populations .
the closest subcells with noticeable sf activity ( above @xmath75 ) at @xmath76 myr are roughly 1 kpc away from the snr
. this would not be noteworthy for a generic type ia snr , but thanks to the work of @xcite and @xcite we know that snr [email protected] was originated by an exceptionally bright type ia sn which synthesized @xmath24 of @xmath21ni .
this seems to be at odds with the conventional wisdom regarding prompt and delayed type ia sn progenitors , which holds that exceptionally bright type ia sne are usually associated with younger stellar populations @xcite , but in fact the local sfh predicts equal contributions from prompt and delayed progenitors even in this remarkably quiet part of the lmc .
if this snr did have a relatively young and massive progenitor , however , it appears to have left the am around it relatively undisturbed @xcite .
[ [ snr-0519 - 69.0 - 1 ] ] snr [email protected] + + + + + + + + + + + + + + + + + + + + + + + + + + + this snr falls in line with [email protected] and dem l71 in having a low sfh at all times , although the local stellar population appears to be more metal rich than in the other type ia snrs ( see figure [ fig-7 ] ) .
other than that , snr [email protected] is unremarkable both in its overall structure and dynamics and in the properties of its ejecta emission .
from the local sfh , delayed type ia progenitors are slightly favored over prompt ones , but not significantly .
the combination of snr studies and local sfhs that we have introduced in this paper is a promising method to further our understanding of cc sn progenitors , but it needs to be refined before it can have a significant impact on this field of research .
two major issues need to be addressed .
first , the techniques used to determine the subtype of cc sne from their snrs are still too crude to provide consistent results , even for well observed objects like the cc snrs in our sample . and
second , the ability of tools like starfish to recover the sfh from mixed stellar populations at the ages that are relevant for the evolution of cc sn progenitors ( below 40 myr ) using photometric data needs to be firmly established . it would be interesting to investigate the possibility of an increased temporal resolution at very early times ( less than 10 myr ) in order to distinguish between type iil / b and type ib / c sn progenitors , but this might require support from spectroscopic surveys like the vlt - flames @xcite .
we hope to address these issues in future publications . despite the limitations of the method
, we have been able to obtian some interesting results . in general
, we can say that the properties of the snrs and the local sfhs are compatible with each other , allowing for the large uncerainties discussed above . among our target objects ,
the sfh around snr n49 seems to indicate that stars more massive than @xmath53 are scarce in this part of the lmc .
this opens two possibilities : either very massive ( @xmath55 ) stars are not necessary to produce magnetars , or the source sgr 0526@xmath066 is not in fact associated with snr n49 , as suggested by @xcite .
the sfh around snr n63a seems compatible with a very massive sn progenitor , maybe a type ib / c sn , but other possibilities can not be discarded . for snr 158a , the temporal resolution of the hz09 maps is too coarse to resolve the issue of the progenitor . in any case ,
our findings stress the importance of revisiting and reanalyzing the x - ray emission from these young cc snrs in detail to learn more about the sn explosions that originated them .
we have found that the combination of snr studies and local sfhs is a powerful tool to explore the properties of type ia sn progenitors . the x - ray emission from type ia snrs
is well understood in terms of explosion physics , and the ability of snr studies to recover the type ia sn subtype ( i.e. , bright vs. dim ) has been demonstrated by @xcite . moreover , the combination of starfish and a data set like the mcps is ideally suited to characterize the stellar population in different parts of the lmc . among the type ia snrs that we have studied , one ( snr n103b )
is found in a region that has experienced vigorous sf in the recent past , but the other three are associated to old and metal - poor stellar populations . on a first inspection
, it would be tempting to establish connections between these two kinds of environments and the prompt and delayed components to type ia progenitors proposed by @xcite , but we have found that the situation is more complex than that . even in regions with a remarkably small amount of recent sf
, it is very hard to isolate objects that arise unambiguously from delayed type ia progenitors .
this is due to the high efficiency of the still unknown mechanisms that turn co wds from relatively young , intermediate - mass stars into type ia sn progenitors of the prompt component @xcite . in order to identify with some confidence a type ia snr as a product of the prompt or delayed component
, it would have to be either in a region with very vigorous sf or in a pristine region with no measurable sf activity in the appropriate range of ages .
since even elliptical galaxies appear to have some residual sf during their entire lifetimes @xcite , isolating the delayed type ia progenitors in the local universe to study their properties in detail might be a very challenging task , unless a type ia sn is found in a nearby globular cluster @xcite .
our results allow us to test specific theoretical predictions about type ia progenitors , like the claim by @xcite that the type ia sn rate should be very low for metallicities lower than a tenth of the solar value .
this is based on the so - called ` accretion wind ' scenario for single - degenerate type ia progenitors , which requires a minimum opacity in the material transferred from the companion to the wd @xcite .
the average _ stellar _ metallicity is close to this value or even much lower around three of the four type ia snrs that we have examined , which seems hard to reconcile with the results of @xcite , although we stress that all the regions that we examined do contain a small number of stars with higher metallicities .
similar concerns about this prediction and its implications have been raised by @xcite , who found several type ia sne in low - metallicity dwarf galaxies .
the accretion wind scenario also makes strong predictions about the shape of the csm around type ia progenitors that are not substantiated by the dynamics of known type ia snrs @xcite . in this context , an interesting possibility is opened by the recent work of @xcite , which allows one to make _ direct _ measurements of the metallicity of type ia sn progenitors using mn and cr lines in the x - ray spectra of young snrs .
if this technique could be applied to the lmc snrs , we would be able to contrast the results with the properties of the stellar populations , and test theoretical ideas about the role of metallicity in different kinds of type ia sne ( e.g. * ? ? ?
two of the snrs we have examined have remarkable properties with important implications for type ia sn progenitors .
the unusual morphology of snr n103b ( see * ? ? ? * and references therein ) , which is strongly suggestive of some kind of csm interaction , has become even more noteworthy in light of the vigorous recent sf revealed by the local sfh .
it appears that snr n103b might be a member of an emerging class of type ia snrs with csm interaction that could be associated with relatively young and massive progenitors that lose an appreciable amount of mass before exploding as type ia sne .
this class would include the kepler snr @xcite in our galaxy and other lmc snrs @xcite , but the local sfhs around these objects should be examined to confirm this possibility .
evident signs of csm interaction , however , can not be found in other well studied type ia objects like tycho and sn 1006 in our own galaxy or the other three lmc snrs that we have analyzed here @xcite , indicating that a majority of type ia progenitors do not modify their surroundings in a noticeable way .
since it is unlikely that _ all _ these other objects have had progenitors from the delayed component , we are left with two possibilities : either the amount of mass loss during the pre - sn evolution of prompt type ia progenitors has a large dynamic range or there is more than one way to produce type ia sne with short delay times .
the properties of snr [email protected] are also remarkable , for entirely different reasons .
@xcite found @xmath77 for this sn , which translates into to a @xmath78 magnitude at maximum light close to @xmath79 @xcite .
yet , the sn exploded in a large region of the lmc with very little sf in the recent past ( see figure [ fig-3 ] ) .
the stars in the subcell that contains this snr are on average very old ( @xmath80 gyr ) and metal - poor ( @xmath81 ) .
@xcite do find some relatively bright type ia sne associated with old stellar populations , but all their objects with peak @xmath78 magnitude above @xmath82 and ages above 5 gyr have large error bars on the age axis ( see their figure 5 ) .
thus , snr [email protected] probably the first _ bona fide _ example of an exceptionally bright type ia sn associated with an old stellar population .
we note that our measurement of @xmath46 should be very reliable , because it has not been derived from a luminosity - weighted spectrum .
it is important to stress that these bulk properties of the stellar population around snr [email protected] do not preclude a relatively young progenitor for this object . during the age range that we have adopted for prompt type ia progenitors , @xmath83 of stars
were formed in the subcell that contains snr [email protected] . with a salpeter imf ,
roughly @xmath84 of this mass is in the 4 to 6 @xmath85 range ( the zams masses that give co wds on timescales shorter than 180 myr according to * ? ? ?
* ) , which results in a few hundred co wds from young stars .
this number may seem small , but observational constraints on the percentage of co wds that eventually explode as sn ia are high ( 2 to 40 % according to * ? ? ?
* ) , making a prompt progenitor for snr [email protected] a perfectly reasonable possibility .
the fact that an object like snr [email protected] appears in a sample of only four snrs implies that bright type ia sne in old stellar populations may not be an exceptional occurrence , which should be taken into account when examining the contribution of bright and dim type ia sne in cosmological studies .
our results underline the dangers of trying to understand the behavior of type ia sn progenitors by studying only the bulk properties of unresolved stellar populations in distant galaxies .
if the lmc had been a distant type ia sn host , two objects with such radically different sfhs as snrs n103b and [email protected] would have been assigned the same age and metallicity .
even average quantities obtained from resolved stellar populations like @xmath46 and @xmath45 can be misleading if they are used by themselves to characterize the properties of sn progenitors - compare the values for cc and type ia snrs from tables [ tab-2 ] and [ tab-3 ] .
in this paper , we have presented the first systematic study of the stellar populations around cc and type ia snrs in the lmc .
our ultimate goal is to use all the available information on the x - ray emitting snrs and the resolved stellar populations of the magellanic clouds to improve our understanding of cc and type ia sn progenitors , their final evolutionary stages , the sn explosions that mark their demise , and the aftermath of these explosions . in that broad context
, this work only represents a first exploration of the many possibilities that are opened by recent theoretical and observational advances in both snr research and stellar population studies .
we plan to pursue this line of research in the future , increasing the sample of objects and refining the techniques that we have presented here .
we have found that the local sfhs around the cc snrs in our sample ( n49 , sn 1987a , n63a , and n158a ) are always dominated by significant episodes of sf in the recent past ( @xmath49 myr ) , as expected from previous observational and theoretical work .
the timing and intensity of these sf episodes can provide interesting constraints on the masses of cc sn progenitors , but more work is needed to explore the full potential of this method .
the local sfhs have also allowed us to study the ages and metallicities of the stellar populations around our target type ia snrs ( dem l71 , n103b , [email protected] , and [email protected] ) in great detail .
we have found that type ia sne explode in a variety of environments , ranging from old and metal - poor populations to sites with vigorous sf in the recent past . using the two - component model proposed by @xcite
, we have explored the relationship between specific properties of type ia sne and their parent stellar populations .
we have seen that extremely bright type ia sne can explode very far away from any significant star formation activity ( snr [email protected] ) , and that type ia sne associated with young stellar populations might sometimes experience significant mass - loss before they explode ( snr n103b ) .
recent studies of extragalactic type ia sn rates and our own findings suggest that reality is probably too complex to be explained with the popular two - component progenitor model .
if this is so , high - quality sfhs for type ia snrs obtained from resolved stellar populations like the ones we present here should provide an excellent observational constraint on new ideas about type ia progenitors .
this work has benefited from many conversations with a large number of colleagues , including steve bickerton , jeremy goodman , jim gunn , jack hughes , raul jimnez , vicky kaspi , dan maoz , maryam modjaz , evan scannapieco , jerry sellwood , and kris stanek .
we are also grateful to the referee , stephen smartt , for many helpful suggestions that improved the quality of our manuscript .
support for this work was provided by nasa through _ chandra _ postdoctoral fellowship award pf6 - 70046 issued by the _
chandra _ x - ray observatory center , which is operated by the smitsonian astrophysical observatory for and on behalf of nasa under contract nas8 - 0360 .
dz acknowledges support from nasa ltsa grant nng05ge82 g .
jlp is supported by nsf grant ast-0707982 .
[email protected] & dem l71 & 05h 05 m 42s & -67d 52 m 39s & 1.4x1.0 ( h03 ) & @xmath22 ( g03 ) & snr dynamics + [email protected] & n103b & 05h 08 m 59s & -68d 43 m 35s & 0.46 ( b07 ) & @xmath16860 ( r05 ) & light echo + [email protected] & & 05h 09 m 31s & -67d 31 m 17s & 0.48 ( b07 ) & @xmath86 ( b08,r05 ) & light echo + [email protected] & & 05h 19 m 35s & -69d 02 m 09s & 0.52 ( b07 ) & @xmath25 ( r05 ) & light echo + [email protected] & n49 & 05h 26 m 00s & -66d 04 m 57s & 1.4 ( p03 ) & @xmath87 ( v06 ) & snr dynamics + sn 1987a & sn 1987a & 05h 35 m 28s & -69d 16 m 11s & 0.03 ( n08 ) & 22 & sn + [email protected] & n63a & 05h 35 m 44s & -66d 02 m 14s & 1.1 ( w03 ) & @xmath88 ( h98 ) & snr dynamics + [email protected] & n158a & 05h
40 m 11s & -69d 19 m 55s & 1.3x0.7 ( h01 ) & @xmath16800 ( c05 ) & pulsar n49 & ob ( c88 ) , co ?
( g01 ) & unknown ( iip ? , b07 ) & 0.0018 & 6.0 & @xmath89 & @xmath90 & @xmath89 + sn 1987a & sn & 1987-like / iipec & 0.0043 & 8.7 & @xmath27 & @xmath91 & @xmath92 + n63a & ob ( c88 ) & unknown ( ib / c ? , h98 ) & 0.0013 & 6.4 & @xmath89 & @xmath11 & @xmath93 + n158a & co ( k89 ) & iip ( w08 ) or ib / c ( c05 ) & 0.0036 & 7.8 & @xmath11 & @xmath94 & @xmath95 dem l71 & x - ray ( h95 ) & normal ? ( bh09 ) & 0.0022 & 8.3 & 28 & 72 + n103b & x - ray ( h95 ) & bright ? ( bh09 ) & 0.0023 & 8.1 & 73 & 27 + [email protected] & x - ray ( h95 ) & very bright ( b08,r08 ) & 0.0014 & 7.9 & 49 & 51 + [email protected] & x - ray ( h95 ) & bright ( bh09 ) & 0.0032 & 7.7 & 41 & 59 | we use the star formation history map of the large magellanic cloud recently published by harris & zaritsky to study the sites of the eight smallest ( and presumably youngest ) supernova remnants in the cloud : sn 1987a , n158a , n49 , and n63a ( core collapse remnants ) , [email protected] , [email protected] , n103b , and dem l71 ( type ia remnants ) .
the local star formation histories provide unique insights into the nature of the supernova progenitors , which we compare with the properties of the supernova explosions derived from the remnants themselves and from supernova light echoes .
we find that all the core collapse supernovae that we have studied are associated with vigorous star formation in the recent past . in the case of sn 1987a
, the time of the last peak of star formation ( 12 myr ) matches the lifetime of a star with the known mass of its blue supergiant progenitor ( @xmath1 ) .
more recent peaks of star formation can lead to supernovae with more massive progenitors , which opens the possibility of a type ib / c origin for snrs n158a and n63a .
stars more massive than @xmath2 are very scarce around snr n49 , implying that the magnetar sgr 0526@xmath066 in this snr was either formed elsewhere or came from a progenitor with a mass well below the @xmath3 threshold suggested in the literature .
three of our four ia snrs are associated with old , metal poor stellar populations .
this includes snr [email protected] , which is known to have been originated by an extremely bright type ia event , and yet is located very far away from any sites of recent star formation , in a population with a mean age of @xmath4 gyr .
the type ia snr n103b , on the other hand , is associated with recent star formation , and might have had a relatively younger and more massive progenitor with substantial mass loss before the explosion .
we discuss these results in the context of our present understanding of core collapse and type ia supernova progenitors . |
the galactic center , located at a distance of 8 kpc @xcite ( 1@xmath4 corresponds to 0.038 pc ) , is our unique laboratory to study the interactions between a supermassive black hole ( 4 @xmath3 10@xmath7 m@xmath8 , @xcite ) and its surrounding environment .
this region is composed of a very dense and warm interstellar medium ( ism ) , mostly condensed in giant molecular clouds ( gmcs ) .
these gmcs have different characteristics than those gmcs found in the galactic disk . in the galactic center ,
the gmcs are warmer , denser , and much more turbulent .
the dynamical center of the milky way is occupied by a nonthermal compact radio source , the supermassive black hole sgr a*. surrounding this source , three arcs of ionized gas are present ( the western arc , the northern arm and the extended bar , @xcite ) , forming the minispiral or sgr a west .
both sgr a * and the minispiral are surrounded by a ring - like structure , the circumnuclear disk ( cnd ) .
the cnd is composed of a mixture of neutral atomic and molecular gas and dust @xcite .
the cnd has an inner radius of @xmath12 pc @xcite and it extends until @xmath1 5 pc @xcite .
it is tilted with respect to the galactic plane ( @xmath1 70@xmath9 ) and its main motion is rotation around sgr a * @xcite .
the inner cavity that the cnd encloses is devoid of dust @xcite , but it is occupied by ionized gas ( minispiral ) @xcite , neutral molecular gas @xcite and a large amount of neutral atomic gas @xcite . also , inside the cavity , occupying the central parsec , stars in different evolutionary stages
have been found @xcite .
the stars are orbiting sgr a * and they could have been formed in - situ @xcite or spiraled inwards after formation @xcite .
@xcite have previously detected the clumpiness of the cnd , calculating a dynamical lifetime of 10@xmath10
10@xmath0 yr . however , later studies by @xcite , @xcite and @xcite estimate that the density is @xmath1 ( 3
4 ) @xmath3 10@xmath11 @xmath12 .
the density of the clumps would allow them to overcome the tidal shear from sgr a * , increasing the lifetime of the cnd to 10@xmath11 yr .
the galactic center has been studied numerous times in different molecular gas tracers .
@xcite studied the emission from four different ammonia rotation inversion transitions ( nh@xmath6(1,1 ) , ( 2,2 ) , ( 3,3 ) and ( 6,6 ) ) using the very large array ( vla ) .
ammonia is a moderately high - density ( @xmath1 10@xmath10 @xmath12 ) , and high - temperature ( 23 to 412 k , ( 1,1 ) to ( 6,6 ) transitions ) tracer .
these studies showed that the colder gas ( nh@xmath6(1,1 ) and ( 2,2 ) ) is highly extended in this region , showing no strong emission in the cnd .
however , nh@xmath6(3,3 ) ( energy above ground @xmath1 125 k ) is more concentrated towards the center , proving to be much more useful for tracing the cnd , although still not found any closer to the supermassive black hole than the cnd .
nh@xmath6(6,6 ) , on the other hand ( tracing material at @xmath1 412 k ) , is found very close to sgr a * , predominantly inside the cnd , well within the inner cavity , while it is much fainter further away from the cnd .
therefore , very dense and warm molecular gas is located in the inner 1.5 pc of the galaxy .
studies of other high - density tracers have also been carried out . in particular , hcn has proved to be a very useful molecular gas tracer in the galactic center environment , providing important results regarding the structure of the cnd .
@xcite studied hcn(1 - 0 ) emission using the hat creek interferometer , obtaining a 10@xmath4 resolution . the emission from this transition
was detected along the cnd , with the emission in the southern part stronger than the emission detected in the rest of the cnd .
also , a gap was found in the eastern part of the cnd .
high - resolution results were also achieved by @xcite using the berkeley - illinois - maryland array ( bima ) .
however , @xcite noted that self - absorption affects hcn(1 - 0 ) ( and hco@xmath13(1 - 0 ) , which is closely correlated ) due to the intervening molecular material along the line of sight to the galactic center .
this problem was also addressed by @xcite , who published their hcn(1 - 0 ) data from the owens valley radio observatory ( ovro ) millimeter interferometer , with a 5@xmath4 resolution . in these latest results ,
the hcn emission is detected along the cnd and also in various narrow structures outside of it , such as the linear filament . however , because of the self - absorption problem , the use of higher transitions , which will be less sensitive to the cooler and more diffuse gas along the line of sight , should lead to improvements .
hcn(3 - 2 ) emission was detected by @xcite using the iram 30-m telescope and by @xcite using the 15-m james clerk maxwell telescope ( jcmt ) . in both cases ,
emission is detected along the cnd , but it is remarkably stronger towards the south of the cnd .
the detection of hcn(4 - 3 ) emission ( which traces material at @xmath1 25 k , and has a critical density value of @xmath1 10@xmath14 @xmath12 ; @xcite ) in the galactic center region was also reported by @xcite using the jcmt , achieving a 15@xmath4 resolution .
the hcn(4 - 3 ) emission is also stronger towards the south of the cnd , with very weak emission towards the north .
moreover , a dynamical model of a rotating torus developed by this group supported the idea that the northern and the southern parts of the cnd are independent structures ( each of them with a different inclination , e.g. a warped disk ) , following a common rotation pattern around sgr a*. the self - absorption problem is much less evident for these higher - excitation transitions , consistent with the a priori expectation that absorption is produced by lower excitation material along the line of sight .
while these higher hcn lines are much more reliable tracers for studying the structure of the cnd , higher - resolution has not been available until now . the submillimeter array ( sma )
is the first interferometer equipped with 350 ghz receivers , and our maps image the cnd in the higher excitation material .
cs has also been studied in the galactic center in various transitions ( ( 2 - 1 ) , ( 3 - 2 ) , ( 5 - 4 ) and ( 7 - 6 ) @xcite ) .
however , no previous results on ( 7 - 6 ) have been reported in the cnd .
our study images the distribution of cs(7 - 6 ) emission ( which traces material at @xmath1 49 k and has a critical density of @xmath1 10@xmath11 @xmath12 , @xcite ) in the central 4 pc of the milky way .
the nh@xmath6 studies allow a determination of kinetic temperatures .
the hcn and cs studies allow estimates of the h@xmath15 densities .
together , they define the high excitation environment of galactic nuclei .
observations of the hcn(4 - 3 ) ( @xmath16 = 354.5054759 ghz ) and cs(7 - 6 ) transitions ( @xmath16 = 342.8830000 ghz ) were made with the submillimeter array @xcite in the compact configuration on 2005 july 1 and 7 and august 13 and in the subcompact configuration on 2007 may 5 .
we produced a 25-pointing mosaic ( figure [ pointings.fig ] ) covering a @xmath1 2@xmath2 @xmath3 2@xmath2 area , which encompasses sgr a * , the minispiral and the cnd .
the central position of the mosaic was at @xmath17 = 17@[email protected]@xmath20 , @xmath21 = -29@[email protected]@xmath4 , and the rest of the pointings were nyquist sampled at half - beam spacings , effectively increasing the sensitivity of the mosaic ( the primary beam for these frequencies is 36@xmath4 ) .
the data consisted of two 2 ghz bandwidths ( upper sideband and lower sideband , usb and lsb , respectively ) divided into 24 spectral windows , each of them composed of 128 channels with a velocity resolution of 0.7 km s@xmath5 .
the hcn(4 - 3 ) was observed in the usb and cs(7 - 6 ) in the lsb .
the velocity coverage for the usb was from -180 km s@xmath5 to 1490 km s@xmath5 , centered on the hcn(4 - 3 ) line at _
v_@xmath22 = 0 km s@xmath5 and from -1605 km s@xmath5 to 129 km s@xmath5 for the lsb , centered on the cs(7 - 6 ) line at
_ v_@xmath22 = 0 km s@xmath5 .
we averaged every 5 km s@xmath5 since molecular lines in the vicinity of sgr a * are very broad ( fwhm @xmath23 100 km s@xmath5 ; @xcite ) .
therefore the degraded velocity resolution was adequate to resolve the lines .
the continuum subtraction was performed in the uv plane .
the total integration time per pointing was 30 minutes , but because of the distribution of the pointings , the actual time for the central 1.5@xmath2 covering the cnd was 90 minutes .
the phase calibrator was 1733 - 130 ( nrao 530 ) , with a flux value of 0.84 jy for the first two tracks and 0.79 jy for the third and fourth tracks ( the fourth track in the subcompact array ) .
also , during the fourth track we used 1751 + 096 as a phase calibrator as well , with a flux value of 1.75 jy .
the fluxes were determined by the sma project in the immediate period around the experiment .
the data were calibrated using the mir software package developed for the owens valley radio observatory ( ovro ) , which was modified for the sma , and the imaging process was done using both the multichannel image reconstruction image analysis and display ( miriad ) and the nrao astronomical imaging processing system ( aips ) .
to produce the dirty map we used the task invert in miriad , with the systemp option to decrease the weight for visibilities with abnormally high system temperatures .
afterwards , we subtracted the continuum using line - free channels from both sides of the line for the hcn(4 - 3 ) , and only one side for the cs(7 - 6 ) , since the latter line was located close to the edge of the passband in order to be able to observe both molecular lines simultaneously .
we cleaned the map using the task mossdi2 in miriad , cleaning to a 1.5 @xmath24 level .
the integrated intensity maps were produced using the momnt task in aips , with a minimum flux cutoff of 0.5 jy and 0.3 jy for the hcn(4 - 3 ) data and the cs(7 - 6 ) data , respectively .
the flux cutoff was chosen based on the noise threshold , to suppress noise contribution , since for a simple summation the noise will add and narrow line features will become effectively filter diluted . natural weighting of the uv data produced an image with a synthesized beam of 4.6@xmath4 @xmath3 3.0@xmath4 with a position angle of 4.1@xmath9 for the hcn(4 - 3 ) emission and 4.5@xmath4 @xmath3 3.1@xmath4 and a position angle of -4.4@xmath9 for cs(7 - 6 ) .
the final rms per channel is 0.3 jy beam@xmath5 and the overall achieved rms sensitivity for the integrated map is 8.4 jy beam@xmath5 km s@xmath5 ( for hcn(4 - 3 ) ) and 0.2 jy beam@xmath5 and 3.3 jy beam@xmath5 km s@xmath5 ( for cs(7 - 6 ) , respectively ) .
the compact - north ( for low declination sources ) configuration of the sma provides a maximum projected baseline distance of 91 m and a minimum of 11 m , while the subcompact configuration has maximum projected baseline of 25 m and a minimum of 6 m. the combination of data from both configurations improves the detection of the extended emission as the inner uv plane is better sampled , while providing a better resolution because of the longest baselines .
the data obtained using only the compact configuration show a better resolution , but the negative bowls due to the missing extended emission are quite severe .
eventually , the ultimate goal would be to combine these data with single - dish data in order to further suppress the missing flux problems which are still evident in our maps .
we detected the extended emission of hcn(4 - 3 ) within 2 pc of sgr a * ( figure [ hcn4_3.fig ] ) .
the southern part of the cnd is clearly detected , tracing the southwest lobe and the southern extension ( using the same nomenclature as @xcite ) , while the emission from the northern part is more sparse and distributed .
@xcite observed a similar distribution of hcn(4 - 3 ) emission using the james clerk maxwell telescope ( jcmt ) ( 15@xmath4 spatial resolution ) , detecting the strongest emission to the south . due to our higher resolution
image , we can directly determine that the emission is very clumpy around sgr a * , i.e. , the cnd is composed of an array of blobs , in a necklace - like manner , especially towards the north , as was first noted by @xcite .
furthermore , the clumpy appearance is highly irregular , in terms of size scale , intensity , spacing and distance to sgr a*. we compare the hcn(4 - 3 ) flux density detected with the sma to the single - dish flux density detected with the jcmt . after smoothing the sma resolution to match that of the jcmt , we find that the interferometer detects 61% of the emission detected with the single - dish over an area that covers 10 jcmt beams .
the missing flux density would be spread over many sma synthesized beams , which are approximately 16 times smaller in area than the jcmt beam ( i.e. 161 sma synthesized beams ) .
the sma data smoothed to the jcmt resolution ( figure [ smooth.fig ] ) , however , show a very striking similarity to the jcmt emission map ( figure 4 in @xcite ) .
this result suggests that the extended emission is a very low level contribution to the final result ( 21 jy km s@xmath5 per sma synthesized beam , which is at the level of the first contour on figure [ hcn4_3.fig ] , 3@xmath24 ) , and most importantly that both the smoothed sma and the jcmt emission maps are dominated by the clumps .
it is widely accepted that the overall motion of the cnd is rotation around sgr a * , as we have previously mentioned .
we plot the spectra at different peaks ( clumps ) along the cnd to show the kinematics ( fig [ spectra_hcn43.fig ] ) . also , we plot the spectra at different `` empty '' locations where no emission was found close to peculiar - looking spectra to check whether the influence of negative bowls ( i.e. the lack of short spacings in our images since extended structures are not recorded ) have affected the emission and , therefore , the final result .
we find that spectra a and q show the largest velocity shifts , with spectrum a redshifted and spectrum q blueshifted .
spectrum a peaks at @xmath1 105 110 km s@xmath5 and spectrum q at @xmath1 -105
-110 km s@xmath5 , which agree with the values previously reported by @xcite and @xcite .
the remainder of the features peak at intermediate velocities , in an orderly manner , consistent with a rotation pattern in the cnd . with an average radial velocity of 110 km s@xmath5 and a 1.5 pc radius ,
if we assume that the velocity is purely rotational , the rotation period would be @xmath1 8 @xmath310@xmath10 yr .
we can extract much more information from the study of spectra .
in particular , there is a group of features worth observing closely .
firstly , spectrum h is slightly different from the rest . it does not seem to follow the rotation pattern and has a different linewidth , being narrower than the other spectra .
these characteristics suggest that this feature may be the `` 70 km s@xmath5 cloud '' , first reported by @xcite .
it seems to be near the cnd , but has a non - negligible radial velocity component ( @xmath1 50 km s@xmath5 ; @xcite ) , much larger than the rest of the clumps , whose predominant motion is rotation . at the same time , spectrum h shows a shock appearance , with a steep blue side and a non - gaussian red tail . also , clump e shows a remarkably different spectrum , with two velocity components , roughly equal in intensity and an overall broader profile than nearby clumps .
clump e lies to the north of the minispiral northern arm ( figure [ hcn43_cont.fig ] ) , believed to be interacting with the cnd and creating a gap where the ionized gas is infalling towards sgr a * @xcite .
in fact , @xcite consider clump e part of the _ cnd northern arm_. @xcite reported the detection of neutral gas also associated with the minispiral northern arm as far as 3 pc from the dynamical center of the milky way , consistent with a scenario of infalling gas through the cnd gap .
clump g , located in close proximity to clump e , also shows a two - velocity component broad profile .
clump g is part of what @xcite called the _ linear filament _ , a ridge of gas that appears to be connecting the _ western streamer _ and the cnd @xcite , though it might be a projection effect . on the other hand
, we observe a very peculiar ( and broad ) double - peak spectrum in clump n , which also seems to be in the path of the minispiral ( figure [ hcn43_cont.fig ] ) .
@xcite reported the western arc of the minispiral to have the same velocity field as the cnd , therefore suggesting a connection between the molecular ring and sgr a west .
however , clump n might be also affected by a different process .
when comparing spectrum m with spectrum n , we observe that both show two dips at roughly the same velocities ( @xmath1 -60 km s@xmath5 , @xmath1 10 km s@xmath5 ) , even though spectrum m was taken a little farther from the region apparently influenced by the minispiral . also , spectrum l shows that there are strong negative bowls in the proximity of clump n because of the lack of short - spacings , so that these spectra are probably affected by this problem .
we will go back to discuss this spectrum a little later .
clump dd does show a very clear two - velocity component broad profile , and may be related to the interaction with the minispiral .
cloud dd is located at the end of the extended bar of sgr a west ( figure [ hcn43_cont.fig ] ) .
we could be observing a similar behavior as the one detected for clump e. also , absorption is probably not the answer for this feature due to its uniqueness among the nearby clumps .
we would have expected the same profile to be shown by the rest of the adjacent clumps ( such as c , d and aa ) , but that is not the case .
we have not found strong negative bowls in the proximity of clump dd . therefore the lack of short - spacings is probably not affecting the line profile for this clump . as @xcite and @xcite previously remarked , we do detect emission in the eastern and northeastern parts of the cnd , a section that is heavily affected by absorption by foreground cold clouds when using a lower j transition line @xcite .
the emission peaks detected in this region ( d and dd ) are indeed strong , demonstrating that the higher temperature gas in this region can be discerned once the cold gas along the line of sight is suppressed .
spectra bb and cc show very narrow profiles , especially spectrum bb .
they are located outside of the main cnd structure , in what we will call the _ northeast arm _
, therefore farther from the supermassive black hole .
the greater distance from sgr a * implies that the gravitational pull from the center will be less important , and the narrower linewidths may reflect a weaker interaction .
spectrum k shows a peculiar profile as well . observing the profile of spectrum l , closely located
, we infer that clump k is affected by the lack of short - spacings in the data .
a similar behavior seems to be affecting clump z , easily understood when comparing to spectrum y. clumps u , w and q are probably also affected by the negative bowls ( see spectra t , v and o ) , which are much stronger in the southern part of the cnd , indicating that the loss of extended emission is more significant in that region than in the northern part , where only clump a seems to be affected .
spectrum r seems to be especially affected by the short - spacings problem , resulting in a very broad but not very strong spectrum . in summary , the hcn(4 - 3 ) data undersample the extended emission .
this is especially remarkable in the southern part of the cnd .
however , such problem does not prevent the very strong detection of this molecular tracer along the cnd and even further away , in the linear filament and the _ northeast arm_. comparing our data with the hcn(1 - 0 ) data from the owens valley radio observatory by @xcite ( figure [ hcn4_3_hcn.fig ] , both maps convolved to the same resolution and integrated over the same velocity range ) we can see that both transitions coincide tracing the southern part of the cnd , and even in some parts of the northern lobe of the cnd , but hcn(1 - 0 ) is stronger to the north as compared to hcn(4 - 3 ) .
this result suggests that the northern and the southern parts of the cnd have different excitation conditions , with the southern part warmer than the northern part .
the strongest peak in hcn(1 - 0 ) in the _ southwest lobe _ does not coincide with the strongest peak in hcn(4 - 3 ) .
it appears that the lower - excitation line is stronger towards the tip of the cnd , where the most blueshifted material is detected , while hcn(4 - 3 ) is strongest north of that position , in what we have called clump n ( figure [ spectra_hcn43.fig ] ) .
however , hcn(1 - 0 ) in the position of clump
n is affected by self - absorption @xcite .
therefore , the difference in gas distribution could be due to this effect .
we overplot the spectra at the locations of various clumps along the cnd for hcn(4 - 3 ) and hcn(1 - 0 ) from @xcite ( figure [ hcn10_spec.fig ] , convolved to the same resolution ) .
we confirm that hcn(1 - 0 ) suffers from strong self - absorption , a problem especially noticeable in clumps k , n , u , x and w. clump n in both transitions seems affected , but hcn(1 - 0 ) suffers a much more dramatic `` dip '' ( i.e. a sharp absorption feature within the emission profile ) than hcn(4 - 3 ) .
the velocity at which the absorption affects the hcn(1 - 0 ) line ( @xmath1 0 km s@xmath5 , produced by a well known cold cloud complex , the `` local gas '' ; @xcite ) is slightly different from the velocity at which we find the dip in the hcn(4 - 3 ) spectrum . also , the nearby clump k shows a similarly non - gaussian profile , but the rest of the clumps are mostly unaffected .
we also overplot the spectra at the location of sgr a * ( figure [ hcn10_spec.fig ] , where we can observe the absorption features in the hcn(1 - 0 ) data noted by @xcite , which are absent from the hcn(4 - 3 ) data .
therefore , hcn(4 - 3 ) is probably not affected by absorption ( and , if it is , not enough to explain the line profiles ) , but by the lack of short - spacings , a problem that can be solved by combining the interferometric data with single - dish data .
hcn(1 - 0 ) could be missing extended emission as well , since the ovro data have not been combined with single - dish data . therefore , both transitions are probably affected by the short - spacings problem , but only in the case of ( 1 - 0 ) the self - absorption seems worrisome .
the ratio of both transitions ( figure [ ratio43_10.fig ] , where the hcn(1 - 0 ) contours have been overplotted to better explain the ratio ) confirms that the southern part of the cnd is more highly excited than the northern part .
we have calculated the ratio using only the points with flux greater than or equal to 3@xmath24 , and in the case where the flux was lower , we have taken 3@xmath24 as the minimum value .
the reasoning behind this treatment of the data was to assess the properties of the northern part of the cnd . from figure [ hcn4_3_hcn.fig ]
we can observe that the emission from hcn(4 - 3 ) in the northern lobe is fairly extended .
since no emission from hcn(1 - 0 ) was observed in that same location , the ratio can not be calculated without assuming an upper limit .
comparing the more excited molecular gas closer to the center with the less excited gas in the northern part of the cnd would not have been possible .
we have , however , not applied the same rule to the absorption features , leaving them blank , thus the void at the position of sgr a*. the missing flux densities problem does introduce some bias in the ratio , however , it is tipically not too severe and it is only for local spots .
the overplotting of the hcn(1 - 0 ) integrated emission map in figure [ ratio43_10.fig ] demonstrates that except for the _ southern extension _ , most of the regions with a high ratio correspond to the peripheral regions where the hcn(1 - 0 ) flux value has been artificially increased .
this means for the hot regions in the northeast side of the cnd , where the hcn emission is getting fainter , the emission must be hotter than what we are able to show in the ratio map .
where both hcn(1 - 0 ) and hcn(4 - 3 ) have been detected , for example the _ northeast lobe _ , the value of the ratio is one of the lowest values along the cnd .
the ratio map therefore shows the _ southern extension _ as a whole , to be more excited that any other part of the cnd .
however , the absence of detection in the eastern part of the cnd can not rule out the presence of hotter gas in the region , which could possibly be more diffuse .
the large ratios in other parts of the cnd , as in the _ southwest lobe _ , are interspersed among regions with smaller ratios .
this is consistent with heating at the boundaries of clumps .
although all of the southern part of the cnd is really affected by self - absorption ( figure [ hcn10_spec.fig ] ) , this is an interesting result since it suggests that the _ southern extension _ is formed in its entirety by more excited gas than other parts of the cnd .
the question arises , then , why is the southeastern part of the cnd more excited ? what is the heating mechanism ?
we will address these questions later , with the help of an extra comparison tool , a high - temperature molecular tracer , nh@xmath6 . at the same time
, the ratio map , especially in the northern part of the cnd , suggests that the ratio increases as the gas approaches the center , i.e. the inner part of the cnd is more excited than the outer part .
@xcite suggested that the uv radiation from the nuclear cluster ( which is located in a cavity void of dust , and therefore transparent to the uv radiation ) could be responsible for the heating of the molecular gas composing the cnd .
the inner edge of the molecular ring would be heated by this radiation , but as the distance from the center increases , the radiation starts to be absorbed by the dust also composing the ring .
thus , the outer parts of the cnd would be less affected by this radiation .
finally , it is interesting to remark that the location of clump e , where we found a broad double - peak profile ( figures [ spectra_hcn43.fig ] and [ hcn10_spec.fig ] ) shows a large ratio value .
this result suggests that clump e has a higher excitation than the surrounding environment .
the interaction with the northern arm of the minispiral may be a possible mechanism . a detailed kinematic study of the velocity distribution in the minispiral is needed to fully elucidate this situation , possibly .
when comparing the hcn(4 - 3 ) emission map with emission maps produced by nh@xmath6 ( ( 3,3 ) and ( 6,6 ) , @xcite and @xcite , respectively ) we can gain a better understanding of the spatial distribution of the high - density molecular gas in the inner 2 pc of the galactic center .
nh@xmath6 has a lower critical density ( 10@xmath0 @xmath12 ) as compared to hcn , but traces gas at higher temperatures .
both nh@xmath6(3,3 ) and ( 6,6 ) have remarkably different distributions in the central region of the milky way as shown by @xcite and @xcite .
a comparison of hcn(4 - 3 ) to nh@xmath6(3,3 ) ( figure [ hcn_33.fig ] ; hcn(4 - 3 ) smoothed to match the nh@xmath6(3,3 ) resolution ) shows that the emission distributions from these two molecular lines are poorly correlated . however , both lines trace the western side of the cnd and show the strongest peak along the cnd at the same position on the _ southwest lobe_. hcn(4 - 3 ) and nh@xmath6(3,3 ) emission coincide in the _ eastern arm _ as well .
it is worth noting that the _ southern streamer _ as traced by nh@xmath6(3,3 ) ( at @xmath17 = 17@xmath1845@xmath1943@xmath20 and @xmath21 = -29@xmath901@xmath240@xmath4 ) seems to connect to the eastern side of the cnd as traced by hcn(4 - 3 ) .
@xcite reported that no kinematic connection could be made between the _ southern streamer _ and the cnd , but at the same time , it was clear that there was a change in the nh@xmath6(3,3 ) emission when ( at least in projection ) it reached the cnd
. therefore , the spatial alignment at the cnd might only be a projection effect along the line of sight , but a physical connection could not be completely ruled out because of the resolution of the nh@xmath6(3,3 ) data .
we do indeed observe the same trend .
the gas traced by both nh@xmath6(3,3 ) and hcn(4 - 3 ) in the eastern part do not overlap , but it seems that the emission from nh@xmath6(3,3 ) does become weaker in the vicinity of the cnd . at present sensitivity , the kinematic and structural information still can not elucidate whether there is a connection between the _ southern streamer _ and the cnd .
the comparison of hcn(4 - 3 ) and nh@xmath6(6,6 ) ( figure [ hcn_66.fig ] ) shows a good correspondence in the eastern and northern parts of the cnd , and even in the _
northeast arm_. however , we can see that nh@xmath6(6,6 ) is detected inside the central cavity of the cnd , apparently approaching sgr a * , while hcn(4 - 3 ) is only detected along the cnd .
this result suggests that the molecular gas that composes the cnd is in general colder and perhaps denser than the gas that is located inside of the cavity . by overplotting the hcn(4 - 3)/hcn(1 - 0 ) ratio and the nh@xmath6(6,6 ) integrated emission we can make an interesting comparison ( figure [ 66_ratio.fig ] ) .
we observe that the southeastern part of the cnd is located south of the strongest peak detected in nh@xmath6(6,6 ) .
it looks like the material traced by nh@xmath6(6,6 ) is `` following '' the path marked by the ratio towards sgr a*. since the large value of the ratio in the southeastern part of the cnd indicates the presence of a larger amount of highly - excited material than of low - excited material , it means that the majority of the molecular gas in that region is warmer . because nh@xmath6(6,6 ) traces very warm gas and it penetrates from the eastern side of the cnd towards the supermassive black hole , the southeastern side of the cnd could be becoming warmer as it flows northwest , towards sgr a*. at the same location where the gas traced by nh@xmath6(6,6 ) heads for sgr a * , the value of the ratio changes radically , as if suddenly the level of excitation in the gas tracing the cnd drops . at that location hcn(1 - 0 )
could be slightly affected by self - absorption ( spectrum z in figure [ hcn10_spec.fig ] ) , but the very low value of the ratio indicates that the amount of hcn(4 - 3 ) is also low , otherwise the ratio would have been large .
the question remains as to why the material is more highly excited precisely in the _
southern extension_. when plotting the hcn(4 - 3)/hcn(1 - 0 ) ratio with the nh@xmath6(3,3 ) integrated intensity map from @xcite ( figure [ 33_ratio.fig ] ) , we observe that the region where the _ southern streamer _ seems to reach the cnd coincides with the location of the northernmost part of the high ratio , where nh@xmath6(6,6 ) becomes stronger before heading northwest . consequently , in terms of projection , there is a location in the southeastern part of the cnd where the _ southern streamer
_ traced by nh@xmath6(3,3 ) , the strongest peak in the nh@xmath6(6,6 ) emission and the edge of the high - ratio area all coincide .
the _ southern streamer _ may impact upon the cnd producing destabilized material which infalls towards the supermassive black hole .
this material may become so highly excited that hcn(4 - 3 ) energy levels may be depopulated so that a higher hcn transition will be needed to trace the infalling gas .
a comparison of spectra of both the hcn(4 - 3 ) and the nh@xmath6 data is needed to determine whether the material traced by these transitions is the same . in order to accomplish this study
we take the spectra at various locations ( figures [ 33_spec.fig ] and [ 66_spec.fig ] ) . from figure [ 33_spec.fig ] we find that the material tracing the _ southwest lobe _ and the _ northeast arm _ ( peaks 1 and 3 ) kinematically coincides both in hcn(4 - 3 ) and nh@xmath6(3,3 ) , but
the same can not be said about peak 2 , located on the region where the material from the `` 20 km s@xmath5 gmc '' approaches the cnd . however
, we do detect a small peak around 20 km s@xmath5 , weaker than the peak detected around -50 km s@xmath5 , but nonetheless present , which indicates the detection of material from the _ southern streamer_. we observe a similar situation in figure [ 66_spec.fig ] when comparing hcn(4 - 3 ) and nh@xmath6(6,6 ) .
the same material is detected by hcn(4 - 3 ) and nh@xmath6(6,6 ) in the _ southwest lobe _ and the _ northeast arm _
( peaks 1 and 4 ) but not in the regions where the material approaches sgr a * ( peaks 2 and 3 ) .
therefore , the material detected in the northern and western parts of the cnd by hcn(4 - 3 ) and nh@xmath6(3,3 ) and ( 6,6 ) coincides , indicating that the denser and colder gas is heavily mixed with the more diffuse and warmer gas at those locations . however , the material detected in the eastern part of the cnd appears to be very different , with the gas approaching the black hole warmer and more diffuse than the gas tracing the cnd and a possible interaction between the gas from the _ southern streamer _ and the gas forming the cnd in the southeasternmost part of the ring .
these results seem to support the previously noted conclusion , where the material in the _ southern extension _ is warmed and pushed towards the dynamical center of the milky way by the action of the gas coming from the `` 20 km s@xmath5 gmc '' .
we employed a correlator setup to sample the cs(7 - 6 ) line at the same time . because of the locations of the two lines , cs(7 - 6 ) had to be placed at the edge of the passband and the velocity coverage was not as broad for this line , from -150 to 128 km s@xmath5 .
as previously noted , this cs transition traces gas at even higher temperature than the observed hcn line but at a slightly lower density .
we clearly detect and resolve the cs(7 - 6 ) emission in the southern part of the cnd , but it is much weaker towards the northern part , although it is clearly detected in the _ northeast arm _ ( figure [ cs7_6.fig ] ) .
this result is consistent with the hcn results in suggesting that the northern and the southern parts of the cnd have different excitation levels , since the warmer gas is absent from the north .
the results for the cs(7 - 6 ) and hcn(4 - 3 ) lines show that these two molecualr lines correlate quite well with each other ( figure [ cs7_6_hcn4_3.fig ] ) .
the emission peaks coincide in the southern part of the cnd , and even in the few places where cs(7 - 6 ) is found in the northern part of the cnd .
also , if we compare the spectra at the same positions ( figures [ spectra_hcn43.fig ] and [ cs_spectra.fig ] , cs(7 - 6 ) spectra plotted in the same velocity range as hcn(4 - 3 ) spectra ) , we observe that both molecular lines have similar line profiles ( fwhm @xmath23 35 km s@xmath5 ) , suggesting that these lines are tracing the same material . however , there are some differences between the two molecular tracers .
first , we observe that hcn(4 - 3 ) is much stronger in the _ southern extension _ than cs(7 - 6 ) . while both molecular tracers seem equally strong along the _ southwest lobe _ , even coinciding in the distribution of the gas , stronger in the northern part of the lobe , at the location of clump n , than in the southern part , cs(7 - 6 ) is obviously weaker in the _
southern extension_. this result would suggest that the _ southern extension _ is composed by denser but colder gas than the _ southwest lobe_. however , we noted in the previous section that the _ southern extension _ is highly excited .
also , because of the strong detection of nh@xmath6(6,6 ) in the northern region of the _ southern extension _ , towards sgr a * , we acknowledge the presence of warm material in the area .
therefore , the combination of the result drawn by the previous section and the picture presented by the weak emission of cs(7 - 6 ) in the _ southern extension _ , which traces warmer but less dense material than hcn(4 - 3 ) , suggests a complicated morphology in the eastern part of the cnd . the gas is denser and colder towards the south , and becomes much more diffuse and warmer heading north .
this point will be discussed in detail later .
another difference between the integrated emission from both molecular lines is that clump e in cs(7 - 6 ) does not show a double - peak profile , unlike its counterpart in hcn(4 - 3 ) .
the detection is much weaker , which could account for the lack of coincidence .
spectrum n shows a two - velocity component , consistent with the result in hcn(4 - 3 ) . the line profile , however , is not as wide .
clump n in the case of cs(7 - 6 ) is located a little further west than in the case of hcn(4 - 3 ) . since we have considered the interaction with the minispiral as probably affecting the profile of clump n in hcn(4 - 3 ) , it could be that the greater distance from clump n in cs(7 - 6 ) to sgr a west and a consequently weaker interaction is responsible for the narrower profile ( compare figures [ cs_cont.fig ] and [ hcn43_cont.fig ] ) . in order to check this conjecture , we extracted a spectrum closer to the inner side of the cnd , 4.4@xmath4 east and 0.8@xmath4 south of clump n in cs(7 - 6 ) ( at the location of clump n in hcn(4 - 3 ) ) and overplotted the two spectra ( figure [ clumpn.fig ] ) .
the broadening in the profile of the spectrum closer to the minispiral is remarkable .
we conclude that the interaction with the minispiral is probably affecting the line profile . at the same time
, the absorption features detected in these spectra can be caused by the negative bowls .
when we examine the spectra l and m , taken nearby , we see that the lack of short - spacings may well affect the spectra towards clump n. clump k presents a very surprising profile , unlike any other
. it is clearly redshifted and very narrow , more so than the `` 70 km s@xmath5 cloud '' , which is barely detected as clump h , but the negative bowls might be seriously affecting it if we pay attention to spectrum l , taken nearby .
nonetheless , the profile is very unexpected and it is probably worth a deeper study .
the comparison of the spectrum in the same position in hcn(4 - 3 ) ( figure [ clumpk.fig ] ) shows a similar profile .
is once again the presence of the minispiral affecting the line profiles in the _ southwest lobe _ ?
both spectra seem to be suffering from absorption around + 50 km s@xmath5 , a situation especially noticeable in the hcn(4 - 3 ) spectrum . however , this absorption feature is not present anywhere else in the cnd .
clump dd produces a very remarkable spectrum , with two clearly separated peaks .
the spectrum at the same position in hcn(4 - 3 ) shows this profile as well .
there could be some interaction involved in this location as absorption does not seem to be responsible for the `` dip '' and the negative bowls detected in the vicinity because of the lack of short - spacings are not so prominent .
as we mentioned in the previous section , interaction with the extended bar of the minispiral is the most probable explanation of this broad profile .
spectrum r shows a double peak profile and , observing the important negative bowls nearby ( spectrum t ) it is very probable that the short - spacings problem is the cause of this double peak profile .
this problem is probably also affecting clumps ff , u and w , if we compare with spectra t and v. therefore , as in the case of hcn(4 - 3 ) , the lack of short - spacings is affecting the data .
however , because of the lack of single - dish observations of cs(7 - 6 ) in the galactic center , we can not calculate the amount of missing flux .
when we compare the cs(7 - 6 ) detected emission with the 6 cm continuum emission from @xcite , we observe that the gap in the cnd in the cs(7 - 6 ) emission towards the north coincides with the position of the northern arm of the minispiral , with clumps e and f delimiting the gap to the west and clump d to the east ( figure [ cs_cont.fig ] ) . the ratio between cs(7 - 6 ) and hcn(4 - 3 ) ( calculated using only the pixels with flux @xmath25 3@xmath24 ) is useful to better understand the distribution of the dense material ( figure [ cs_ratio.fig ] ) .
we can observe that the ratio increases towards the inner edge of the cnd , especially in the western part .
coincidentally , the western arc of the minispiral is observed in the same region ( figure [ cs_cont.fig ] ) . both cs(7 - 6 ) and hcn(4 - 3 )
have similarly high critical densities , with hcn(4 - 3 ) tracing slightly denser gas than cs(7 - 6 ) , but the temperature traced by cs(7 - 6 ) is double the temperature traced by hcn(4 - 3 ) ( 49 k vs. 25 k ) .
we could interpret the ratio distribution as being ruled mostly by the temperature , although the density should also be a factor .
therefore the molecular gas in the inner side of the ring is probably warmer and could be less dense . as we mentioned before
, the temperature increase can be due to the absorption of uv radiation emitted by the nuclear stellar cluster .
high - ratio values found to the west , outside of the cnd , can also be related to the presence of higher - temperature gas , as the _ western streamer _ is very clearly detected in nh@xmath6(3,3 ) ( tracing gas at @xmath1 125 k ) .
the comparison of cs(7 - 6 ) and nh@xmath6(3,3 ) ( figure [ cs7_6_33.fig ] ) , cs(7 - 6 ) smoothed to match the nh@xmath6 resolution , nh@xmath6(3,3 ) data from @xcite ) shows that both molecular tracers coincide at the location of the strongest peak within the cnd ( as was also noted regarding hcn(4 - 3 ) ) and the _
northeast arm_. however , the coincidences stop there , with cs(7 - 6 ) and nh@xmath6(3,3 ) having in general very different distributions in the region observed in both tracers .
the nh@xmath6(3,3 ) map by @xcite is much larger than the cs(7 - 6 ) map , and it is not represented here in its entirety .
the correlation between the high - ratio values that we found previously west of the cnd and the _ western streamer _ can not be checked in this figure because of the lower resolution of the nh@xmath6(3,3 ) data , and the small clumps detected in cs(7 - 6 ) have been smoothed out . when comparing cs(7 - 6 ) and nh@xmath6(6,6 ) we observe that the eastern part of the cnd and the _ eastern arm _ are the only regions where both molecular tracers overlap and are well correlated ( figure [ cs7_6_66.fig ] ) .
the comparison is similar to that obtained with hcn(4 - 3 ) and nh@xmath6(6,6 ) , except for the northern part of the cnd , which is barely detected in cs(7 - 6 ) .
once again , we note that nh@xmath6(6,6 ) is detected inside the central cavity , whereas cs(7 - 6 ) is limited to the cnd .
this result is consistent with the idea that the cnd is better traced by higher - density and lower - temperature tracers since only a minimum part of the cnd is traced by nh@xmath6 , while cs ( and hcn ) are detected along most of the structure .
similarly to the way we proceeded in the previous section , we overplot the nh@xmath6(6,6 ) map and the cs(7 - 6)/hcn(4 - 3 ) ratio ( cs(7 - 6)/hcn(4 - 3 ) ratio convolved to the nh@xmath6(6,6 ) resolution , figure [ 66_ratiocs.fig ] ) .
unlike in the case of the hcn(4 - 3)/hcn(1 - 0 ) ratio map , the cs(7 - 6)/hcn(4 - 3 ) ratio does not correlate with the nh@xmath6(6,6 ) emission .
the highest ratio values , as mentioned before , are found in the western side of the cnd , overlapping with the western arc of the minispiral ( when higher - resolution is utilized since the nh@xmath6(6,6 ) resolution is not sufficient to observe such details ) .
if the cs(7 - 6)/hcn(4 - 3 ) ratio is tracing the region with the highest temperature - lowest density combination , then the western part of the cnd is either warmer or less dense or both . since nh@xmath6(6,6 ) is detected more strongly in the eastern part of the cnd , it indicates the presence of high - temperature gas in this region of the cnd .
therefore , the western part of the cnd may be less dense than the eastern part , and the high - ratio could be in this case a sign of lower density instead of higher temperature . the material detected by the high - temperature tracers might be unrelated to the material observed in high - density but lower - temperature tracers .
this case is indeed true in the eastern part of the cnd , as proved by the comparison of the hcn(4 - 3 ) and nh@xmath6(6,6 ) spectra , showed previously in figure [ 66_spec.fig ] .
however , the kinematic study indicates the coincidence of the material traced in the western part of the cnd by both hcn(4 - 3 ) and nh@xmath6(6,6 ) . at the same time
, the strength of the detection of the warmer gas is much smaller in the western part of the cnd than in the eastern part .
this indicates that the warmer and more diffuse gas is well mixed with the colder and denser gas in the western part , but not in the eastern part , where the amount of warmer gas is much larger , although not in the totality of the _ southern extension _ where it is only detected in the northernmost region . the lack of strong hcn(1 - 0 ) emission in the _ southern extension _
, tracing the same material as hcn(4 - 3 ) ( figures [ hcn4_3_hcn.fig ] and [ ratio43_10.fig ] ) , but characterized by a lower temperature and critical density , indicate a higher temperature and density than in the _ southwest lobe_. furthermore , if a transition defined by a lower temperature and critical density , such as hcn(1 - 0 ) , coincides in its spatial distribution in the southern part of the cnd with a transition characterized by a higher - temperature but also lower critical density , cs(7 - 6 ) ( both of them compared to hcn(4 - 3 ) ) , the _ southern extension _ will be then denser than the southwestern part of the cnd . since the material coming from the _ southern streamer _ is approaching the cnd in the region of the _ southern extension _ , if the latter structure appears as denser it may be related with the possible interaction previously mentioned between the cnd and the _
southern streamer_. because of this interaction , the material forming the southeastern part of the cnd may be undergoing a compression process , appearing as denser than the adjacent region .
in the previous section we noted the clumpy nature of the cnd . as mentioned , @xcite and @xcite found that the different clumps ( or cores ) within the cnd were dense enough to overcome the tidal shear produced by the central supermassive black hole , and therefore the cnd might be a more stable structure than previously thought @xcite .
for our results , we use the virial calculation to estimate the required masses and densities for a clump to be stable against tidal shear .
we consider then , the clumps to be gravitationally bound against the motions of the gas in the clump , with uniform density and where optical depths do not play a significant role in line - broadening @xcite .
@xmath26 where @xmath27_v@xmath28 _ is the fwhm velocity linewidth and _ r _ is the radius of the clump .
the results are displayed in tables [ tab : hcn ] and [ tab : cs ] .
table [ tab : hcn ] shows the masses measured for the clumps marked in figure [ hcn4_3.fig ] and table [ tab : cs ] displays the results for the clumps in figure [ cs7_6.fig ] .
the virial masses calculated by @xcite range from 2 @xmath3 10@xmath29 to 89 @xmath3 10@xmath29 m@xmath8 , while the results obtained by @xcite are slightly smaller ( 3 @xmath3 10@xmath29 to 45 @xmath3 10@xmath29 m@xmath8 ) .
our measurements are similar , although a little bit higher than these previous results(4 @xmath3 10@xmath29 to 595 @xmath3 10@xmath29 m@xmath8 ) .
we have not calculated the masses for all the clumps due to some difficult non - gaussian profiles ( like clump r ) , that prevented us from being able to calculate a linewidth .
the mean virial density of the clump , assuming spherical symmetry , can be calculated with : @xmath30 therefore , the internal virial density of the clump is : @xmath31 the critical density for a clump to survive the tidal shear from a 4 @xmath3 10@xmath7 m@xmath8 black hole is calculated using the model from @xcite , which assumes the mass distribution in the inner region of the milky way to be described by a spherically symmetric approximation : @xmath32 where d is the radius from the center of the galaxy ( i.e. distance to sgr a * ) . the critical density for a clump to be tidally stable is then : @xmath33(cm^{-3 } ) } \label{eqn : criticaldensity}\ ] ] for a 1.5 pc distance , using equation [ eqn : criticaldensity ] , we can calculate the critical density to be 1.4 @xmath3 10@xmath11 @xmath12 .
therefore , we can conclude that the clumps we have detected using both hcn(4 - 3 ) and cs(7 - 6 ) are tidally stable , as it can be seen in tables [ tab : hcn ] and [ tab : cs ] .
@xcite reported that the total ionized gas of the minispiral is 10@xmath34 m@xmath8 , and @xcite measured a neutral gas mass in the northern arm of the minispiral of 3 @xmath3 10@xmath34 m@xmath8 .
we can then assume the minispiral with a total mass of @xmath1 4 @xmath3 10@xmath34 m@xmath8 .
at the same time , kinematic models by @xcite , using an infalling gas velocity of half the rotation velocity ( mentioned previously as 110 km s@xmath5 ) , were able to explain the majority of the characteristics of the minispiral .
therefore , the infalling time should be @xmath1 3 @xmath3 10@xmath10 yr .
inflow models support the idea of the minispiral as being formed by a cloud that lost angular momentum , probably due to cloud - to - cloud collisions and then fell towards the center .
while infalling , the uv radiation from the nuclear stellar cluster dissociated the molecular gas and later ionized the neutral atomic gas @xcite .
consequently , if the material composing sgr a west comes directly from the cnd , which has a total molecular mass of 1.3 @xmath3 10@xmath7 m@xmath8 ( calculated with the hcn(4 - 3 ) results ) , and the current mass of the minispiral takes 3 @xmath3 10@xmath10 yr to reach the center , if a mere 10% of the cnd is stripped off and becomes part of the minispiral ( as @xcite considered ) , the lifetime of the cnd is @xmath1 9 @xmath310@xmath7 yr , much longer that its rotation period ( 8 @xmath3 10@xmath10 yr ) . therefore , the cnd is not a transient structure , since the amount of time needed for the cnd to disappear , `` swallowed '' by the inner cavity , is longer than the amount of time needed to circle it .
however , @xcite found the molecular gas approaching sgr a * , as traced by nh@xmath6(6,6 ) , not to follow the minispiral path .
consequently , the previous lifetime value can be an overestimation , since not all the molecular gas approaching the dynamical center of the milky way is contained in the minispiral .
nonetheless , if the cnd is a non - transient structure , the clumpiness should be expected to disappear and the cnd to become a homogeneous structure
. however , the observations do not show such homogeneity , instead , the clumps present a wide range of velocities .
this dispersion in velocities could result in various interactions between the clumps ( cloud - to - cloud collisions ) , with the more diffuse material spiraling inwards towards the center .
finally , recent studies by @xcite showed that the accretion rate of sgr a * is between 2 @xmath3 10@xmath35 and 2 @xmath3 10@xmath36 m@xmath8 yr@xmath5 . if the infalling rate is 1.5 @xmath3 10@xmath37 m@xmath8 yr@xmath5 ( considering for now only the material confined in sgr a west ) , there is clearly a surplus of material inside the cavity that is not accreting into the black hole .
the remaining material could be undergoing star formation processes .
@xcite reported that 10@xmath29 - 10@xmath10 m@xmath8 in the inner cavity became stars around 10@xmath7 years ago . with the infalling rate that we are considering ,
an accumulation of 10@xmath10 m@xmath8 will take @xmath1 7 @xmath3 10@xmath0 yr ( and even less if we consider the material traced by nh@xmath6(6,6 ) ) .
therefore , the material infalling towards sgr a * could be forming stars in the inner cavity , and not only accreting into the black hole .
we have successfully detected hcn(4 - 3 ) and cs(7 - 6 ) within 2 pc of sgr a * , effectively tracing the cnd .
we demonstrate that the higher hcn transition minimizes the self - absorption problem observed in lower transitions @xcite , therefore providing a much more reliable sampling of the kinematics and structure of the cnd .
the emission from both molecular tracers , hcn(4 - 3 ) and cs(7 - 6 ) , appears in clumps , forming a necklace - like structure around cnd .
the clumps have various sizes , from @xmath1 3@xmath4 to @xmath1 13@xmath4 . in the case of hcn(4 - 3 ) , the emission detected from these clumps amounts to 61% of the single - dish flux detected by @xcite .
problems due to missing short spacings , however , continue to affect the observed kinematics .
nevertheless , when smoothing our data to match the resolution of the jcmt results from @xcite , we can easily note the remarkable similarity between the smoothed ( to a 15@xmath4 resolution ) hcn(4 - 3 ) emission map and the hcn(4 - 3 ) integrated emission map from @xcite .
this result suggests that the extended emission does not greatly contribute to the final result and the molecular gas is mostly concentrated in clumps .
therefore , the missing short spacing problem probably does not seriously affect the final morphological results .
moreover , we confirm the stability of the cnd based on the density of the clumps , which is large enough for the clumps to overcome the tidal shear produced by the supermassive black hole . furthermore , the lifetime value of the cnd is far longer than the rotation value .
this result supports the conclusions of @xcite and @xcite , agreeing that the cnd is not a transient structure .
both hcn(4 - 3 ) and cs(7 - 6 ) are found to be much more abundant in the southern part of the cnd than in the northern part of the cnd . in order to determine the geometrical distribution of the gas , the location of the colder and warmer gas , as well as the denser and more diffuse gas , and the implications of such distribution , line ratio measurements have been used .
our results indicate that the northern and the southern parts of the cnd have different excitation levels , as well as different densities , with the southern part of the cnd warmer and denser than the northern part .
also , comparing our results with those from nh@xmath6 , a high - density , high - temperature molecular tracer , we conclude that the molecular gas forming the cnd is denser and colder than the molecular gas inside the inner cavity .
more precisely , the southeastern part of the cnd is denser than the southwestern part .
however , the southeastern part seems to become more diffuse or warmer as the material heads northwest , approaching the supermassive black hole , as detected by nh@xmath6(6,6 ) .
the comparison of the linewidths supports this conclusion , since both the nh@xmath6(6,6 ) and hcn(4 - 3 ) present similar line profiles along the cnd , kinematically tracing the same material , except in the region where the nh@xmath6(6,6 ) emission is detected closer to sgr a * , where it shows a clear line - broadening effect , absent from the hcn(4 - 3 ) profiles .
this result indicates a probable infall of the material forming the cnd towards the dynamical center through the eastern part of the ring - like structure .
however , the material detected inside the cnd by the nh@xmath6(6,6 ) is hotter and denser than the material forming the minispiral , and both structures appear to be unrelated , as indicated by @xcite . at the same time , the detection of an interaction between the _ southern streamer _ , as traced by nh@xmath6(3,3 ) , and the cnd indicates that the material forming the ring may be undergoing a compression process due to the material approaching the cnd before the gas spirals inward towards the center .
therefore , the gas detected in the southeasternmost part of the cnd may be pushed towards the northwest and heated in the process , explaining the lack of hcn detection .
in addition , we have detected a correlation between the minispiral and the cnd .
line - broadening has been detected in the spectra of the clumps that spatially coincide with the arcs of the minispiral .
we have observed this phenomenon in the eastern , western and northern parts of the cnd .
the western arc of the minispiral and the inner western part of the cnd seem to overlap .
also , we have observed that the eastern end of the extended bar of the minispiral spatially coincides , at least in projection , with the location of a clump characterized by a broad spectrum .
a similar situation has been observed regarding the northern part of the cnd and the northern arm of the minispiral , where line - broadening has also been found in the spectra of the clumps seemingly located at the very end of the northern arm .
furthermore , we have detected a gap in the northern region of the cnd , which the northern arm of sgr a west appears to be traversing .
these results suggest that the cnd and the minispiral may be physically related .
sgr a west could be gas which has been stripped from the cnd . alternatively , the spiral arms could be features infalling from beyond the cnd . while more detailed kinematics are needed , we detect the influence of the minispiral on the western inner part , the northern part , and the eastern part of the cnd , where we can see line - broadening in the vicinity of sgr a west . finally , we have observed that the inner edge of the cnd seems more highly - excited than the outer part of the ring , as traced by the ratio of hcn(4 - 3)/hcn(1 - 0 ) .
most probably the nuclear stellar cluster is responsible for the excitation of the inner side of the cnd .
we thank m. christopher for providing the hcn(1 - 0 ) data we have used for comparison .
we also thank the referee , t. wilson , for his helpful suggestions to improve the manuscript . during the development of this study , m.m .- c . has been supported by a smithsonian institution visiting student grant and an academia sinica institute of astronomy and astrophysics fellowship .
becklin , e.e . ,
gatley , i. & werner , m.w .
1982 , , 258 , 135 choi , m. , evans ii , n.j . , tafalla , m. & bachiller , r. 2000 , , 538 , 738 christopher , m.h . , scoville , n.z . ,
stolovy , s.r .
& yun , m.s . 2005 , , 622 , 346 coil , a.l .
& ho , p.t.p .
1999 , , 513 , 752 coil , a.l . & ho , p.t.p .
1999 , , 533 , 245 gatley , i. , jones , t.j . ,
hyland , a.r . ,
beattie , d.h .
& lee , t.j . 1984 , , 210 , 565 genzel , r. , watson , d.m . ,
crawford , m.k .
& townes , c.h .
1985 , , 297 , 766 ghez , a.m. , salim , s. , hornstein , s.d . ,
tanner , a. , lu , j.r . ,
morris , m. , becklin , e.e . & duchne , g. 2005 , , 620 , 744 grkin , m.a .
& rasio , f.a .
2005 , , 628 , 236 gsten , r. , genzel , r. , wright , m.c.h . ,
jaffe , d.t . ,
stutzki , j. & harris , a.i .
1987 , , 318 , 124 harris , a.i . ,
jaffe , d.t . , silber , m. & genzel , r. 1985 , , 294 , l93 herrnstein , r.m .
& ho , p.t.p .
2002 , , 579 , l83 herrnstein , r.m .
& ho , p.t.p .
2005 , , 620 , 287 ho , p.t.p . , moran , j.m . & lo , k.y .
2004 , , 616 , l1 httemeister , s. , wilson , t.l . ,
bania , t.m . & martn - pintado , j. 1993 , , 280 , 255 httemeister , s. , dahmen , g. , mauersberger , r. , henkel , c. , wilson , t.l . & martn - pintado , j. 1998 , , 334 , 646 jackson , j.m . , geis , n. , genzel , r. , harris , a.i . ,
madden , s. , poglitsch , a. , stacey , g.j .
& townes , c.h .
1993 , , 402 , 173 jackson , j.m .
, paglione , t.a.d . ,
carlstrom , j.e . & rieu , n .- q .
1995 , , 438 , 695 krabbe , a. , genzel , r. , drapatz , s. & rotaciuc , v. 1991 , , 382 , l19 lapinov , a.v . , schilke , p. , juvela , m. & zinchenko , i.i .
1998 , , 336 , 1007 lugten , j.b . ,
genzel , r. , crawford , m.k .
& townes , c.h .
1986 , , 306 , 691 marrone , d.p . ,
moran , j.m . , zhao , j .- h .
& rao , r. 2007 , , 654 , 57 marshall , j. , lasenby , a.n . &
harris , a.i .
1995 , , 277 , 594 mcgary , r.s . , coil , a.l .
& ho , p.t.p .
2001 , , 559 , 326 mezger , p.g .
, zylka , r. , salter , c.j . ,
wink , j.e .
, chini , r. , kreysa , e. & tuffs , r. 1989 , , 209 , 337 oort , j.h .
1977 , ara&a , 15 , 295 paumard , t. , genzel , r. , martins , f. , nayakshin , s. , beloborodov , a.m. , levin , y. , trippe , s. , eisenhauer , f. , ott , t. , gillesen , s. , abuter , r. , cuadra , j. , alexander , t. & sternberg , a. 2006 , , 643 , 1011 reid , m.j .
1993 , ara&a , 31 , 345 roberts , d.a .
& goss , w.m .
1993 , , 86 , 133 rohlfs , k. & wilson , t.l .
2000 , tools of radio astronomy , fourth edition , springer - verlag sanders , r.h .
1998 , , 294 , 35 schdel , r. & eckart , a. 2005 , memorie della societ astronomica italiana , 76 , 65 serabyn , e. , gsten , r. & evans , n.j .
1989 , in iau symp .
136 , the center of the galaxy , ed .
m. morris ( dordrecht : kluwer ) , 417 serabyn , e. & lacy , j.h .
1985 , , 293 , 445 serabyn , e. , lacy , j.h . &
achterman , j.m .
1992 , , 395 , 166 shukla , h. , yun , m.s . &
scoville , n.z .
2004 , , 616 , 231 takakuwa , s. , ohashi , n. , bourke , t.l . ,
hirano , n. , ho , p.t.p .
, jrgensen , j. , kuan , y .- j . ,
wilner , d.j .
& yeh , s.c.c . 2007 , , 662 , 431 vollmer , b. & duschl , w.j .
2000 , newa , 4 , 581 wright , m.c.h . , coil , a.l . , mcgary , r.s .
, ho , p.t.p .
& harris , a.i . 2001 , , 551 , 254 yusef - zadeh , f. & morris , m. 1987 , , 320 , 545 | we present the first large - scale mosaic performed with the submillimeter array ( sma@xmath0 ) in the galactic center .
we have produced a 25-pointing mosaic , covering a @xmath1 2@xmath2 @xmath3 2@xmath2 area around sgr a*. we have detected emission from two high - density molecular tracers , hcn(4 - 3 ) and cs(7 - 6 ) , the latter never before reported in this region .
the data have an angular resolution of 4.6@xmath4 @xmath3 3.1@xmath4 , and the spectral window coverage is from -180 km s@xmath5 to 1490 km s@xmath5 for hcn(4 - 3 ) and from -1605 km s@xmath5 to 129 km s@xmath5 for cs(7 - 6 ) .
both molecular tracers present a very clumpy distribution along the circumnuclear disk ( cnd ) , and are detected with a high signal - to - noise ratio in the southern part of the cnd , while they are weaker towards the northern part . assuming that the clumps are as close to the galactic center as their projected distances , they are still dense enough to be gravitationally stable against the tidal shear produced by the supermassive black hole . therefore , the cnd is a non - transient structure .
this geometrical distribution of both tracers suggests that the southern part of the cnd is denser than the northern part .
also , by comparing the hcn(4 - 3 ) results with hcn(1 - 0 ) results we can see that the northern and the southern parts of the cnd have different excitation levels , with the southern part warmer than the northern . finally , we compare our results with those obtained with the detection of nh@xmath6 , which traces the warmer and less dense material detected in the inner cavity of the cnd .
we suggest that we are detecting the origin point where a portion of the cnd becomes destabilized and approaches the dynamical center of the milky way , possibly being impacted by the _ southern streamer _ and heated on its way inwards . |
h.s.g . would like to acknowledge support from the national science council , taiwan , under grant no .
97 - 2112-m-002 - 012-my3 , support from the frontier and innovative research program of the national taiwan university under grant nos . 97r0066 - 65 and 97r0066 - 67 , and support from the focus group program of the national center for theoretical sciences , taiwan .
c.a.y . and s.j.h .
would like to acknowledge support from the national science council , taiwan , under grants nos .
97 - 2221-e-239 - 022- and 95 - 2221-e-011 - 032-my3 . | in this comment , we show that the special attack [ s .- j . qin , f. gao , q .- y .
wen , f .-
c .
zhu , opt .
commun .
* 281 * ( 2008 ) 5472 . ] , which claims to be able to obtain all the transmitted secret message bit values of the protocol of the multiparty quantum secret sharing of secure direct communication using single photons with random phase shift operations , fails .
furthermore , a class of similar attacks are also shown to fail to extract the secrete message . quantum secure direct communication , quantum secret sharing , quantum random phase shift operation 03.67.hk , 03.65.ud , 03.67.dd in a recent paper , qin _
et al . _
@xcite pointed out that the hllz protocol @xcite is vulnerable to a special participant attack strategy ( hereafter called the qgwz attack ) where alice s secret message can be extracted by any dishonest agent without any help of the other agents .
however , in this paper , we comment that some errors may exist in the qgwz attack and thus it fails to attack hllz protocol successfully .
the six main stages of the hllz protocol are _ preparation phase _ , _ encryption phase _ , _ the first detection _ , _ encoding phase _ , _ recovery phase _ and _ the second detection_@xcite . in short ,
based on the linear property of the quantum random phase shift operations ( rpsos ) or equivalently random single - qubit rotations along the @xmath0-axis , the encryption , encoding and recovering phase of this protocol can be expressed as the addition / subtraction of the rotation angles on each photon ( qubit ) at different stages : @xmath1 here @xmath2 represents a rotation on photon @xmath3 with @xmath4 , and @xmath5 is the rotation angle made by alice on the photon @xmath3 according to the bit value @xmath6 or @xmath7 of her secret message @xmath8 . the measurement action in the recovery phase by the agent ( say zach ) , to which alice sent the photons with encoded message , reflects basically whether there is a rotation of an angle @xmath9 on the photon ( qubit ) @xmath3 or not .
if there is , the bit value of the message bit @xmath3 is @xmath10 ; otherwise , @xmath6 . grasping the main steps of the hllz protocol , the scenario for an dishonest agent ( say charlie ) in the qgwz attack is to entangle the transmitted photons with prepared ancilla photons firstly , to escape from the detection secondly , and to attain the secret message without being discovered lastly @xcite .
however , the errors that existed in the ref .
@xcite will doom charlie to fail to extract the secret message even if he can escape from the detection .
the first typo in ref .
@xcite is in the last sentence of the first paragraph of step ( e2 ) in the second column on page 5473 . the correct sentence for that should read as `` then charlie sends the fake sequence @xmath11 to alice . '' . in other words , the sequence that was sent out by charlie is @xmath11 rather than @xmath12 stated in ref .
@xcite .
but this typo is very minor as it is obvious from the subsequent description in ref
. @xcite that it is the @xmath11 sequence that was sent out .
the possible serious error is in the expression of eq .
( 9 ) of ref .
@xcite which was derived from eq .
( 7 ) of ref .
@xcite by a controlled - controlled-(@xmath13 ) operation ( referred to as @xmath14 operation in @xcite ) with photons @xmath15 and @xmath16 being the control qubits and photon @xmath17 being the target qubit .
the action of @xmath14 operation on computational basis states is given by @xmath18 , where @xmath19 and @xmath20 represent the control qubits and @xmath21 stands for the target qubit on which the @xmath13 operation acts .
so the correct expression for eq . ( 9 ) of ref .
@xcite should read as @xmath22 comparing eq .
( [ qgwz-0-y ] ) above with eq .
( 10 ) of ref .
@xcite , one can find that the two @xmath23 photon states in those two equations are identical rather than orthogonal .
the statement that the two states of photons @xmath23 are orthogonal in ref .
@xcite comes from comparing the incorrect eq .
( 9 ) of ref .
@xcite with eq .
( 10 ) of ref .
@xcite .
therefore , charlie can not , by measuring photons @xmath23 , distinguish the state @xmath24 of the correct eq .
( [ qgwz-0-y ] ) from the state @xmath25 of eq .
( 10 ) of ref .
@xcite , and thus can not extract alice s secret bit value in practice . as a result , the qgwz attack fails to extract alice s secret message bits from the hllz protocol .
the main idea of the qgwz attack seems ( i ) to use an entanglement operation to entangle each of the photon in sequence @xmath26 with prepared ancilla photons to escape the first detection of the hllz protocol , ( ii ) to use the inverse entanglement operation to separate the entangled states back to separable state to obtain the secret message and to escape the second detection of the hllz protocol .
the particular entanglement operation in the ref .
@xcite is @xmath14 and we have shown that it fails to extract the secret message .
one may wonder whether charlie can extract the secret with other operations instead of @xmath14 in this qgwz - like attack .
in fact , we show below that charlie will still fail to extract the secret even if he can apply other entanglement operations in the qgwz - like attacks . without loss of generality ,
suppose the photon @xmath3 in the @xmath26 sequence sent out by the last agent , zach , is in the state @xmath27 where @xmath4 . in this way
, charlie may prepare his ancilla qubits sequence @xmath28 each in the state @xmath29 , and apply a particular kind of entanglement operation @xmath30 on @xmath31 such that @xmath32_{s_j } , \label{e_operation}\end{aligned}\ ] ] where @xmath33 and @xmath34 , and @xmath35 is the rotation operation along the @xmath0 axis defined in ( [ superposition ] ) .
after these operations , charlie sends sequence @xmath26 to alice , and he may escape from alice s first detection by telling the value of his rotation angle of either @xmath36 or @xmath37 . that is , when alice selects the photon @xmath3 to check the security of the quantum channel used , charlie can first measure the corresponding ancilla qubit @xmath38 in the @xmath39 basis .
he then announces his rotation angle of either @xmath36 or @xmath37 when the measured result is in @xmath40 or in @xmath41 , respectively . as a result
, alice will not discover charlie s dishonest action during her first detection step . from this point of view , the qgwz attack of ref .
@xcite is just a simplified special case of above description where charlie s @xmath30 is the controlled - controlled-(@xmath13 ) operation and the rotated phase shift angle @xmath42 in eq .
( [ e_operation ] ) is @xmath43 . in short , charlie
hence can escape from the first detection in the hllz protocol with this qgwz - like attack .
after that , in the hllz protocol , alice will encode her secret message bit value on each of the remaining photons in sequence @xmath26 with a rotation operation along the @xmath0-axis by an angle of @xmath9 if the bit value is @xmath7 .
otherwise , no operation is performed if the bit value is @xmath44 .
this message - encoding operation can be expressed as @xmath45 when it is applied to the state of eq .
( [ e_operation ] ) , let us consider the following transformation .
@xmath46 where @xmath47 . since @xmath48 of eq .
( [ u_encoding ] ) commutes with @xmath49 in eq .
( [ e_operation ] ) , the resultant state becomes @xmath50_{s_j } \label{1_encode}\end{aligned}\ ] ] if alice s secret message bit value is @xmath7 . otherwise , if alice s secret message bit value is @xmath44 , the resultant state remains in the state of eq .
( [ e_operation ] ) . after intercepting the transmitted photon again
, charlie will apply the inverse entanglement operation @xmath51 to separate the entanglement between photon @xmath38 and @xmath17 in the qgwz - like attack .
he then will try to figure out alice s secret message by making an measurement on his ancilla qubits in the orthonormal basis of @xmath39 . since @xmath52 , if charlie applies the operator @xmath51 to eq .
( [ e_operation ] ) , the state of the photons @xmath38 and @xmath17 will become @xmath53_{s_j } ) = { \mbox{$ | \varepsilon \rangle $ } } _ { a_j } \otimes { \mbox{$ | \chi \rangle $ } } _ { s_j}. \label{0_es}\end{aligned}\ ] ] it follows that if the operator @xmath51 is applied to eq . ( [ 1_encode ] )
, it will become @xmath54_{s_j } ) = { \mbox{$ | \varepsilon \rangle $ } } _ { a_j } \otimes { \mbox{$ | \chi^u \rangle $ } } _ { s_j}. \label{1_es}\end{aligned}\ ] ] since the photon state @xmath29 is the same in both eq .
( [ 0_es ] ) and eq .
( [ 1_es ] ) , it is obvious that charlie can not distinguish whether the state of photon @xmath17 has been rotated by an angle of @xmath9 or not from the measurement result of the ancilla photons @xmath38 .
thus charlie can not extract the transmitted secret bits even if he can apply the entanglement operations defined in eq .
( [ e_operation ] ) in the qgwz - like attacks . in summary ,
the qgwz attack as well as the qgwz - like attacks which apply entangled operations other than the @xmath14 operation fail to obtain the transmitted secret message bits of the hllz protocol . |
the main goal of heavy - ion collision ( hic ) physics is the study of the behavior of nuclear matter at high temperatures and/or high densities . in particular , the major interest is the investigation of the nuclear matter phase diagram as a function of temperature and baryon chemical potential . according to quantum - chromo - dynamics ( qcd ) , at large temperatures and densities the hadrons can not survive anymore as bound states and they dissolve forming the so - called quark - gluon - plasma ( qgp ) . at relativistic heavy - ion collider ( rhic )
energies for the first time the creation of a qgp , identified as an almost perfect fluid @xcite , has been proven experimentally .
actually , the properties of this deconfined state of matter are still debated as well as the phase boundaries to the hadronic phase . in order to shed some light on this issue ,
many heavy - ion experiments are performed at the super - proton synchrotron ( sps ) , rhic , the large hadron collider ( lhc ) and will be performed at the future facility for antiproton and ion research ( fair ) as well as the nuclotron - based ion collider facility ( nica ) .
the crucial challenge is to identify in the final particle distributions those signatures , which allow to disentangle the qgp contribution that is impossible to observe directly or independently .
the strange particle production has always been suggested as one of the most sensitive observables that could spot out the creation of a qgp during the early stages of a hic .
the earliest suggested signature is the strangeness enhancement in a+a collisions with respect to elementary p+p collisions @xcite .
later on , gazdzicki and gorenstein @xcite proposed that a sharp rise and drop in the excitation function of the @xmath4 ratio ( so called `` horn '' ) should show up due to the appearance of a qgp phase at a center - of - mass energy @xmath7gev .
several statistical models @xcite have succeeded in reproducing the trend of the experimental observation of the @xmath4 ratio and other strange to non - strange particle ratios , but they can provide only a statistical description of the heavy - ion collision process . on the other hand there was no conclusive interpretation of the `` horn '' from dynamical approaches for hic , like microscopic transport models @xcite .
only recently , the parton - hadron - string dynamics ( phsd ) , a transport approach describing hic on the basis of partonic , hadronic and string degrees - of - freedom , obtained a striking improvement on this issue when including chiral symmetry restoration ( csr ) in the string decay for hadronic particle production @xcite . within the phsd approach
it has been suggested that the horn feature emerges in the energy - dependence of the @xmath4 ratio , both due to csr , which is responsible for the rise at low energies , and to the appearance of a deconfined partonic medium at higher energies , which is responsible for the drop at top sps energies .
apart from deconfinement the chiral symmetry restoration addresses another aspect of the qcd phase diagram in the ( @xmath8)-plane as an additional transition between a phase with broken and a phase with restored chiral symmetry . as in case of the qcd deconfinement phase transition ,
the boundaries of the csr phase transition line are not well known .
lattice qcd ( lqcd ) calculations show that at vanishing baryon chemical potential @xmath9=0 the csr takes place at roughly the same critical temperature and energy density as the deconfinement phase transition which is a crossover . at finite baryon
chemical potential lqcd calculations can not be performed due to the sign problem and one must rely on effective models ( or extrapolations ) in order to study the qcd phase transitions @xcite .
different models support the idea that at finite chemical potential a partially restored phase is achieved before the deconfinement occurs @xcite . in order to distinguish the two phases of such a transition
, effective models use the scalar quark condensate @xmath10 as an order parameter . as the baryon density and temperature increase , the scalar quark condensate @xmath10 is expected to decrease from a non - vanishing value in the vacuum to @xmath11 which corresponds to csr .
since @xmath12 is not a measurable quantity , it is crucial to determine experimental observables which are sensitive to this feature .
since long the dilepton spectroscopy has been in the focus in this respect since in a chirally restored phase the spectral functions of the the @xmath13- and the @xmath14-meson should become identical .
however , no clear evidence has been achieved so far @xcite . on the other hand , the strangeness production at alternating - gradient synchrotron ( ags ) and
lower sps energies was suggested to be a signature of csr @xcite . in this work
we will perform a systematic study within the phsd approach on effects of the csr on final particle distributions in hics in the energy range @xmath15gev .
the phsd is particularly suited to this aim since it includes both hadronic and partonic degrees - of - freedom @xcite and has been extensively used to describe hic observables from sps to lhc energies @xcite .
this study is organized as follows : in sec .
ii we briefly recall the basic ingredients of the phsd approach , while in sec .
iii we illustrate the string fragmentation included in phsd and the most recent extension of this particle production mechanism to incorporate essential aspects of chiral symmetry restoration .
we also discuss the sensitivity of our results on different parameter settings for the nuclear equation of state ( eos ) within the non - linear @xmath1 model and examine the role of three - meson channels for strangeness production ( @xmath16 ) . in sec .
iv we present the calculated results from phsd with and without the inclusion of csr for particle spectra , yields and ratios in central a+a collisions from 2 to 158agev , with a special focus on the strange particle production . a comparison with experimental data
is performed wherever possible . in sec .
v we explore new aspects of csr in heavy - ion collisions , i.e. the system size and the centrality dependence of strange particle yields and ratios . in sec .
vi we focus on the equilibrium stages of the time evolution of central au+au collisions in connection to strangeness production in the ( @xmath17)-plane and discuss the perspectives to identify a critical point in the phase diagram .
we conclude this work with a summary in sec.vii .
the parton - hadron - string dynamics ( phsd ) is a microscopic covariant dynamical approach for strongly interacting systems in and out - of equilibrium @xcite .
it is a transport approach which goes beyond the quasi - particle approximation , since it is based on kadanoff - baym equations for the greens functions in phase - space representation in first - order gradient expansion @xcite . including both a hadronic and a partonic phase as well as a transition between the effective degrees - of - freedom , phsd is capable to describe the full time evolution of a relativistic heavy - ion collision . the theoretical description of the partonic degrees - of - freedom ( quarks and gluons ) is realized in line with the dynamical - quasi - particle model ( dqpm ) @xcite which reproduces lqcd results in thermodynamical equilibrium and provides the properties of the partons , i.e. masses and widths in their spectral functions . in equilibrium
the phsd reproduces the partonic transport coefficients such as shear and bulk viscosities or the electric conductivity from lqcd calculations as well @xcite .
an actual nucleus - nucleus collision in phsd follows the following steps : * primary hard scatterings between nucleons take place and produce excited color - singlet states , denoted by `` strings '' ( as described within the fritiof lund model @xcite and realized also in pythia 6.4 @xcite ) .
these strings decay into `` prehadrons '' with a formation time @xmath18fm / c and `` leading hadrons '' , which originate from the string ends and can re - interact with hadrons almost instantly with reduced cross - sections ( according to the constituent quark number ) . * in case the local energy density is above the critical value of @xmath19gev/@xmath20 , the deconfinement is implemented by dissolving the newly produced hadrons into the massive colored quarks / antiquarks and mean - field energy from the dqpm . * within the qgp phase , the partons ( quarks , antiquarks and gluons ) scatter and propagate in a self - generated mean - field potential .
they are described as off - shell quasi - particles with temperature - dependent masses and widths , which are given by the dqpm . *
the expansion of the system is associated to a decrease of the local energy density and , once the local energy density becomes close to or lower than @xmath21 , the massive colored off - shell quarks and antiquarks hadronize to colorless off - shell mesons and baryons .
the hadronization process is defined by covariant transition rates and fulfills the energy - momentum and quantum number conservation in each event . * in the hadronic corona as well as in the late hadronic phase , the particles are still interacting and propagating .
elastic and inelastic collisions between baryons , mesons and resonances are implemented in phsd and the corresponding backward reactions are included through detailed balance for all channels .
we further note that the pure hadronic phase in phsd is equivalent to the hadron - strings dynamics ( hsd ) model @xcite .
accordingly , the comparison between phsd and hsd calculations allows us to disentangle the role of the qgp phase in heavy - ion collisions . even at low center - of - mass energies ( @xmath22gev ) the phsd and
the hsd results slightly differ due to the appearance of qgp droplets in central cells of the system , which are characterized by high baryon and energy densities .
the phsd approach has been tested for different colliding systems ( p+p , p+a , a+a ) in a wide range of bombarding energy , from ags to lhc energies , and has been able to describe a large number of experimental observables , such as charged particle spectra , collective flow coefficients @xmath23 as well as electromagnetic probes such as photons and dileptons @xcite .
more recently , in ref .
@xcite , it has been also shown to provide a microscopic description of the maximum in the @xmath4 ratio in central nucleus - nucleus collisions . in the latter work we also found that the inclusion of csr in the hadronic sector via string decay is crucial in order to reproduce the strangeness enhancement at ags and lower sps energies .
the string formation and decay represents the dominant particle production mechanism in nucleus - nucleus collisions for bombarding energies from @xmath24agev to @xmath25agev . in phsd ,
the primary hard scatterings between nucleons are described by string formation and decay in the fritiof lund model @xcite .
a string is an excited color - singlet state , which is composed of two string ends corresponding to the leading constituent quarks of the colliding hadrons and a color flux tube in between . as the string ends recede , virtual @xmath26 or @xmath27 pairs
are produced in the uniform color field , causing the breaking of the string .
finally , the string decays into mesons or baryon - antibaryon pairs with formation time @xmath18fm / c ( in the rest - frame of the string ) .
+ in the string decay , the flavor of the produced quarks is determined via the schwinger formula @xcite , which defines the production probability of massive @xmath28 pairs with respect to light flavor production @xmath29 pairs : @xmath30 with @xmath31gev@xmath32 representing the string tension and @xmath33 denoting the constituent quark masses for strange and light quarks . for the constituent quark masses
@xmath34gev and @xmath35gev in the vacuum , the production of strange quarks is suppressed by a factor of @xmath36 with respect to the light quarks , which is the default setting in the fritiof routines . the relative production factors in phsd / hsd have been readjusted in 1998 as follows @xcite : @xmath37 the probability ratio @xmath38 has been increased to @xmath39 at ags energies in order to correctly reproduce the strangeness yield for p+be collisions at ags energies @xcite . a smooth transition between the two values of @xmath38
is ensured by a linear interpolation as a function of the center - of - mass energy @xmath40 . + a further ingredient to fix the rapidity distribution for the newly produced hadrons is the fraction of energy and momentum that they acquire from the decaying string .
this is defined by the fragmentation function @xmath41 , which is the probability distribution for a hadron with transverse mass @xmath42 to be produced with an energy - momentum fraction @xmath43 from the fragmenting string : @xmath44 where @xmath45 and @xmath46gev@xmath47 as reliable settings for p+p and p+a collisions . as becomes evident from eq .
( [ frag0 ] ) the meson @xmath42-scaling from string decay is included by default . in ref .
@xcite the phsd has been extended to include csr in the string decay in a hadronic environment of finite baryon and meson density .
here we recall the main aspect of this extension which is based on the hellman - feynman theorem for the scalar quark condensate @xcite .
accordingly , a linear decrease of the scalar quark condensate @xmath48 which is nonvanishing in the vacuum due to a spontaneous breaking of chiral symmetry is expected with baryon density @xmath49 towards a chiral symmetric phase characterized by @xmath50 0 @xcite .
this decrease of the scalar quark condensate is expected also to lead to a change of the hadron properties with density and temperature , i.e. in a chirally restored phase the vector and axial vector currents should become equal @xcite ; the latter implies that e.g. the @xmath13 and @xmath14 spectral functions should become identical ( as addressed above in the context of dilepton production ) .
since the scalar quark condensate @xmath51 is not a direct observable , its manifestations should also be found indirectly in different hadronic abundancies and spectra or particle ratios like @xmath4 , @xmath52 etc . as advocated in ref .
@xcite . in leading order
the scalar quark condensate @xmath10 can be evaluated in a dynamical calculation as follows @xcite : @xmath53 where @xmath54 stands for the @xmath55-commutator of the relevant mesons @xmath56 , @xmath57 represents the vacuum condensate , @xmath58mev is the pion - nucleon @xmath2-term and @xmath59 and @xmath60 are the pion decay constant and pion mass , respectively .
note , however , that the value of @xmath61 is not so accurately known ; a recent analysis points towards a larger value of @xmath62 mev @xcite while actual lqcd results @xcite suggest a substantially lower value .
accordingly , our following calculations - based on @xmath63 mev - have to be taken with some care although in some sense it represents a world average ( cf
. fig . 3 in ref .
@xcite ) . according to the light quark content , the @xmath2-term for hyperons is decreased by a factor of 2/3 for @xmath64 and @xmath2 hyperons and by a factor of 1/3 for @xmath65 baryons .
furthermore , for mesons made out of light quarks and antiquarks , we use @xmath66 , whereas for mesons with a strange ( antistrange ) quark we consider @xmath67 .
we mention here that improved results for the sigma - commutator for kaons can be obtained from chiral perturbation theory as in ref . @xcite and
alternative assumptions for non - pseudoscalar mesons can be employed as e.g. suggested by cohen et al . in ref .
@xcite . in view of the subleading contributions of these mesons to the ratio in eq .
( [ scalar_condensate ] ) we keep the simple estimates noted above for our present study and look forward to a clarification of the present tension between the results from lqcd and dispersive approaches @xcite . in eq .
( [ scalar_condensate ] ) , the quantities @xmath68 and @xmath69 denote the nucleon scalar density and the scalar density for a meson of type @xmath56 , respectively .
the scalar density of mesons @xmath56 is evaluated in the independent - particle approximation as : @xmath70 where @xmath71 denotes the meson phase - space distribution ( @xmath72 ) and @xmath73 refer to the discrete spin and isospin quantum numbers , respectively .
moreover , the vacuum scalar condensate @xmath74 can be computed according to the gell - mann - oakes - renner ( gor ) relation @xcite , @xmath75 and gives @xmath76@xmath77 for the bare quark masses @xmath78mev . finally , in eq .
( [ scalar_condensate ] ) the nucleon scalar density @xmath68 has to be determined in a suitable model with interacting degrees - of - freedom in order to match our knowledge on the nuclear eos at low temperature and finite density .
a proper ( and widely used ) approach is the non - linear @xmath1 model for nuclear matter where @xmath68 is defined as : @xmath79 where @xmath80 and @xmath81 denote the effective mass and momentum , respectively , and @xmath82 the phase - space occupation of a nucleon while the degeneracy factor is @xmath83=4 .
in fact , in the non - linear @xmath1 model the nucleon mass is modified due to the scalar interaction with the medium : @xmath84 where @xmath85 denotes the nucleon mass in vacuum and @xmath86 is the scalar field which mediates the interaction between the nucleons and the medium with the coupling @xmath87 . in order to calculate @xmath68
, we need to determine the value of the scalar field @xmath86 at each space - time point @xmath43 .
this is done via the non - linear gap equation @xcite : @xmath88 @xmath89 since for matter at rest we have @xmath90 . in eq .
( [ gap_equation ] ) the self - interaction of the @xmath55-field is included up to the forth order .
the parameters @xmath91 are fixed in order to reproduce the values of the nuclear matter quantities at saturation , i.e. the saturation density , the binding energy per nucleon , the compression modulus , and the effective nucleon mass . actually , there are different sets for these quantities that lead to slightly different saturation properties .
we defer a discussion on the uncertainties of our results to the following section [ eos ] .
the main idea in ref .
@xcite is to consider effective masses for the dressed quarks in the schwinger formula ( [ schwinger - formula ] ) for the string decay in a hot and dense medium .
the effective quark masses can be expressed in terms of a scalar coupling to the quark condensate @xmath10 in first order as follows : @xmath92 @xmath93 with @xmath94mev and @xmath95mev for the bare quark masses . in eq .
( [ effective_masses ] ) the effective masses decrease from the vacuum values with decreasing scalar condensate @xmath10 to the constituent masses .
this adaptation of the schwinger formula in case of a hot and dense medium implies a modification of the flavor production factors in eq .
( [ schwinger_factors ] ) . in an actual nucleus - nucleus collision
, phsd incorporates a dynamical calculation of all these features for each cell in space - time : * the scalar density @xmath68 is determined by solving the gap equation ( [ gap_equation ] ) for the @xmath55-field ; * the scalar condensate @xmath10 is then computed via eq .
( [ scalar_condensate ] ) ; * the effective masses @xmath96 are calculated according to eqs . (
[ effective_masses ] ) , ( [ mus ] ) and plugged in the schwinger formula ( [ schwinger - formula ] ) in order to compute the flavor production ratios for the string decay .
we stress that , once the nucleon scalar density @xmath68 and @xmath97 are fixed , there is no additional parameter in the phsd3.3 compared to the previous version phsd3.2 that has been employed for a couple years for the analysis of relativistic heavy - ion reactions @xcite . in this section
we analyze in more detail the flavor production ratios from the schwinger formula in the presence of a hot and dense nuclear medium .
as mentioned in the previous section there are different sets for the parameters @xmath91 in the gap equation ( [ gap_equation ] ) .
in fact , these parameters are fixed within the non - linear @xmath1 model in order to reproduce empirical values of nuclear matter quantities at saturation ( i.e. saturation density , binding energy per nucleon , compression modulus , effective nucleon mass etc . ) , but sizeable uncertainties remain with respect to the high density properties . in table
i we display the values of @xmath98 together with the vector coupling @xmath99 , the vector meson mass @xmath100 , the compression modulus @xmath101 and the ratio between the effective and the bare nucleon mass @xmath102 at saturation density for two sets commonly indicated as nl1 and nl3 .
[ tab1 ] .parameter sets nl1 and nl3 for the non - linear @xmath1 model employed in the transport calculations from ref . @xcite . [ cols="^,^,^,^",options="header " , ] in the context of the string decay , the most important parameter to focus on is the scalar coupling @xmath87 , which is lower for the nl1 set with respect to the corresponding values in the nl3 set . in fig .
[ plots_eos ] we show the dependence of the nucleon scalar density @xmath68 on energy density @xmath103 in panel ( a ) , the ratio between the scalar quark condensate and its value in the vacuum @xmath104 in panel ( b ) , the light and strange quark effective masses @xmath105 in panel ( c ) , and the production probability of massive @xmath28 relative to light flavor production @xmath38 in panel ( d ) . note that an analogous dependence is observed with respect to the baryon density @xmath49 since the energy density @xmath103 in leading order is just the nucleon mass times the baryon density . ( a ) , the ratio between the scalar quark condensate and its value in the vacuum @xmath104 ( b ) , the light and strange quark effective masses @xmath105 ( c ) , and the production probability of massive @xmath28 relative to light flavor production @xmath38 ( d ) as a function of the energy density @xmath103 for the parameter sets nl3 and nl1 at @xmath106 and with @xmath107=45 mev.,scaledwidth=46.0% ] the results shown in fig .
[ plots_eos ] are obtained at vanishing temperature @xmath106 , but all the following considerations can be extended to a more realistic picture at finite temperature ( meson density ) and illustrate the consequences of csr in the phsd results for heavy - ion collisions . the energy density @xmath103 here
is calculated within the non - linear @xmath1 model by : @xmath108 with @xmath109 @xmath110 where @xmath111 represents the nucleon density and @xmath112 and @xmath113 are the particle / antiparticle occupation numbers at fixed momentum @xmath114 , respectively . the scalar density @xmath68 increases with increasing energy density @xmath103 as displayed in panel ( a ) of fig .
[ plots_eos ] .
we find a moderate sensitivity to the nuclear equation of state up to energy densities of @xmath115 0.5 gev/@xmath20 that are of relevance for the hadronic phase .
in fact , both lines referring to the two parameter sets nl1 and nl3 show a very similar behavior as a function of @xmath103 , but the nl3 ( solid line ) set is always characterized by lower values of the scalar density @xmath68 relative to the nl1 parametrization ( dashed line ) .
this is due to the smaller value of the scalar coupling @xmath87 in case of the nl1 parameter set .
in fact , smaller values of @xmath87 are associated with larger values of the nucleon effective mass @xmath80 and consequently to larger values of the scalar density @xmath68 , as it is evident from eqs .
( [ rhos_nucleon ] ) and ( [ mass ] ) . in panel ( b ) of fig .
[ plots_eos ] the ratio @xmath104 is presented as a function of @xmath103 . at @xmath116 the scalar condensate corresponds to the vacuum value @xmath117 and for fixed @xmath118=45 mev it decreases almost linearly with increasing energy density and almost vanishes for the critical energy density @xmath119 0.5 gev/@xmath20 . in this case , the order between nl1 and nl3 results is reversed : the nl3 parametrization for the nuclear eos is associated to higher values of the scalar quark condensate with respect to the nl1 set .
this feature can be easily explained looking at the definition of the ratio @xmath104 ( [ scalar_condensate ] ) : at @xmath106 , there are no thermal mesons , thus the last term of the relation vanishes and the ratio is entirely fixed by the scalar density @xmath68 ; hence , higher values of @xmath68 correspond to lower values of @xmath104 .
therefore , the nl1 parametrization is characterized by lower values of the scalar quark condensate with respect to the nl3 parameter set .
it follows that the light and strange quark effective masses @xmath105 in panel ( c ) of fig .
[ plots_eos ] show a very similar dependence on the energy density . at vanishing energy density @xmath103 , the quark effective masses
are equal to their vacuum values , @xmath120gev and @xmath35gev ; with increasing @xmath103 the quark masses decrease in line with the scalar quark condensate up to their bare values @xmath94mev and @xmath95mev for vanishing @xmath104 .
the decrease of both @xmath121 and @xmath122 is approximately linear in energy density where the slope associated to the light quark is flatter in comparison to the strange quark mass .
concerning the comparison between the two different choices for the nuclear equation of state , we find also for these masses a non - negligible sensitivity and the same hierarchy as for the scalar quark condensate ( the results associated to nl1 are always below the corresponding results for nl3 ) .
we recall that the effective masses of the quarks enter the schwinger formula ( [ schwinger - formula ] ) for the hadronic particle production via the string decays . in panel
( d ) of fig .
[ plots_eos ] the strangeness ratio @xmath38 is shown as a function of energy density for the two parameter sets .
the factor @xmath38 increases from the vacuum case ( @xmath123 0.3 ) with increasing energy density up to values of 0.8 0.9 for @xmath124 .
thus the production of a @xmath125 pair relative to a light quark pair is no longer suppressed close to the phase boundary for csr as it is in vacuum .
the reason of this increase is the steeper decrease of the effective strange quark mass ( with energy density ) in comparison to the effective light quark mass as mentioned above .
furthermore , the nl1 parametrization gives larger values of @xmath38 as the nl3 parametrization due to a faster change of the masses with @xmath103 ( cf . panel ( c ) ) .
we note in passing that this scheme for csr in the string decay mechanism can be applied not only to the light and strange quarks , but also to diquark combinations that are additionally produced in the fragmentation of the string and lead finally to baryon - antibaryon pairs . here
the default jetset ratios fix the diquark mass in the vacuum , e.g. a light diquark mass in vacuum of @xmath126= 0.65 gev leads to a suppression of a light diquark pair relative to a light quark - antiquark pair of @xmath127 for the creation of a diquark ( @xmath128 ) one employs @xmath129 0.725 gev which leads to the ratio for a ( @xmath128)-diquark pair relative to a light ( @xmath130)-diquark pair of @xmath131 within the same line the vacuum mass of a @xmath132-diquark can be determined from jetset .
the schwinger mechanism of string decay in vacuum thus requires the following dressed vacuum masses : @xmath133 gev , @xmath134 0.5 gev , @xmath135 0.65 gev , @xmath129 0.725 gev and @xmath136 0.87 gev to comply with experimental observation in nucleon - nucleon collisions .
the production probability of diquarks ( @xmath128 ) relative to @xmath130-diquarks ( [ 2 ] ) and ( @xmath132 ) relative to @xmath130-diquarks does not change very much in the dense medium in line with ( indirect ) experimental observation and since @xmath137 , the bare diquark masses @xmath138 can be fixed and give @xmath139 0.5 gev , @xmath140 0.593 gev and @xmath141 0.763 gev .
the explicit variations of the flavor ratios with the energy density are displayed in fig .
[ fig_ratios ] and show that the diquark ratios only very moderately change with the energy density whereas the @xmath142 ratio steeply rises with @xmath103 .
we specify once more that these results have been obtained within a pure hadronic system at vanishing temperature , but the conclusions are valid also for finite temperatures ( meson densities ) . as evaluated within the non - linear @xmath1 model for the parameter set nl3 for @xmath106.,scaledwidth=45.0% ] the variations of the ratios in fig .
[ fig_ratios ] , however , are limited to the hadronic phase with energy densities below @xmath21 . above @xmath143gev/@xmath144 strings
can not be formed anymore due to a vanishing string tension @xmath145 in the qgp and the hadrons dissolve to partonic degrees - of - freedom ( and mean - field energy ) .
consequently , the @xmath142 factor finally shows an increase for @xmath146 ( associated to csr ) and in correspondence of @xmath147 it drops to the value @xmath148 ( fixed by comparison with the strangeness production at rhic and lhc energies observed experimentally ) .
the energy dependence of the @xmath142 ratio including the qgp phase has been shown already in fig . 2 of ref .
we recall that in the partonic phase the @xmath142 ratio remains constant as a function of the energy density . as a result
we can identify a `` horn '' structure in the @xmath142 ratio as a function of @xmath103 , where the initial increase is related to chiral symmetry restoration in the hadronic phase and the subsequent sharp decrease is associated to the formation of the qgp .
we thus expect that csr modifies the particle abundancies and spectra ( especially in the strange particle sector ) from heavy - ion collisions , where increasing energy densities @xmath103 in the overlap region can be achieved with increasing bombarding energy ( at the same centrality of the collision ) . for a solid study of the differential spectra of strange and non - strange particles
it is essential that both are treated on the same level in many - body theory .
whereas in earlier phsd studies the three - body channels incorporating three mesons in the production of baryon - antibaryon pairs in line with ref .
@xcite ( e.g. @xmath150 ) have been incorporated by default in the non - strange sector the corresponding channels in the strangeness sector ( e.g. @xmath151 etc . ) had been discarded in phsd3.2 . nevertheless , the possible impact of such channels on the observables of interest in this study has to be examined in order to control the consistency of the approach .
accordingly , in this subsection we present the influence of an extended description of @xmath152 reactions to the strange sector on the rapidity spectra from central heavy - ion collisions with and without csr .
baryons in 5% central au+au collisions for the energies of 8 ( a ) , 30 ( b ) and 158agev ( c ) with ( solid lines ) and without csr ( dashed lines ) .
the blue lines stand for the results that include the new three - meson reactions with strangeness content while the red lines display results when incorporating only non - strange three - meson channels.,scaledwidth=45.0% ] to this aim we have generalized the quark rearrangement model , first presented in ref .
@xcite for the light ( @xmath153 ) sector , to the strange sector .
the details of the extended model will be given in an upcoming study where also the explicit tests of detailed balance relations on a channel by channel basis are demonstrated as well as the particular impact on multi - strange antibaryon production and annihilation . by including the strangeness sector
the number of individual three - body ( mass channels ) amounts to more than 2500 including also the fusion of three kaons ( or @xmath154 s ) to the @xmath155 channel .
we also consider the hidden strangeness of the @xmath156 and the @xmath157 mesons with proper weights for channels where either the light quarks or @xmath28 are rearranged . as in ref .
@xcite the matrix elements for the transitions are provided by experimental data on baryon - antibaryon annihilation . without going into further details
we show in fig .
[ fig::old - new-3body ]
exemplary for other particles the rapidity spectra of @xmath158 baryons at 8 , 30 and 158agev in central au + au collisions with ( solid lines ) and without ( dashed lines ) chiral symmetry restoration as well as with ( blue lines ) and without ( red lines ) the extended strangeness @xmath159 reactions . in general
, the inclusion of chiral symmetry restoration enhances the number of produced @xmath158 baryons at all energies .
this is most pronounced for the lower energies of 8 and 30 agev ( cf .
the detailed studies in the next section ) . in particular for the lower energies there is no visible difference in the rapidity spectra for the extended @xmath160 reactions . only at 158agev
one finds a slight difference ( @xmath115 1 % ) in the central rapidity region .
the reason for these findings is related to the low abundance of mesons carrying a strange quark at low bombarding energies and additionally in the heavier masses of the hyperons which induce a lower phase - space for production , since the transition probability is directly proportional to the available phase - space in the flavor rearrangement model of ref .
we conclude that the extended three - body reactions with strangeness have no crucial impact at ags and sps energies .
in this section we study observables from heavy - ion collisions with respect to different aspects .
we present results obtained within the phsd for the rapidity distribution and transverse mass spectra of the most abundant particles at ags and sps energies as well as excitation functions of particle yields and ratios .
moreover , we show the time evolution of the number of strange particles and the corresponding production rates .
the following scenarios will be explored : * default phsd calculations without csr ; * phsd calculations including csr with nl3 as parameter set for the nuclear eos ; * phsd calculations including csr with nl1 as parameter set for the nuclear eos . at ags and sps energies
the strange particle production takes place at the early stages of the collision process as it is seen from fig .
[ number_strange ] , where the number of particles containing @xmath161-quarks @xmath162 is plotted ( green solid and dashed lines ) as a function of time in central au+au collision at a bombarding energy of 30 agev for 5% most central collisions . as a function of time in 5% central au+au collision at 30 agev .
the solid lines show the results from phsd including csr with nl1 parameters , while the dashed lines show the results from phsd without csr .
the green lines refer to the total number of strange particles , while the blue and red lines refer to the hadronic and partonic contributions of the strangeness content , respectively.,scaledwidth=45.0% ] more than @xmath163 of the strange content of the system is created in the time interval between 0 and 4fm / c after the collision and already at about 8fm / c strange particles are not produced anymore . in fig .
[ number_strange ] the total number of strange particles has also been separated in the corresponding hadronic and partonic contributions , represented by the blue and red lines , respectively . at the energy of 30agev
, the hadronic strangeness content is dominant relative to the partonic one , which can reach @xmath164 of the total strangeness content in correspondence of the maximum value of the partonic contribution .
the strange quarks in the partonic phase appear in the system not immediately after the collision , but after about 1.5fm / c .
in fact , the primary interactions within the phsd are realized via string excitation and after that in the cells with energy density @xmath165gev/@xmath144 the hadrons are dissolved into partons and mean field energy .
the partonic @xmath162 distribution initially increases as a function of time , reaching a maximum at about 3fm / c . then at larger times
the energy density of the system decreases , as a result the partons hadronize by dynamical coalescence which lasts up to about 8fm / c , when the strange hadron number is basically fixed .
we can conclude that at 30agev the steep increase of the number of strange particles as a function of time has to be attributed dominantly to the hadronic production , which occurs in phsd via string formation and decay .
the other hadronic scattering processes , which are not negligible at this bombarding energy , do not create further strangeness in the system , but they are only responsible for strange flavor exchanges .
it is , furthermore , interesting to compare the strange particle amount computed in phsd including and excluding csr in the string dynamics . in fig .
[ number_strange ] we show phsd calculations only in the limits : without csr ( dashed lines ) and including csr with nl1 as parameter set ( solid lines ) .
the restoration of chiral symmetry causes a sizable increase ( @xmath166 ) of the total strangeness content .
we notice that csr does not modify the time evolution of @xmath167 , but it only affects the hadronic contribution to the strange particle production .
there is a slight difference between the partonic results with and without csr since the strange particle number in the partonic phase is slightly higher when including csr .
this is , however , not due to a higher strangeness production in the qgp , but stems from particles , which are produced by string decay in the hadronic corona and travel to cells with energy density above @xmath21 during their propagation .
thus such strange particles , even if produced by the string decay , dissolve into partonic degrees - of - freedom . in this respect the enhancement of the strange particle number in
the hadronic phase drives a small increase also in the partonic contribution of @xmath162 . to provide a more complete illustration of the time evolution of the strangeness content in a heavy - ion collision we show in fig .
[ rate_strange ] the rate @xmath168 of strange particles at 8 ( a ) , 30 ( b ) , 158agev ( c ) in central collisions .
we show again the total strange particle rate in green , the hadronic contribution in blue and the partonic contribution in red , comparing in all cases the calculations with and without csr by the solid and dashed lines , respectively .
we mention that these rates include production as well as losses either due to dissolution of hadrons in the qgp or due to hadronization of strange partons .
accordingly these rates become negative when dissolution or hadronization dominates . as a function of the time in 5% central au+au collisions at 8 ( a ) , 30 ( b ) , 158agev ( c ) .
the coding of the lines is the same as in fig .
[ number_strange ] .
note that these rates become negative when strange hadron dissolution or strange parton hadronization dominate , respectively .
, scaledwidth=45.0% ] as already seen in fig .
[ number_strange ] , the strange particle production occurs at the first stages of the collisions and the strangeness production rate increases with increasing bombarding energy .
the hadronic contribution is dominant at the collision energies @xmath169 and 30agev , while the partonic production rate is larger than the hadronic one at @xmath170agev . the strangeness enhancement associated to csr
is most clearly visible at lower energies while at 158agev it is very moderate .
we will find the same result on the final particle rapidity spectra ( see next subsection ) .
the total strange particle rate remains positive during the entire time evolution in panels ( a ) and ( b ) and shows a small negative rate only for @xmath170agev due to the dissolution of strange hadrons in the qgp ; on the other hand the partonic rate becomes negative at larger times since the hadronization ( loss ) superseds the strange quark production .
the negative rate on the partonic side is balanced by the positive rate on the hadronic side due to strangeness conservation . comparing the rates at the different energies
, we can see that the strangeness production is slower at lower energies .
in fact , the peak of the total @xmath168 shifts to smaller times with increasing energies , and for @xmath171agev the whole strangeness production occurs within @xmath24fm / c . on the other hand
the duration of the hadronization process becomes longer at higher bombarding energies , where a larger volume of the systems turns into the qgp phase .
in general , after @xmath172fm / c the strangeness content is fixed and within phsd the creation of strange particles ceases .
equilibrium aspects of the strangeness production will be discussed in sec.vi .
we present in figs .
[ rapidity1 ] to [ rapidity3 ] the phsd results for the rapidity distribution of protons , @xmath173 s , pions and kaons for central nucleus - nucleus collisions at different energies ( from ags to top sps energies ) in comparison to the experimental data from refs . @xcite .
the following scenarios will be explored at different energies : default phsd calculations without csr , represented by the dotted blue lines ; phsd calculations including csr with nl3 as parameter set for the nuclear eos , represented by the solid red lines ; phsd calculations including csr with nl1 as parameter set for the nuclear eos , represented by the dashed green lines .
s , pions and kaons for 5% central au+au collisions at 10.7agev in comparison to the experimental data from refs . @xcite .
the solid ( red ) lines show the results from phsd including csr with nl3 parameters , the dashed green lines show the results from phsd including csr with nl1 parameters and the blue dotted lines show the result from phsd without csr.,scaledwidth=43.0% ] s , pions and kaons for 5% central au+au collisions at 30agev in comparison to the experimental data from ref .
the coding of the lines is the same as in fig .
[ rapidity1].,scaledwidth=43.0% ] s , pions and kaons for 5% central au+au collisions at 158agev in comparison to the experimental data from ref .
the coding of the lines is the same as in fig .
[ rapidity1].,scaledwidth=43.0% ] the results of the first two scenarios are almost equivalent to the phsd calculations shown in the previous study @xcite , with a slight difference regarding the proton and meson spectra .
this is due to a few recent upgrades in phsd : the first concerns the extension of inelastic meson - meson scattering above 1.3 gev of invariant energy @xmath40 to string formation and decay with a cross section of @xmath115 10 mb ; the second improvement is related to p - wave scattering in the reaction channel @xmath174 which had been treated isotropically before .
first , we compare the results with and without csr at @xmath175agev ( fig .
[ rapidity1 ] ) , to point out the general effect of this mechanism on the final particle rapidity distributions .
furthermore , we show the dependence on the parameter sets nl1 and nl3 for the non - linear @xmath1 model for the nuclear eos .
the restoration of chiral symmetry gives an enhancement of the strange particle yields both for mesons and baryons . on the other hand , it produces a slight decrease in the number of pions at midrapidity due to
the suppression of pions in the string decays in favor of strange hadrons .
the proton rapidity spectra do not present any sensible variation , in fact the csr as implemented in phsd modifies essentially the chemistry of the newly produced particles in the string decay and has a minor impact the dynamics of the nucleons , which in the string picture are associated to the string ends of the primary interactions in the system .
the inclusion of the csr is essential in order to correctly reproduce the strange particle rapidity spectra , as we can see especially for @xmath173 hyperons and @xmath176 mesons .
furthermore , our calculations for the proton rapidity spectra are in good agreement with experimental observation .
next , we discuss the results from phsd with csr using two different parametrization for the nuclear equation of state , i.e. nl3 and nl1 .
the general features of the strangeness enhancement hold for both parametrizations ; in particular the nl1 set provides larger values for all strange particle rapidity spectra at midrapidity in line with the discussion of fig .
1 . the difference between the two parametrizations represents the uncertainty of our results related to csr as implemented in phsd .
we stress that we do not tune the parameters of the equation of state to fit the data , but we employ different nuclear eos to compute the scalar density ( as explained in section [ eos ] ) in order to explore the uncertainties of our approach . in fig .
[ rapidity2 ] the rapidity spectra of various hadrons at @xmath177agev are shown .
we find the same features as for @xmath175agev concerning the strangeness enhancement and the comparison between the two parameter sets for the equation of state ; the differences are slightly smaller at this energy . at midrapidity ,
both protons and pions are very slightly over - estimated in all explored scenarios which suggests that the nuclear stopping is still a bit overestimated . finally , at the top sps energy @xmath171agev ( fig .
[ rapidity3 ] ) the csr does not play a significant role , since the dynamics is dominated by the qgp phase .
thus , there is no appreciable difference between the results with and without csr for the two different eos .
our results for @xmath178 and @xmath176 are lower with respect to the experimental data , however , the @xmath173 and @xmath179 as well as the protons are correctly reproduced .
it is presently unclear where these final differences stem from , since strangeness conservation is exactly fulfilled in the phsd calculations . in this section
we show the phsd results for the transverse mass spectra of protons , pions and kaons for different energies in central au+au collisions in figs .
[ mtp1 ] and [ mtm1 ] and pb+pb collisions in figs .
[ mtp2 ] and [ mtm2 ] in comparison with ags and sps data , respectively .
we do not show the third scenario ( phsd calculations including csr with nl1 ) since the difference with respect to the parameter set nl3 for csr is practically not visible .
agev in comparison to the experimental data from ref .
the solid ( red ) lines show the results from phsd including csr with nl3 parameters , the dotted ( blue ) lines show the results from phsd without csr.,scaledwidth=45.0% ] agev in comparison to the experimental data from ref .
the coding of the lines is the same as in fig .
[ mtp1].,scaledwidth=45.0% ] agev in comparison to the experimental data from refs .
we show the results from phsd including csr with nl3 parameters by solid ( red ) lines and those from phsd without csr by dotted ( blue ) lines.,scaledwidth=48.0% ] agev in comparison to the experimental data from ref .
the coding of the lines is the same as in fig .
[ mtm1].,scaledwidth=48.0% ] at ags energies ( fig .
[ mtp1 ] ) , i.e. @xmath180agev , our calculations for the proton spectra show the same trend as the experimental data . however , we observe that the computed spectra are softer than the experimental data in this energy regime .
in fact , our results over - estimate the data at low transverse mass @xmath42 and under - estimate the data at larger @xmath42 .
note , however , that in our present calculations hadronic potentials have not been included .
we leave the study of explicit hadronic potentials in the propagation of the degrees - of - freedom to a future work .
we notice that csr produces no change in the transverse mass spectra of the protons , both at ags energies ( fig .
[ mtp1 ] ) and at sps energies ( fig .
[ mtp2 ] ) . at larger energies , i.e. @xmath181agev ( fig .
[ mtp2 ] ) , the phsd results are in a good agreement with the experimental data for protons . in the latter cases ,
a sizeable volume of the system performs a phase transition to the qgp and the final particle spectra are not sensitive to hadronic potentials anymore because the baryon densities in the final hadronic phase are rather low . in figs .
[ mtm1 ] and [ mtm2 ] we display the transverse mass spectra for pions and kaons in central au+au collisions at ags energies and in pb+pb collisions at sps energies , respectively .
we focus on the role played by the csr on the mesons transverse mass spectra . at the lower energies , @xmath182agev there is no appreciable difference between the calculation with and without csr , since the energy density reached by the system is not high enough to produce a vanishing scalar quark condensate . instead ,
in the energy range @xmath183agev , we notice a small difference between the scenarios with and without csr .
as already mentioned , the csr acts directly on the chemistry and not so much on the dynamics of the schwinger mechanism , thus the effect of the partial restoration of chiral symmetry is rather small on the transverse mass spectra .
the kaon spectra are harder when csr is included , while the pion spectra remain essentially unchanged . at the higher sps energies
@xmath184agev the dynamics of the system is ruled dominantly by the qgp phase and our calculations do not show any sensitivity on the inclusion of csr .
the agreement of our phsd calculations with the data in figs .
[ mtm1 ] and [ mtm2 ] is good in all cases studied .
even if the focus of this work is the study of csr , we point out that additional kaon potentials might modify this picture at low energies .
in particular , the attractive potential for @xmath179 in the hadronic phase should improve our calculations at @xmath185agev producing a softening of the spectra .
we will report on the effect of hadronic potentials in a forthcoming study . in this subsection
we study the excitation function of the particle ratios @xmath4 , @xmath186 and @xmath187 at midrapidity from 5% central au+au collisions . in fig .
[ horn ] we show the calculations for the following three scenarios : the default phsd without csr ( blue dotted line ) , phsd including csr with nl3 and nl1 as parameter sets for the nuclear eos from the non - linear @xmath1 model ( red solid and green dashed lines , respectively ) .
the shaded area displays the uncertainties of our calculations from the two scenarios for the nuclear eos . ,
@xmath186 and @xmath187 at midrapidity from 5% central au+au collisions as a function of the invariant energy @xmath0 up to the top sps energy in comparison to the experimental data from @xcite .
the coding of the lines is the same as in fig .
[ rapidity3 ] .
the grey shaded area represents the results from phsd including csr taking into account the uncertainty from the parameters of the @xmath1-model for the eos.,scaledwidth=48.0% ] as already described in ref .
@xcite , the inclusion of csr in phsd is responsible for the strong strangeness enhancement at ags and low sps energies .
the experimental observations of the ratios @xmath4 and @xmath187 show the well - known `` horn '' structure , which is reproduced by the phsd calculations with csr .
in fact , csr gives rise to a steep increase of these ratios at energies lower than @xmath188gev , while the drop at larger energies is associated to the appearance of a deconfined partonic medium .
as anticipated by the considerations in sec .
[ eos ] , the nl1 parameter set produces a sharper peak both in the @xmath4 and in the @xmath187 excitation functions with a @xmath189 maximum increase with respect to the nl3 result .
we point out that even adopting different parametrizations for the @xmath1 model , we recover the same `` horn '' feature .
this supports the reliability of the csr mechanism as implemented in the phsd model . at ags energies ,
the energy dependencies of the ratios @xmath4 and @xmath187 are closely connected , since @xmath176 and @xmath64 ( or @xmath190 ) are mostly produced in pairs due to strangeness conservation . on the other hand , the excitation function of the @xmath186 ratio does not show any peak , but it smoothly increases as a function of @xmath0 .
in fact , especially at ags energies , the antikaon production differs substantially from the production of @xmath176 and @xmath64 , which occurs dominantly via string formation .
in fact , the antikaons are produced mainly via secondary meson - baryon interactions by flavor exchange and their production is suppressed with respect to the @xmath64 hyperons that carry most of the strange quarks .
this is the reason why the inclusion of chiral symmetry restoration provides a substantial enhancement of the @xmath4 and @xmath187 excitation functions and a smaller change on the @xmath186 ratio .
we also notice that there is no sizeable difference between the nl1 and nl3 results for the @xmath186 ratio . at top sps
energies the strangeness is produced predominantly by the hadronization of partonic degrees - of - freedom , thus our results for all the ratios do not show an appreciable sensitivity to the nuclear eos and the calculations with and without csr tend to merge at @xmath191gev . finally , in fig .
[ hyperons2 ] we present the yields of @xmath173 and @xmath192 at midrapidity from 5% central au+au collisions as a function of the invariant energy @xmath0 in comparison to the available data from refs .
@xcite . and @xmath192 at midrapidity from 5% central au+au collisions as a function of the invariant energy @xmath0 up to the top sps energy in comparison to the experimental data from refs .
the coding of the lines is the same as in fig .
[ horn].,scaledwidth=48.0% ] we recover a `` horn '' structure , similar to that shown in fig .
[ horn ] for the energy dependence of the strange to non - strange particle ratios .
a sensitivity on the nuclear model parametrizations persists at low energy , while in the top sps energy regime the results corresponding to the different scenarios merge .
the comparison with the available data at @xmath193gev supports the validity of the csr picture , while at larger energies we under - estimate the experimental observations .
we mention that this discrepancy is not due to the csr mechanism , since it does not play an essential role in the high - energy regime as pointed out above .
in this section we explore new aspects of csr in heavy - ion collisions .
first , we analyze the dependence of the strange to non - strange particle ratios on the size of the colliding system ( cf . also ref .
@xcite ) . second , we investigate the effects of csr on the strange particle yields for different centralities of au+au collision . in fig . [ horn_systems ]
we present the particle ratios @xmath4 , @xmath186 and @xmath187 from phsd for three types of collision systems .
the aim here is to explore how the variation of the system size modifies the excitation functions shown in the previous section . in fig .
[ horn_systems ] we display the calculations for @xmath194au + @xmath194au in blue , for @xmath195ca + @xmath195ca in green and for @xmath196c + @xmath196c in red . ,
@xmath186 and @xmath187 at midrapidity from 5% central symmetric a+a collisions as a function of the invariant energy @xmath0 .
the solid lines show the results from phsd including csr with nl1 parameters , the dashed lines show the result from phsd without csr .
the blue lines refer to au+au collisions , the green lines to ca+ca collisions and the red lines to c+c collisions.,scaledwidth=48.0% ] the scenarios considered are the default phsd without csr ( dashed lines ) and phsd including csr with nl1 as parameter set for the nuclear equation of state from the non - linear @xmath1 model ( solid lines ) .
the inclusion of csr gives a strangeness enhancement also in case of smaller system size with respect to au+au collisions and this holds for all three particle ratios .
in fact , when considering central collisions , a sizeable volume of the system is affected by the partial restoration of chiral symmetry even in case of light ions .
we notice that , for the @xmath4 , @xmath186 ratios , the discrepancy between the calculations with and without csr remains sizable even at high sps energies for ca+ca and c+c collisions .
in particular the spread between the scenarios with and without csr is larger when the size of the system is smaller .
this can be explained by the fact that in ca+ca and c+c collisions the fraction of the system , which performs the phase transition to the qgp , is smaller with respect to au+au collisions , and the string excitations and decays still have a large strangeness production rate even at larger energies .
these characteristics are evident also in the observation that at large energies the ratio @xmath4 is smaller for the au+au collisions and larger in c+c collisions .
in fact , we recall that the drop of the @xmath4 ratio in fig .
[ horn ] is due to the appearance of the qgp , since the strangeness production in the qgp phase is suppressed with respect to the hadronic production at fixed energy density . concerning the `` horn '' structure in the @xmath4 ratio , we notice that the peak of the excitation function becomes less pronounced in case of ca+ca and it disappears completely in case of c+c collisions . with decreasing system size
the low energy rise of the excitation functions becomes less pronounced .
we can see also that the peak for ca+ca is shifted to larger energies with respect to the au+au case .
differently from the @xmath4 , the @xmath187 ratio preserves the same structure for all three colliding systems . in order to produce @xmath64s the threshold energy of @xmath197gev ( for @xmath190 @xmath198gev )
must be reached , so the @xmath187 ratio increases when the system easily exceeds this value .
the peak of the @xmath64 production is not exactly in correspondence of the threshold energy , since we are considering a+a collisions where the available collision energy is distributed among participants and where secondary and even higher order interactions take place .
however , it is interesting to notice that the peak position in this excitation function does not move for different systems , different from the @xmath4 ratios . at large energies the @xmath187 ratio decreases as a function of the energy , since the pion production is enhanced in the hadronic re - scattering .
finally , we observe no peak structure in the energy dependence of the @xmath186 ratio in any of the scenarios studied .
we notice that the results for the different sizes of the system present an opposite hierarchy with respect to the @xmath4 and the @xmath187 ratios .
in fact , for c+c and ca+ca collisions the pion production is suppressed , since in the small systems the hadronic re - scattering can not develop as in au+au collisions . furthermore , in fig . [ e30yields ] the abundances of pions , kaons and the most abundant hyperons are plotted as a function of the number of participants @xmath199 at midrapidity from au+au collisions at 30agev and in fig .
[ e30ratios ] the ratios @xmath4 , @xmath186 , @xmath52 and @xmath200 are shown for the same collision configuration . in both cases we show the calculations from phsd including csr with nl3 as parameter set by solid red lines and the calculations from phsd without csr by dotted blue lines . , @xmath176 , @xmath179 , @xmath158 and @xmath192 at midrapidity from au+au collisions at 30agev as a function of the number of participants .
the solid ( red ) lines show the results from phsd including csr with nl3 parameters , the dotted ( blue ) lines result from phsd without csr.,scaledwidth=45.0% ] , @xmath186 , @xmath52 and @xmath200 ( the last one increased by a factor of 10 ) at midrapidity from 5% central au+au collisions at 30agev as a function of the number of participants .
the coding of the lines is the same as in fig .
[ e30yields].,scaledwidth=45.0% ] all particle yields decrease with decreasing number of participants . on the other hand ,
the ratios appear to be almost constant as a function of the centrality for @xmath201 50 .
only the @xmath200 ratio smoothly decreases with decreasing @xmath202 .
the inclusion of csr produces a strangeness enhancement in the whole range of centralities investigated .
note , however , that very peripheral reactions are not considered for @xmath201 50 . at @xmath177agev the interaction volume of the two colliding nuclei reaches high energy densities such that practically all central cells are influenced by the csr mechanism independently of the centrality of the collision .
future heavy - ion collision experiments are expected to shed further light on the dynamics of the chiral symmetry restoration by exploring these kind of additional observables .
the aim of this section is to study which parts of the phase diagram in the ( @xmath8)-plane are probed by heavy - ion collisions with special focus on the strangeness production . in general
it is not straight forward ( or even impossible ) to connect non - equilibrium dynamics from microscopic transport studies to macroscopic equilibrium properties like temperature and chemical potentials . for this purpose
one needs the exact qcd - equation of state that relates the energy and the conserved charges to temperature and the chemical potentials of an equilibrated system .
as long as lattice calculations at finite chemical potential are prevented by the sign - problem one has to rely on effective models , making a study of the phase diagram model dependent .
another issue is the question if the system reaches a local equilibrium during the heavy - ion collision .
a common method to decide on kinetic equilibration is the pressure equilibration of the energy momentum tensor @xmath203 . in the local rest - frame it takes the form @xmath204 where @xmath103 is the energy density and @xmath205 , @xmath206 and @xmath207 are the pressure components in @xmath43- , @xmath208- and @xmath209-direction . in the center of the collision
they are often labeled as @xmath210 and @xmath211 , when the beam is in the @xmath209-direction . due to the initial asymmetry of the collision
the longitudinal and the transverse pressure differ significantly .
a necessary requirement for kinetic equilibrium is the coincidence of the pressure components @xmath212 .
the behavior of the pressure components in the central region of the collision zone has been studied in ref .
@xcite with the urqmd transport model @xcite .
it was found that the pressure equilibrates at @xmath213fm / c after the initial impact of central au+au collisions at ags energies .
additionally , a good agreement between the energy spectra of different hadron species with the predictions of statistical models was found at this time .
this indicates that one can indeed find an equilibrated system at ags energies for times larger than @xmath115 10 fm / c . to fix points in the ( @xmath8)-plane we have to determine the temperature @xmath214 and the baryon chemical potential @xmath9 of the medium in the expanding fireball
this is usually done by comparing the energy density and the conserved charges in the local cell to the corresponding equation of state @xcite . for hadronic matter
it is common to use a hadron - resonance gas equation of state .
however , it is unclear which hadronic resonances should be included in such a model .
we will therefore determine a temperature @xmath214 and a baryon chemical potential @xmath9 from the energy density and particle density of nucleons and pions which are directly accessible within our transport simulations . instead of examining the whole fireball we will focus on local cells where strangeness is produced ; these cells may be close to thermal equilibrium ( at late times ) or out - of equilibrium ( at early times ) .
the focus on local cells with strangeness production excludes free streaming cells as well as everything that happens after chemical freeze - out .
whenever a new @xmath215-pair is produced in the hadronic medium we take the nucleon and pion energy densities @xmath216 and @xmath217 and determine the temperature @xmath214 and the baryon chemical potential @xmath9 using the expressions for a noninteracting hadron gas ( in equilibrium ) : @xmath218 where @xmath219 is the energy of the respective particle and @xmath220 and @xmath221 are the degeneracy factors of pions and nucleons . by additionally evaluating the densities @xmath222 and @xmath223 we can check if the local cell is in approximate thermal equilibrium or not . in practical terms
: if the temperatures obtained from both methods differ by more than @xmath224mev or if the chemical potentials differ by more than @xmath225mev we consider the cell to be out - of equilibrium . within this procedure
we can eliminate further cells from the phase diagram , which are out - of equilibrium , however , it does not change the probed area if one considers only events that happen @xmath226fm / c after the collision .
agev for different time intervals .
each point belongs to a cell where strange quarks were produced .
the color of the points indicates the time of the events within some varying interval . for times
@xmath227 10 fm / c the strangeness production occurs in cells that are in approximate thermodynamic equilibrium while the cells in panels ( a ) and ( b ) are dominantly out - of equilibrium . ]
10 fm / c .
each point belongs to a cell where strange quarks were produced .
the color of the points indicates the time of the events within some varying interval . ]
[ pic : pd_vs_time ] shows the reconstructed temperatures and chemical potentials for a au+au collision at @xmath228agev with an impact parameter @xmath229fm .
this reaction is clearly dominated by local cells with partonic content since hadrons in qgp cells dissolve since the energy density is above critical .
each point in fig .
[ pic : pd_vs_time ] stands for a local cell in which new strange quarks where produced .
the color of the points indicates the time of the production ( red points early , yellow and green points at intermediate times , blue points at the end of the time interval ) .
the panel ( a ) shows the events from @xmath230fm / c after the initial collision .
these points cover all chemical potentials up to @xmath231mev .
the maximum temperature is @xmath232mev when adopting only baryons and pions as degrees - of - freedom .
these cells in panel ( a ) are dominantly out - of thermodynamic equilibrium and for low @xmath9 and high @xmath214 correspond to partonic cells with some pion content .
the panel ( b ) shows the events from @xmath233fm / c where the majority of the points are located at baryon chemical potentials larger than @xmath234mev .
the blue points in panel ( b ) belong to cells that are already in equilibrium .
the panel ( c ) shows the events that happen at times @xmath235fm / c after the initial collision and are approximately in equilibrium .
they cover an area between @xmath236 and @xmath237 .
the panel ( c ) describes indeed a proper phase diagram while the upper two ( a ) and ( b ) suffer from non - equilibrium effects .
we stress again that the temperatures and baryon chemical potentials that are shown in the panels ( a ) and ( b ) can not be related to any equilibrium properties . in spite of the restrictions imposed by thermodynamic equilibrium we find that even central au+au collisions at 10.7 agev explore a wide range of points in the @xmath8-plane for @xmath9 essentially below 650 mev .
[ pic : pd_vs_energy ] , furthermore , shows the occupation of the phase diagram extracted from au+au collisions at @xmath238 ( a ) , @xmath239 ( b ) and @xmath240agev ( c ) with an impact parameter @xmath229fm .
all points belong to events that happened at times @xmath241fm / c after the initial collision .
one sees that the probed region shifts to larger baryon chemical potentials and smaller temperatures when lowering the beam energy .
the maximum baryon chemical potential in these plots is @xmath242mev , the lowest temperature @xmath243mev , however , with a very large spread in @xmath214 and @xmath9 .
it is important to discuss how these occupancies in the phase diagram relate to the real qcd - phase diagram . due to
the model dependent eos the extracted temperatures and chemical potentials do not represent the real ones for qcd .
we recall that the phsd transport approach uses a critical energy density of @xmath244gev/@xmath20 to distinguish between a hadronic and a partonic medium .
if the local energy density is above this threshold the hadrons dissolve into quarks .
the critical energy density @xmath21 marks the largest energy density a hadronic system can reach in our simulations .
when compared to recent lattice results from the wuppertal - budapest collaboration @xcite at @xmath245 this translates to a temperature of around @xmath246mev . on the other side
the model used to extract the temperature and chemical potential is based on the hadron resonance gas with a reduced number of degrees - of - freedom .
we mention that a hadron resonance gas ( hrg ) , that contains all the hadronic particles included in phsd , reaches the critical energy density at a temperature @xmath247mev .
this indicates that the temperatures shown in figs .
[ pic : pd_vs_time ] and [ pic : pd_vs_energy ] are too large compared to full qcd . as a rough guide one should divide the temperatures in these figures by a factor @xmath115 1.1 in order to obtain an estimate closer to full qcd .
for the baryon chemical potential we note that the baryon number susceptibilities @xmath248 of the hrg are smaller than the lattice results @xcite .
this implies the corresponding baryon densities , in first order given by @xmath249 , exceed the hrg densities , thus overestimating the extracted baryon chemical potentials @xmath9 in comparison to full qcd .
admittedly we can not give a definite rescaling for finite chemical potentials , nevertheless , the general trend should be the same in the whole @xmath250 -plane shifting the probed area to smaller temperatures and chemical potentials .
nevertheless , it becomes apparent from fig .
[ pic : pd_vs_energy ] that it will be very hard to identify a critical point in the ( @xmath8)-plane experimentally since the spread in @xmath214 and @xmath9 is very large at all bombarding energies of interest .
in this work we have analyzed the effects of chiral symmetry restoration ( csr ) on observables from heavy - ion collisions in the energy range @xmath15gev .
our results have been obtained within the parton - hadron - string dynamics transport approach @xcite , where essential aspects of csr have been incorporated in the schwinger mechanism for the string decay @xcite .
since the phsd approach includes both hadronic and partonic degrees - of - freedom and has been tested in a wide energy regime , it represents a powerful tool to study nucleus - nucleus collisions on a microscopic basis .
the csr , as implemented in phsd , affects only the hadronic particle production and it does not imply modifications in the quark - gluon plasma ( qgp ) phase .
we have found that csr induces an enhancement of the strange quark fraction @xmath38 produced via the string decay , while we predict no sensible changes in the diquark production and accordingly in baryon - antibaryon production .
the @xmath142 ratio , as defined by the schwinger formula ( [ schwinger - formula ] ) , increases as a function of the energy density due to csr and this is reflected in an enhancement of the strange particle abundances with respect to the non - strange ones .
this has been observed explicitly in the particle spectra at ags and lower sps energies ( @xmath251 to 30agev ) . on the other hand , at top sps energies ( e.g. @xmath171agev ) the results from phsd with and without csr merge , since the dynamics of the system is dominated by the qgp phase where csr does not play a significant role .
we found that the transverse mass spectra are only slightly modified by the inclusion of the csr .
in fact , the csr mechanism acts predominantly on the chemistry and not on the kinematics of the string decays .
we stress that our phsd calculations provide a microscopic interpretation of the `` horn '' structure in the excitation function of the @xmath4 ratio in central au+au ( or pb+pb ) collisions .
the steep rise of this ratio at ags energies is associated to csr , while the drop at higher sps energies is due to the appearance of the qgp phase in an increasing volume of the interaction region .
we have found an analogous energy dependence for the @xmath187 ratio , while the excitation function of the @xmath186 ratio does not show any explicit peak .
furthermore , we have studied the sensitivity of our results on the parametrizations for the nuclear equation of state .
in fact , we adopted the non - linear @xmath1 model to compute the baryon scalar density @xmath68 in the space - time cells of the system .
the scalar density @xmath68 is an essential ingredient of the csr mechanism within the phsd and it is needed for the calculation of the scalar quark condensate in the nuclear medium ( [ scalar_condensate ] ) .
there is a moderate sensitivity of the particle ratios and of the rapidity spectra at ags and lower sps energies to the nuclear eos , in particular the nl1 parameter set gives a stronger strangeness enhancement relative to the nl3 set .
in general , the phsd results obtained with the inclusion of csr are in a good agreement with the available data for all observables analyzed , while calculations without csr fail substantially .
we have investigated also different sizes of the colliding ions ( @xmath194au , @xmath195ca and @xmath196c ) and computed the strange to non - strange particle ratios for these configurations with and without csr .
it is found that the `` horn '' feature in the @xmath4 ratio appears only for larger system sizes , i.e. au+au and ca+ca , while the `` horn '' disappears in case of c+c collisions .
the @xmath187 excitation function maintains the peak - structure as we have observed in case of au+au collisions also for smaller sizes of the system .
furthermore , we have analyzed the strange particle abundances in au+au collisions at @xmath252agev as a function of the number of participants in the collision .
as mentioned above , when including csr in the phsd calculations we obtain an increase of the strange particle yields with respect to the results from phsd without csr and this feature is valid in case of central collisions as well as moderate peripheral collisions .
more experimental observations are needed to extract information about the centrality and system size dependence of the csr .
we also have addressed the question whether the strangeness production in hic occurs in thermodynamical equilibrium or not and have found that strange particles are produced dominantly at the early stages of the collisions , when the system is not in thermal and chemical equilibrium . at ags energies
only a few percent of the total strangeness production happens in approximate local thermodynamical equilibrium .
with decreasing bombarding energy lower temperatures and higher baryon chemical potentials are reached , however , the spread in @xmath214 and @xmath9 is very large such that a search for a critical point in the phase diagram becomes very difficult experimentally . in conclusion ,
our microscopic studies support the idea that csr occurs in hadronic systems with high temperatures and densities before the deconfinement phase transition takes over .
we suggest that the strange particle spectra and yields are suitable signatures to study the properties of csr in hics in future also as a function of system size and centrality .
the authors acknowledge inspiring discussions with j. cleymans , m. gazdzicki , m. gorenstein and o. linnyk and thank the helmholtz international center for fair ( hic for fair ) , the helmholtz graduate school for hadron and ion research ( hgs - hire ) , the helmholtz research school for quark matter studies in heavy - ion collisions ( h - qm ) , and the bundesministerium fr bildung und forschung ( bmbf ) for support .
the computational resources have been provided by the loewe - csc .
e. l. bratkovskaya _
_ , phys . rev .
c * 69 * , 054907 ( 2004 ) .
e. l. bratkovskaya , s. soff , h. stcker , m. van leeuwen , w. cassing , phys .
* 92 * , 032302 ( 2004 ) .
w. cassing , a. palmese , p. moreau and e.l .
bratkovskaya , phys .
c * 93 * , 014902 ( 2016 ) . c. s. fischer , j. luecker , and c. welzbacher , phys .
d * 90 * , 034022 ( 2014 ) . c. s. fischer , l. fister , j. luecker , and j. m. pawlowski , phys .
b * 732 * , 273 ( 2014 ) .
g. eichmann , c. s. fischer , and c. a. welzbacher , phys .
d * 93 * , 034013 ( 2016 ) .
t. k. herbst , j. m. pawlowski , and b .- j .
schaefer , phys .
d * 88 * , 014007 ( 2013 ) .
l. mclerran and r. d. pisarski , nucl .
a * 796 * , 83 ( 2007 ) .
y. nambu and g. jona - lasinio , phys . rev . *
122 * , 345 ( 1961 ) .
e. l. bratkovskaya , w. cassing , v. p. konchakovski and o. linnyk , nucl .
a * 856 * , 162 ( 2011 ) .
o. linnyk _
et al . _ ,
c * 92 * , 054914 ( 2015 ) ; phys .
c * 89 * , 034908 ( 2014 ) ; phys .
c * 88 * , 034904 ( 2013 ) ; phys .
c * 87 * , 014905 ( 2013 ) ; phys .
c * 85 * , 024910 ( 2012 ) ; phys .
c * 84 * , 054917 ( 2011 ) ; nucl .
a * 855 * , 273 ( 2011 ) . v. koch , int .
e * 6 * , 203 ( 1997 ) .
j. v. steele , h. yamagishi , and i. zahed , phys .
b * 384 * , 255 ( 1996 ) ; phys .
d * 56 * , 5605 ( 1997 ) .
g. e. brown , prog . theor . phys . * 91 * , 85 ( 1987 ) .
g. e. brown , c. m. ko , z. g. wu , and l. h. xia , phys
c * 43 * , 1881 ( 1991 ) . v. koch and g. e.
brown , nucl . phys .
a * 560 * , 345 ( 1993 ) .
b. friman , w. nrenberg and v. d. toneev , eur .
j. a * 3 * , 165 ( 1998 ) .
m. bando , t. kugo and k. yamawaki , phys .
rep . * 164 * , 217 ( 1988 ) .
t. d. cohen , r. j. furnstahl , d. k. griegel , and x. jin , prog . part .
. phys . * 35 * , 221 ( 1995 ) .
j. boguta and a. r. bodmer , nucl .
a * 292 * , 413 ( 1977 ) .
w. cassing , nucl . phys .
a * 700 * , 618 ( 2002 ) .
j. stachel , nucl .
a * 610 * , 509c ( 1996 ) .
s. albergo _ et al .
_ , phys .
lett . * 88 * , 062301 ( 2002 ) . c. alt _ et al . _
[ na49 collaboration ] , phys .
c * 77 * , 024903 ( 2008 ) ; c. alt _ et al . _
[ na49 collaboration ] , phys .
c * 73 * , 044910 ( 2006 ) ; c. alt _ et al .
_ [ na49 collaboration ] , phys . rev .
c * 78 * , 034918 ( 2008 ) .
s. v. afanasiev _ et al . _
[ na49 collaboration ] , phys .
c * 66 * , 054902 ( 2002 ) ; t. anticic _ et al .
_ [ na49 collaboration ] , phys . rev .
c * 83 * , 014901 ( 2011 ) ; t. anticic _ et al . _
[ na49 collaboration ] , phys .
lett . * 93 * , 022302 ( 2004 ) .
l. ahle _ et al . _
[ e866 and e917 collaboration ] , phys .
b*476 * , 1 ( 2000 ) ; ibid . * 490 * , 53 ( 2000 ) ; phys .
c * 58 * , 3523 ( 1998 ) .
j. l. klay _ et al . _
[ e895 collaboration ] , phys . rev .
c*68 * , 054905 ( 2003 ) .
l. v. bravina _ et al .
_ , nucl . phys .
a * 661 * , 600 ( 1999 ) .
s. a. bass _ et al .
. part .
phys . * 41 * , 255 ( 1998 ) .
m. bleicher _ et al .
_ , j. phys .
g * 25 * , 1859 ( 1999 ) .
s. endres , h. van hees and m. bleicher , phys .
c * 93 * , 054901 ( 2016 ) .
s. borsanyi , z. fodor , c. hoelbling , s. d. katz , s. krieg and k. k. szabo , phys . lett .
b * 730 * , 99 ( 2014 ) . | we study the effect of the chiral symmetry restoration ( csr ) on heavy - ion collisions observables in the energy range @xmath0=320gev within the parton - hadron - string dynamics ( phsd ) transport approach .
the phsd includes the deconfinement phase transition as well as essential aspects of csr in the dense and hot hadronic medium , which are incorporated in the schwinger mechanism for the hadronic particle production .
we adopt different parametrizations of the nuclear equation of state from the non - linear @xmath1 model , which enter in the computation of the quark scalar density for the csr mechanism , in order to estimate the uncertainty in our calculations . for the pion - nucleon @xmath2-term
we adopt @xmath3 45 mev which corresponds to some world average. our systematic studies show that chiral symmetry restoration plays a crucial role in the description of heavy - ion collisions at @xmath0=320gev , realizing an increase of the hadronic particle production in the strangeness sector with respect to the non - strange one .
we identify particle abundances and rapidity spectra to be suitable probes in order to extract information about csr , while transverse mass spectra are less sensitive .
our results provide a microscopic explanation for the `` horn '' structure in the excitation function of the @xmath4 ratio : the csr in the hadronic phase produces the steep increase of this particle ratio up to @xmath5 7 gev , while the drop at higher energies is associated to the appearance of a deconfined partonic medium .
furthermore , the appearance / disappearance of the horn-structure is investigated as a function of the system size and collision centrality .
we close this work by an analysis of strangeness production in the ( @xmath6)-plane ( as extracted from the phsd for central au+au collisions ) and discuss the perspectives to identify a possible critical point in the phase diagram . |
in the last decades , an increasing interest has arisen towards analytic tools able to create singular bundles for interesting energies , starting with donaldson s breakthrough linking @xmath2-dimensional topology to the study of @xmath3-bundles which are critical points of the @xmath4-norm of the curvature studied by uhlenbeck @xcite .
this study is especially interesting in the case of supercritical dimensions , where it gave rise to new questions , among which is the one regarding the possibility of constructing new nontrivial bundles satisfying the yang - mills equations by minimizing the natural energy under some topological constraint ; see @xcite , @xcite and the references therein . for nonabelian bundles in supercritical dimensions ,
a setting in which the direct method of the calculus of variations can be applied to obtain critical bundles with singularities is not yet well established ( see however the definitions in @xcite , @xcite ) .
together with tristan rivire , we started @xcite an approach which should give the _ suitable setting for posing a minimization problem for weak bundles in supercritical dimensions _
, in the abelian case of complex @xmath5-line bundles . since the yang - mills energy in the nonabelian case is the @xmath4-norm of the curvature , we considered the energies defined by @xmath0-norms also in the abelian case . under such constraint
a suitable setting should consist of the following ingredients : * _ a class of weak bundles which is closed by sequential weak-@xmath0 convergence of the curvatures : _ since as we said the natural energy is the @xmath0-norm of the curvature , the topology giving precompactness of sublevelsets is the weak @xmath0-topology . in particular any minimizing sequence will have a weakly convergent subsequence , and a suitable class of bundles should be closed under this topology .
* _ a suitable notion of boundary value : _ if @xmath6 is the curvature of our bundle , we want to be able to state the minimization problem which could be formally written as follows @xmath7 in a meaningful way . in particular , we would like the weak convergence in the previous point not to disrupt our boundary condition , and to reduce to the usual boundary restriction for locally smooth bundles . as we briefly describe in sections
[ closureintro ] and [ closuresec ] , the first point above was solved by the main theorem of @xcite .
the solution of the second point is one of the main results of the present work ( see section [ defbdry ] ) another possible approach for the creation of nontrivial bundles which are critical for our energy is by minimizing a _ relaxed energy _ instead ,
as suggested in @xcite and @xcite , and in analogy with the case of harmonic maps @xcite . in our case
a good candidate for such energy would for example be given by @xmath8 consider the aim of making a minimization problem as in rigorous for supercritical @xmath5-bundles .
the natural topological invariant of @xmath5-bundles is the first chern class , @xmath9 , which , for an @xmath5-bundle @xmath10 over a compact surface @xmath11 is expressible via chern - weil theory as @xmath12 where @xmath6 is a curvature on @xmath10 ( see@xcite ) . by identifying the lie algebra @xmath13 with @xmath14
, we can identify @xmath6 with a @xmath14-valued @xmath15-form on @xmath11 .
then , in the `` supercritical '' dimension @xmath1 , a @xmath15-form can be interpreted as a curvature if it gives integer volume to almost every closed surface ( that integer corresponds to the @xmath9 of a line bundle restricted to the surface ) . in @xcite
the following class was first defined : [ defnfzabel ] we call an @xmath0-integrable @xmath15-form @xmath6 defined on a @xmath1-dimensional domain @xmath16 a * curvature of a weak line bundle with group @xmath5 * , if for all @xmath17 and for almost all @xmath18 such that @xmath19 , there holds @xmath20 where @xmath21 is the inclusion map .
we call @xmath22 the class of such @xmath6 .
[ vfif ] we could associate to a @xmath15-form @xmath6 the vectorfield @xmath23 satisfying @xmath24 so that @xmath25 , @xmath26 being the outer normal to @xmath16 and @xmath27 being the oriented unit volume @xmath15-form on @xmath28 . via this correspondence
, we can identify curvatures of @xmath0 weak @xmath5-bundles as in definition [ defnfzabel ] with _ @xmath0-vectorfields having integer fluxes _ through almost all spheres . note that in definition [ defnfzabel ] no assumption is made a priori , regarding the existence of an underlying topological bundle structure , and we only concentrate on the datum present in our target minimization problem , namely the curvature form @xmath6 , while the presence of an underlying bundle is witnessed just by the chern class requirement .
this is the natural setting where to construct new bundles by minimizing the energy , because the supercritical case is precisely characterized by the possibility of creation of topological singularities , and imposing an initial smooth structure precludes this possibility .
the sets along which we slice , in definition [ defnfzabel ] , are just spheres . by the density result of @xcite however , it follows from such definition that automatically such integrality condition is valid on all codimension @xmath29 closed generic surfaces . in a similar way ,
the same result for generic surfaces can be achieved starting from a definition which involves slicing sets different from spheres , e.g. cubes , or some other family of sets becoming arbitrarily fine at each point , and allowing a similar density result .
the above definition [ defnfzabel ] gives a description of bundles in terms of their slices on spheres , and therefore a suggestive parallel can be made with the theory of _ scans _ present in @xcite,@xcite , @xcite .
a first fruit of this parallel is the idea leading to the proof of the closure theorem in @xcite : [ closureweakbdl ] suppose @xmath30 as in definition [ defnfzabel ] are weakly convergent to some @xmath15-form @xmath6 . then @xmath31 .
the main achievement of the present paper is the definition of the boundary trace in section [ defbdry ] . on one hand such definition
gives a nontrivial trace , while on the other hand it is preserved under weak convergence .
this is due to the intervention of the slice distance directly in its definition , together with the properties described in section [ regandconv ] .
an interesting consequence of that definition is the following result on the existence of minimizers for the problem : [ minimexist ] consider @xmath32 . fix an @xmath0-integrable @xmath15-form @xmath33 on @xmath34 having integer degree .
then the infimum in problem with is achieved , if we interpret the boundary condition @xmath35 as @xmath36 , with notations as in section [ defbdry ] .
fix a minimizing sequence @xmath37 in the class @xmath38 .
up to extracting a subsequence we may suppose that @xmath39 , which by weak semicontinuity of the @xmath0-norm has energy at most equal to the infimum in problem .
theorem [ closureweakbdl ] gives then the fact that @xmath40 , while lemma [ boundarys ] of section [ defbdry ] gives @xmath36 as wanted .
the same result and proof hold in the case of domains @xmath41 such that @xmath16 is just bilipschitz equivalent to a smooth domain @xmath42 . in that case
we will have to first define the distance analogous to our @xmath43 on slices along the sets @xmath44 which foliate in the usual way a tubular neighborhood of @xmath45 , then using that distance define the boundary trace exactly as in section [ defbdry ] .
the proof of hlder dependence on the parameter @xmath46 proceeds as in section [ regandconv ] , and this is enough to prove the results in section [ defbdry ] . proceeding as in section [ lipslices ] we
can then obtain the same result for @xmath16 , simply by using the bilipschitz equivalence .
the fact that minimizers in have finitely many singularities in any compact @xmath47 is proved in @xcite .
the boundary regularity ( and thus the fact that singularities of minimizers are isolated ) will be addressed in @xcite .
we now build an analogy between @xcite ( see also the more recent development @xcite ) , @xcite and the case treated here .
consult @xcite for the notations of the next paragraph .
+ recall that in @xcite normal currents @xmath48 on a metric space @xmath49 were identified by the property that their slices by lipschitz functions @xmath50 , having values in the space @xmath51 of @xmath52-dimensional currents @xmath53 were metric bounded variation ( mbv ) functions , where @xmath51 is endowed with the flat metric @xmath54 for mbv slice functions it was then proved that the union of the atoms of the slices constituted a _
rectifiable _ set .
such construction of a rectifiable set tailored on a normal current was the main step for the later closure theorems .
the proof of the rectifiability was based on the following estimate valid for @xmath55 functions @xmath56 , where @xmath57 is a weakly separable metric space : @xmath58 where @xmath59 is a @xmath60-function related to @xmath61 ( see @xcite , p.42 for a precise definition ) . + in our case , we slice along spheres , as in definition [ defnfzabel ] .
consider a form @xmath31 , and a ball @xmath62 .
then if @xmath63 is the identification by dilation and translation , we can define the slice function as @xmath640,r[\to \mathcal y , \quad \rho\mapsto i_{x,\rho}^*f,\ ] ] and the natural choice for the space @xmath65 is indicated again by definition [ defnfzabel ] : @xmath66 we could also define the slice function on the space of all balls @xmath67 by defining @xmath68 .
this function takes almost everywhere values on @xmath65 .
the choice of the suitable distance on @xmath65 is actually justified by the need of a property like .
the picture justifying the definition of the distance depends on the density result of @xcite ( see @xcite , section 6 ) , and is based on the following result : [ densitykessel ] suppose that @xmath31
. then there exists a @xmath29-dimensional rectifiable current @xmath69 such that @xmath70 for a constant @xmath71 independent of @xmath6 and @xmath72 for all @xmath73 .
the situation between two slices centered at @xmath74 is shown in figure [ fig : slices ] .
therefore we would expect that the property that @xmath75 coincides with @xmath76 in the weak sense be preserved by some projection along the radial segments connecting the slices .
given this picture and this idea , the definition of the following analogous of the flat distance is justified : let @xmath77 .
we define then the following function @xmath78 where the infimum is taken over all triples given by a @xmath0-integrable @xmath29-form @xmath79 , an integer @xmath29-current @xmath69 of finite mass , and an @xmath80-ple of couples @xmath81 , where @xmath82 and @xmath83 . in @xcite
it was proved that @xmath43 as above defines a distance , and satisfies an inequality like , where @xmath59 is replaced by a @xmath84-function depending on @xmath6 and on the ratio of the two radii defining the slices . the closure theorem [ closureweakbdl ]
was proved using the interplay between this metric and the function @xmath85 in the next section we discuss some definitions of metrics on slices , relying mainly on the density result @xcite . in section [ regandconv ]
we prove that the slice function @xmath86 is actually hlder with respect to the above distance , and that it metrizes the weak sequential convergence on slices of bounded norm .
a parallel is drawn with the classical case of slices of @xmath87-vectorfields . in section [ lipslices ]
we show how to extend the distance to different kinds of slices , where the space @xmath88 in the definition of @xmath65 is replaced by a lipschitz domain .
this result is used to give a suitable definition of the boundary trace for curvatures belonging to @xmath22 in section [ defbdry ] . then , in the final section , we state several open questions and indicate some further directions of research .
i wish to thank my advisor prof .
tristan rivire for introducing me to the topics of this paper and for the many enlightening discussions we had on the topics of this paper .
i would also like to thank prof .
bernd kirchheim for pointing out the proof of lemma [ kirchlemma ] .
we will compare here the distance on @xmath89 defined by @xmath90 as in the introduction , to the following function : @xmath91 we define also the following function , in the struggle to free our distance @xmath43 from the presence of an unknown sum of dirac masses : @xmath92\subset a , \:a\text { open } , |a|\leq \epsilon\right\}.\ ] ] the motivations for introducing these objects are as follows : 1 .
sometimes in applications , as for example in section [ lipslices ] , it is easier to deal with definitions in terms of finite , rather than infinite , sets of `` topological '' singularities .
this justifies the introduction of @xmath93 .
it is natural to ask whether or not our distance @xmath43 is induced by a norm on @xmath94 .
candidates for such norms would then be norms which `` do nt see small sets '' , in particular they should not be sensible to the presence of the singular measures defining @xmath43 . more importantly , having an underlying norm could perhaps help to define new and more natural notions of critical points for our energy . investigating
the relationship between @xmath43 and @xmath95 seems a reasonable first step in that direction of research .
we will use the following density result : [ density ] fix an exponent @xmath96 and consider the space @xmath97 consisting of all @xmath0-integrable @xmath29-forms @xmath79 on @xmath88 such that @xmath98 where @xmath69 is an integer rectifiable @xmath29-current of finite mass on @xmath88 .
then its subspace @xmath99 given by the @xmath29-forms which are smooth outside a discrete ( thus finite ) set , is dense in the @xmath0-norm .
[ eqdistances ] on @xmath65 we have @xmath100 .
we clearly have @xmath101 since the infimum defining @xmath93 is taken on a smaller class . to prove the opposite inequality , fix @xmath102 and consider a minimizing sequence @xmath103 as in the minimization defining @xmath43 .
we then have @xmath104 we consider the function @xmath105 satisfying the following equation : @xmath106 by standard elliptic theory we have that @xmath107 .
it follows that @xmath108 and proposition [ density ] applies then to @xmath109 , giving us a decomposition @xmath110 where @xmath111 is smooth outside of a finite set and @xmath112 in @xmath0-norm .
in particular there exists a measure @xmath113 of the form @xmath114 as in the definition of @xmath93 , for which @xmath115 therefore @xmath116 thus @xmath117 are competitors in the infimum defining @xmath118 , and as @xmath119 , their @xmath0-norms converge to @xmath120 .
this concludes the proof of @xmath100 .
+ a possible choice for the set @xmath122 in the definition of @xmath95 ( if we interpret @xmath122 as the set on which @xmath123 `` avoids '' as much @xmath0-norm of @xmath124 as possible ) could be some neighborhood of a superlevelset of @xmath125 , which gives us a third distance @xmath121 : @xmath126 @xmath95 is a distance and for @xmath77 there holds @xmath127 .
the inequality @xmath127 follows easily once we know from proposition [ eqdistances ] that @xmath128 , since we can take as the set @xmath122 a small neighborhood of the singularities in the definition of @xmath93 .
in particular , it follows that @xmath129 .
being the triangular inequality and the symmetry evident for @xmath95 , and since @xmath130 follows directly from the lebesgue continuity property of @xmath0-forms , we deduce that @xmath95 is indeed a distance .
the other inequalities are still to be investigated : is it true that @xmath131 ? [ rembcs ]
we mention here an interesting analogy . a simpler distance similar to @xmath95
was studied in @xcite , where for probability measures @xmath132 on @xmath133 bounded open with smooth boundary the following distance was defined : @xmath134 for a class of functions @xmath135 including the case @xmath136 . the connection between our distances and the class of distances @xmath137 would give an interesting connection to the theory of optimal transportation , which would strongly echo with the use of basic optimal transportation for `` minimal connections '' connecting singularities of harmonic maps in @xcite .
if for functions @xmath138 we have that @xmath139 is bounded , @xmath140 and @xmath141 weakly in @xmath0 , then @xmath142 .
we observe that from the weak convergence it follows @xmath143 for all @xmath144 .
we define then the potentials @xmath145 by @xmath146 and we observe that @xmath147 satisfies @xmath148 and @xmath149 .
+ now take @xmath29-forms @xmath150 such that @xmath151 where @xmath152 is a sum of dirac masses with integer coefficients , and let @xmath153 .
call also @xmath154 and observe that @xmath155 is bounded and @xmath156 .
then @xmath157 and denoting @xmath158 we obtain @xmath159 therefore ( see the main theorem of @xcite ) we can find @xmath160 such that @xmath161 , where @xmath162 is the normalized volume @xmath29-form of @xmath163 .
the end of the proof goes as in @xcite , proposition 3.4 , where at the level of the @xmath164 it is possible to find a converging subsequence , and by sard theorem it is concluded that for a rectifiable @xmath29-current of finite mass @xmath165 there holds @xmath166 , thus @xmath167 .
see @xcite for the details .
[ metrizes ] if @xmath168 are equibounded in @xmath0 , then @xmath169 using the fact that a sequence has a limit @xmath170 if and only if each subsequence has a subsequence converging to @xmath170 and the previous lemma , we obtain immediately the `` @xmath171 '' implication .
+ suppose now @xmath172 .
then take the potential such that @xmath173 . by the elliptic estimates and the rellich - kondrachov theorem , after extracting a subsequence , @xmath174 in @xmath175 .
the limit is zero independent of the subsequence , so @xmath176 satisfies @xmath177 which implies @xmath178 .
here we consider a form @xmath179 and we want to compare its slices along @xmath180 .
this will be possible only under some condition on @xmath181 m but we formulate the condition later .
+ the slices will be given by a function ( defined a.e . )
@xmath182 , where @xmath183 is the function on @xmath88 corresponding to the restriction of @xmath184 to @xmath185 , after a homothety and an identification @xmath186 .
+ we consider the following function @xmath187\to \omega$ ] : @xmath188\sigma:=x_t+r_t\sigma.\ ] ] suppose that @xmath122 is a diffeomorphism onto its image .
then @xmath189)$ ] , and we will build out of it a competitor for the infimum in the definition of @xmath190 .
consider @xmath191 here @xmath192 indicates the component of @xmath193 parallel to the volume form of the sphere @xmath194 .
along the lines of proposition 4.1 of @xcite ( see also proposition [ densitykessel ] and figure [ fig : slices ] in the introduction ) , we can show that this gives a competitor .
we then introduce the reparameterization @xmath195 and we compute : @xmath196 in order to compare this with the norm of @xmath184 , we first find @xmath197id_{ts^2}|x - x ' + ( r - r')\sigma\right)=\left(r_tid_{ts^2}|x - x ' + ( r - r')\sigma\right).\ ] ] then ( assuming @xmath198 for the moment ) we pull back the function @xmath199 : @xmath200 we can now formulate our hypothesis on the slices : @xmath201 under this hypothesis , ( since @xmath202 ) we can estimate @xmath203 observe that @xmath204 is the poincar dual of @xmath205 .
we introduce the function @xmath206 and we can finally write @xmath207 combining the basic estimate above for a couple of segments , we obtain hlderianity . [ holder ] the slice - function @xmath208 defined above is hlder-@xmath209-continuous on with respect to the distance @xmath43 , and its hlder constant is bounded by the @xmath0-norm of @xmath184 .
we want to see how our estimates worsen if instead of connecting @xmath210 along a segment , we use a polygonal curve .
consider then @xmath211 , consisting in a union of segments @xmath212 , each ow which satisfies ( h ) .
for a given segment @xmath213 $ ] ( where @xmath214 is the end with the largest radius ) we call @xmath215 . for a given @xmath86 , call @xmath216 the difference of the radii of @xmath217 .
we then have the following estimate , following the same reasoning as above : @xmath218 and summing up and using the triangle inequality , @xmath219 because of this estimate , the question is how we can join @xmath220 by some polygonal @xmath211 which stays in the allowed set @xmath221 and is made of segments verifying ( h ) , such that @xmath222 is as small as possible and @xmath223 is bounded above .
+ we will see that @xmath80 can be bounded by @xmath2 because we do nt need more than @xmath2 segments , and that @xmath224 is bounded by @xmath225 ( also in this case it s optimal to have a few long segments rather than many short ones ) . we just briefly describe the kind of @xmath211 we use for the estimates .
+ the worst case that we can face is the one where @xmath220 are on @xmath226 , have the same @xmath227-coordinate , and are as far from each other as possible . if they are on the part where @xmath228 with @xmath229 then we can take @xmath211 to start from @xmath230 and go up in the @xmath227-direction with slope @xmath15 until it touches @xmath231 , then down until close to @xmath52 radius and center @xmath232 , then do the same symmetrically , building up an @xmath233-shaped graph .
if @xmath234 , then it s better to first go down then up , making a symmetric @xmath235-shaped graph .
if instead @xmath236 then again a @xmath235-shaped graph is the best option , and if @xmath227 is large enough a @xmath237-shaped graph will be even better . + it is easy ( but tedious ) to verify that the above constructions verify the estimate on @xmath216 .
we thus end up with the following bound : @xmath238 [ distancefig ] we observe that in general , even though @xmath43 is hlder on the slices , proposition [ metrizes ] does not apply , to give weak continuity on the slices , because the norm boundedness is not verified .
this is already clear in the case where the form @xmath6 is the radial form @xmath239 .
then consider the slices @xmath240 along @xmath241 $ ] ( see figure [ slicesunbound ] ) .
since these spheres look almost flat near @xmath242 for small @xmath243 and the integral of @xmath6 on the portion of a given slice just depends on the solid angle covered by that region , we easily see that the @xmath0-norm of the slice @xmath240 on a small ball near the singularity grows like @xmath244 , i.e. blows up . propositions [ holder ] and [ metrizes ] allow a simplification of the proof of the closure theorem [ closureweakbdl ] , which avoids using the stronger theorem 5.1 of @xcite .
we state here the crucial result from which theorem [ closureweakbdl ] follows at once , and we give a new proof of it .
let @xmath2451,3/2[$ ] as above .
suppose that the @xmath15-forms @xmath30 are weakly convergent to a @xmath15-form @xmath246 .
given @xmath247 , consider the slice functions go @xmath6 , @xmath248\to l^p(s^2)$ ] , given by @xmath249 .
then for almost all @xmath250 $ ] , @xmath251 , i.e. the integer flux condition is preserved .
we suppose w.l.o.g . that @xmath232 . by lower semicontinuity of the norm
, we may suppose that @xmath252 . by proposition [ holder ] , the slice functions @xmath253 defined as @xmath86 in the statement of the lemma , but with @xmath254 instead of @xmath6 , are equi - hlder with respect to our metric @xmath43 .
this means that we can extract a pointwise convergent subsequence .
+ it is evident that the deformation factor of the @xmath0-norm , coming from the fact that @xmath255 are dilations , is bounded .
fubini s and chebychev s theorems imply then , that we may restrict to a subset of @xmath256 $ ] on which the @xmath0-norms of the @xmath257 stay bounded .
this is just the situation where proposition [ metrizes ] applies .
therefore the slices converge weakly almost everywhere , and testing them on the constant function @xmath29 , we see that their integer degrees also converge .
therefore @xmath258 has integer degree on @xmath88 , as wanted .
we consider here the problem of extending the definition of the distance @xmath43 to the case of slices different from spheres .
the main motivations for this extension are the following : * a natural question regarding the class @xmath259 is whether or not the condition that the integrality is required on spheres can be replaced by a condition on different kinds of surfaces .
a particularly interesting case would be one in which the slicing sets tile space , as is the case for the surfaces of cubes . *
the definition of the boundary condition in section [ defbdry ] will be based on slicing .
therefore having more general slice models will allow defining the trace on more general domains . given a bilipschitz map @xmath260 , we thus define the following distance between @xmath0-integrable @xmath15-forms on @xmath11 : @xmath261 we observe that the pullback by bilipschitz functions preserves the integrability class , since @xmath262 and the same holds with @xmath263 instead of @xmath264 .
analogous estimates imply that different bilipschitz maps induce equivalent distances : [ bilipeq ] suppose @xmath265 are bilipschitz maps . then @xmath266 are distances and they are equivalent : @xmath267 moreover the constant @xmath71 depends only on the lipschitz constants of @xmath268 .
the fact that @xmath266 satisfy the triangular inequality and the reflexivity follow at once from the analogous properties of @xmath43 .
the nondegeneracy @xmath269 is a consequence of the equivalence stated in the proposition , for the case @xmath270 , and @xmath271 .
we also observe that if we prove the proposition for this special case , the general case will follow by transitivity of the equivalence between distances .
thus we will consider just this case .
+ we will work with the equivalent definition of @xmath93 as in section [ eqdef ] . + fix @xmath272 , and consider a competitor @xmath79 in the definition of @xmath118 .
in other words we have , in the @xmath15-form setting , that if @xmath102 and @xmath11 represents a finite sum of dirac masses , then @xmath273 where @xmath274 represents the hodge star operator with respect with the standard metric .
the crucial observation is that all the objects above extend naturally to the space of lipschitz functions , and it is equivalent to use just @xmath275 instead of @xmath276 above . by replacing @xmath33 by @xmath277 and changing variable , we obtain ( recall that @xmath278 is the image measure ) : @xmath279 since @xmath264 is bilipschitz , it is a bijection of @xmath280 into itself , and thus we see that @xmath281 is a competitor for the distance @xmath282 . + now observe as above that @xmath283 , which leads to the conclusion that @xmath284 the same holds also with @xmath263 instead of @xmath264 , so the infimum in the definition of @xmath43 is comparable with the one in the definition of @xmath285 .
let @xmath286 be an open bounded smooth domain .
we consider here the class @xmath22 as described in section [ lipslices ] .
such class consists of all @xmath0-integrable @xmath15-forms @xmath6 such that for generic @xmath15-cycles @xmath86 bilipschitz - equivalent to @xmath88 , there holds @xmath287 in this section we would like , given a smooth @xmath15-form @xmath288 on @xmath289 , to find a suitable class @xmath290 , which satisfies the following three conditions : * * ( closure)*[wellp ] for any @xmath0-regular @xmath15-form @xmath288 on @xmath289 , the class @xmath290 is closed by sequential weak @xmath0-convergence . * * ( nontriviality)*[nontr ] if @xmath291 are two @xmath0-regular @xmath15-forms on @xmath289 , then @xmath292 . * * ( compatibility)*[compa ] for any smooth @xmath15-form @xmath288 , @xmath293 are exactly the @xmath15-forms @xmath294 such that @xmath295 , where @xmath296 is the inclusion map . for general @xmath0-forms ( i.e. without the restriction of belonging to @xmath259 )
no such class can exist , even if in the closure requirement above we had required _
strong _ convergence .
indeed , let @xmath297 be different smooth forms , and consider @xmath298 , f_n=\chi_{[1/n,\infty[}$ ] .
then @xmath299 satisfy @xmath300 . then by compatibility @xmath254 and @xmath6 should have the same trace , and so by closure @xmath301 and @xmath6 should have the same trace , contradicting nontriviality .
+ at the other extreme , for _ locally exact _
@xmath0-forms , using the poincar lemma , we have @xmath302 thus we can impose the boundary condition directly on the restrictions to @xmath28 of @xmath303-regular `` local primitives '' @xmath122 , using classical trace theorems , and we easily obtain all the above properties . + our new space @xmath22 is an intermediate space between the two extrema above , escaping both the above reasonings .
we therefore use the approach which is a natural consequence of @xcite , namely we use the distance @xmath304 between @xmath15-forms on cycles @xmath86 , as in section [ bilipeq ] . up to applying a bilipschitz deformation , we may assume that we have @xmath32 , and we will define the boundary condition in this case first .
+ we use the distance @xmath43 to compare the boundary datum with the slices of our forms belonging to @xmath259 .
we call @xmath305 with variable @xmath306 the form on @xmath88 corresponding to the restriction to @xmath307 of the form @xmath6 .
we define the right class @xmath308 via the continuity requirement @xmath309 it is clear that the definition satisfies the _ nontriviality _ and _ compatibility _ conditions above , since @xmath310 is a distance and since for @xmath311 having smooth boundary datum implies that in a neighborhood of @xmath34 the slices are smooth and converge in the smooth topology to @xmath288 .
the validity of the _ well - posedness _ is a bit less trivial , therefore we prove it separately .
[ boundarys ] if @xmath312 are converging weakly in @xmath0 to a form @xmath40 then also @xmath6 belongs to @xmath313 . by weak semicontinuity of the @xmath0 norm
we have that @xmath254 are bounded in this norm .
we then have ( since @xmath314 ) for all @xmath315 : what we need is precisely @xmath316 . +
therefore by proposition [ holder ] the @xmath254 are @xmath43-equicontinuous , so a subsequence ( which we will not relabel ) of the @xmath254 converges to @xmath317 , i.e. for all @xmath318 $ ] the forms @xmath319 are a cauchy sequence for the distance @xmath43 .
this is enough to imply that @xmath320 .
we observe that @xmath6 is just defined up to zero measure sets , but it has a @xmath43-continuous representative . by continuity
it is clear that @xmath6 still satisfies .
the same proof also gives an apparently stronger result : [ boundarys2 ] if @xmath321 are converging weakly in @xmath0 to a form @xmath40 then the forms @xmath322 converge with respect to the distance @xmath43 to a form @xmath288 and also @xmath6 belongs to @xmath313 . at a previous stage in the preparation of this paper , before remarking the @xmath43-hlderianity of the slices , we had prepared a different definition of the boundary value , based on the @xmath84-bound of the modulus of lipschitzianity of @xmath43 given in theorem 5.1 in @xcite .
assuming just such bounds , with @xmath96 , proposition [ boundarys2 ] remains true , if we replace condition by the following approximate continuity requirement @xmath323\cap a_\epsilon|}{\rho}=0,\text { where } a_\epsilon:=\{\rho':\:d(f(\cdot+\rho),\varphi))>\epsilon\}.\ ] ] we give the result also in the formalism of vectorfields with integer fluxes mentioned in remark [ vfif ] : let @xmath324 be the class of vectorfields with integer fluxes described in remark [ vfif ] .
let @xmath325 be the radial vectorfield defined outside the origin of @xmath326 .
for @xmath3270,1[$ ] , define @xmath328 . for a given @xmath0-regular function @xmath33 defined on @xmath34
we define then the class @xmath329 via the continuity requirement @xmath330 where we identify @xmath15-forms on @xmath88 to functions via the hodge - star duality with respect to the standard metric .
+ with this definition we have the following two properties : 1 .
if @xmath331 are converging weakly in @xmath0 to a form @xmath332 then also @xmath23 belongs to @xmath333 .
2 . if @xmath334 are converging weakly in @xmath0 to a form @xmath332 then the forms @xmath335 ( where @xmath274 is the hodge star with respect to the standard metric ) converge with respect to the distance @xmath43 to a function @xmath336 and @xmath23 belongs to @xmath337 .
the correspondence described in remark [ vfif ] translates the language of forms into that of vectorfields . under this translation
the restriction operation @xmath338 corresponds to the operation @xmath339 .
the closure of @xmath324 under weak convergence being proved in @xcite , we are just left to prove the preservation and convergence of the boundary condition .
the needed results are proved in the language of differential forms in lemma [ boundarys ] and in proposition [ boundarys2 ] respectively . as already pointed out
, the definition of the distance as in section [ lipslices ] allows to extend the definition of the boundary value to arbitrary domains
since we used slices to define our class of weak curvatures @xmath340 , it is natural to try to go one step further and try to represent a form @xmath341 by its slices .
this kind of problem seems to represent an unexplored area of research , related perhaps to integral geometry .
we were not able to find any example of similar problems in the literature .
therefore in the following subsections we will attempt to formalize the main questions which have arisen .
consider a @xmath15-form @xmath342 which is bounded in @xmath0-norm . given a lipschitz @xmath15-cycle @xmath343 on @xmath16 , chosen in a `` generic '' way such that @xmath344 is in @xmath345 and that ( in the duality between @xmath15-cycles and @xmath15-forms )
@xmath346 we can associate @xmath347 where @xmath348 is the set of @xmath15-forms @xmath349 such that we have still to explain what the requirement that @xmath349 be defined only for `` generic '' cycles should mean .
for that purpose , denote by @xmath353 the _ fixed set of lipschitz cycles _ on which our compatibility theory will be defined ( useful choices may vary from the set of all spheres to the set of all lipschitz cycles ) .
the domain of definition of @xmath349 above should then given by @xmath354 for some set @xmath355 , possibly depending on @xmath6 , which belongs to an admissible class of residual sets defined as follows .
* suppose that @xmath357}$ ] is a lipschitz foliation by lipschitz cycles @xmath358 i.e. there is a lipschitz cycle @xmath359 and a bilipschitz parameterization @xmath360\to\cup_{x\in[-\epsilon,\epsilon ] } c_x$ ] sending @xmath361 to @xmath362 .
then there exists @xmath363 such that the intersection with @xmath353 of set of cycles @xmath362 corresponding to choices of @xmath74 inside a subset of of zero lebesgue measure of @xmath364-\delta,\delta[$ ] , should form a set belonging to @xmath365 .
consider the question of slice compatibility for a class of slicing cycles @xmath353 .
the starting observation is that even in simple cases , not all applications @xmath366 can be represented as slices @xmath349 of an underlying form @xmath367 : [ notcomp ] assign to each cycle @xmath368 $ ] the form @xmath369 equal to @xmath370 of a fixed nonzero @xmath15-form @xmath371 , where @xmath372 is the similitude bijection .
the so - obtained function @xmath373:\:x\in\omega , r\in]0,{{\rm{dist}}}(x,\partial\omega)\}\to\mathcal y_{\mathcal c}\ ] ] can not satisfy @xmath374 for generic @xmath359 .
assume for a moment that there exists such @xmath15-form @xmath31 .
then for any fixed @xmath375 we would have @xmath376 almost everywhere on @xmath16 . indeed fix @xmath377 such that @xmath378 , where @xmath379 .
then consider the sets @xmath380 , with the constraint @xmath381 .
these sets form a fine covering of @xmath16 , and if @xmath382 for almost all @xmath71 in the definition of @xmath383 , then @xmath384 must be larger than @xmath233 almost everywhere on @xmath385 . by extracting a ( not necessarily disjoint ) countable cover of @xmath16 by sets @xmath385 up to zero lebesgue measure , we obtain that @xmath376 almost everywhere . by the arbitrariness of @xmath233 we obtain that @xmath6 can not be in @xmath0 , thus contradicting our assumption . [ star ] suppose that @xmath353 is a family of cycles such that for almost all @xmath17 the tangent spaces @xmath386 span the grassmannian @xmath387 of @xmath15-planes .
then for any @xmath388 as in there is at most one @xmath15-form @xmath6 such that @xmath389 .
indeed , fixing the restrictions @xmath390 at some point @xmath74 along three linearly independent tangent planes relative to three choices of @xmath71 , automatically fixes the value of @xmath6 at @xmath74 .
the compatibility requirement between @xmath388 and @xmath6 following from remark [ star ] depends on the pointwise behavior of the single slices .
we would like to find a more geometric condition @xmath391 which can be tested by looking only at the function @xmath388 as in .
see question [ star5 ] for such an example . the wanted condition @xmath391 should also satisfy the following properties . 1 . if @xmath342 for some @xmath2451,3/2[$ ] then the function @xmath392 which to a generic @xmath393 associates the slice of @xmath6 along @xmath71 , satisfies @xmath391 .
if @xmath394 are a sequence converging @xmath0-weakly to a form @xmath6 , then also @xmath395 satisfies @xmath391 .
3 . whenever @xmath388 satisfies @xmath391 ,
there exists a @xmath342 such that @xmath396 .
[ tooweak ] in @xcite , a @xmath391 satisfying the first condition was given , and consisted in asking that for @xmath397 ( where the generic sets are the algebra generated by the ones of the form @xmath398 s.t .
@xmath399 ) , the integral of @xmath400 be an integer .
this is just the definition of @xmath259 .
as shown by lemma [ notcomp ] , this candidate for condition @xmath391 is too weak to satisfy the second property above .
we consider still the case where @xmath353 consists of all spheres contained in @xmath16 .
if we imagine that our form @xmath6 has only finitely many singularities , then the integral @xmath401 along each cycle corresponds to an algebraic sum of the degrees associated to the singularities situated in the interior of @xmath71 .
now consider two intersecting spheres , @xmath402 and suppose that their intersection is a circle @xmath403 .
if we assume that none of the singularities of @xmath6 is on @xmath404 , we will have then that near @xmath403 the forms @xmath405 can be represented respectively as @xmath406 , for suitable @xmath29-forms @xmath407 .
it is easy to see by using stokes theorem that the difference @xmath408 must then be an integer , and must equal the algebraic sum of the degrees of all the singularities contained inside @xmath409 .
it is thus natural to formulate the following compatibility condition more in general : @xmath410 it is easy to see that condition @xmath411 of definition [ compatdef ] is satisfied , while condition @xmath412 will probably be achievable using the techniques leading to the closure theorem [ closureweakbdl ] .
the third condition is however still to be investigated .
we thus formulate the following example [ tooweak ] suggests considering a stronger form of condition @xmath391 than just the requirement that spherical slices have integer degree . in order to give a second candidate for a compatibility condition , we will now suggest how to extend the class @xmath353 here , to include all boundaries of bounded sets writable as finite intersections of balls and of complements of balls ( since we are interested in the boundaries , and just finite intersections are involved , it is not relevant whether we use closed or open balls ) .
we will call such boundaries * convex spherical polyhedra * , in analogy with the case when balls are replaced by half - spaces .
consider a convex spherical polyhedron ( where @xmath414 and @xmath415 are balls included in @xmath16 ) : @xmath416.\ ] ] a natural notion of genericity , which is also easily seen to be admissible according to definition [ residual ] , can defined as follows : if @xmath417 then generic sets of perturbations of @xmath71 will be the ones formed by @xmath418,\ ] ] with @xmath419r_i(1-\epsilon),r_i(1+\epsilon)[\setminus n_i$ ] , @xmath420 , and with @xmath421 , @xmath422 for all @xmath423 ( and similarly for @xmath424 ) .
+ it will also be useful to consider the natural induced cell complex structure on each cycle @xmath71 ; the dimensions of the different faces will agree with the hausdorff dimension of the underlying sets .
we will use as @xmath15-skeleton the sets @xmath425 such that @xmath426 are the connected components of the interior of @xmath427 for some @xmath230 among the ones in the definition of @xmath71 .
the lower - dimensional skeletons can be then defined by intersection .
+ given this cell complex structure , we define a spherical cell complex as in singular homology theory , the only differing feature being that all our @xmath1-cells are required to be obtained via intersections of ( generic ) balls .
we thus also have a way of making sums and differences of our cycles . given these data
, it will be enough to define the compatibility condition on couples of convex spherical polyhedra having exactly one common face , then extend the definition by taking sums to more general cases .
+ we first define a compatibility property for two spherical polyhedra having just one common face of highest dimension , then we extend this to all polyhedra .
the whole construction is done in the case of forms @xmath428 having finitely many singularities , which is a very special and easy case . + * compatibility for neighboring cells . * given @xmath429 where the @xmath15-skeletons @xmath430 have exactly one common face , we describe a candidate compatibility condition for @xmath431 and @xmath432 as follows .
* on the face @xmath433 we ask that @xmath434 a.e . * on the faces neighboring @xmath403 in @xmath435 the form @xmath436 can be expressed locally as @xmath437 , and similarly @xmath438 .
we then ask also that @xmath439 , where the orientations on @xmath440 in the two integrals are coming from the orientations of @xmath441 respectively .
* definition of our candidate condition @xmath442 . *
we define the candidate property @xmath442 for a function @xmath388 as in in the case of a form @xmath31 which is smooth outside a finite number of singularities @xmath443 and : @xmath444 for @xmath31 which has just isolated singularities , one can find local representations outside the singularities via potentials @xmath445 as above , and the integer @xmath446 is equal to the number of charges inside the sum cycle @xmath447 .
we formulate the future steps to be taken from here on , as an open question : [ [ regularity - of - critical - points - and - of - minimizers - of - the - lp - energy . ] ] regularity of critical points and of minimizers of the @xmath0-energy .
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ we have already recalled in proposition [ densitykessel ] of the introduction that to any @xmath448 we can associate an integral rectifiable current of finite mass such that @xmath449 as distributions , and the map @xmath450 is bounded with respect to the mass norms . + it is an intriguing direction of investigation to study how much of the information about @xmath6 is encoded in such an @xmath69 , and to try and use instruments of the theory of currents to study the forms in @xmath259
. we will now consider the regularity question for minimizers of the @xmath0-energy in the class @xmath259 as in , or more generally for critical points @xmath451 this equation is usually complemented by the stationarity requirement @xmath452\\ \text{such that } \forall t\:\phi_t|_{\partial\omega}=id_{\partial\omega}\text { and } \phi_0=id_\omega,\\ \text{there holds } \left.\frac{d}{dt}\int_\omega|\phi_t^*f\right|^pdx^3|_{t=0}=0 .
\end{array}\ ] ] it is tempting to imitate the blow - up - and - monotonicity ( or stationarity ) approach in order to study the singularity points , and to prove regularity results , in the spirit of @xcite , @xcite . besides the harmonic maps , another model problem for us is the regularity of minimal surfaces @xcite .
the use of the fact that the current is boundaryless in the problem of mass - minimizing integral currents ( in the case of harmonic maps , the studied quantity is a gradient , thus the property of being boundaryless is implicit there too ) , would correspond here to the use of the fact that @xmath453 is constant for a.e .
@xmath18 , which in our case is not necessarily true ; so we must content ourselves of a weaker result , relying on the above euler - lagrange equations and their analogues , and trying to use the integrality of the boundary to reduce to the boundaryless case .
one of the main steps in the study of our problem would be the description , and perhaps even the classification , of tangent maps , as is done in the case of minimizing hypersurfaces ( where the tangent spaces are seen to be hyperplanes without much effort ) , and as was achieved ( with more effort ) in the case of harmonic maps for example in @xcite .
the best - behaving local models of singularities in our case are up to rotation @xmath15-homogeneous and symmetric , but proving that they are the only possible tangent maps of critical forms @xmath6 is so far just a conjecture .
+ suppose @xmath6 is a minimizer of the energy , in particular it satisfies , , and a suitable comparison principle .
assume that the @xmath0-weak limit of a blow - up sequence @xmath454 , @xmath455 exists .
then prove that such limit equals , up to a rotation , one of the forms @xmath456 .
in particular the limit is unique . for minimizers
the conjecture is true , as will be proved in some future work .
our proof however uses the results of @xcite , i.e. the special properties available for minimizers .
note that the same kind of properties are used also in the result @xcite about harmonic maps , therefore the following question is meaningful also in that case .
supposing that the tangent maps are classified , we have still one more step to achieve until the local regularity becomes provable , at least if the classification is done in the sense of the above conjecture : we must use the information given by the tangent maps on degrees , in order to eliminate the possibility that singularities ( i.e. points of the support of the above boundary @xmath75 of the current of proposition [ densitykessel ] ) are present in the regions of small rescaled energy : this would indeed be the analogue of an @xmath243-regularity theorem , in the setting of a classical approach to regularity ( as opposed to the alternative combinatoric approach of @xcite ) .
we state the following open problem in this spirit : we spend the rest of this subsection to give some hints encouraging the idea that the answer to the above question is positive . + the above problem can transformed into an abstract question involving only the more handy current @xmath69 , rather than the mysterious form @xmath6 .
we will concentrate on the following property involving the boundary of a finite mass integral @xmath29-current @xmath69 ( the idea will be to use this property in relation to the current @xmath69 of proposition [ densitykessel ] ) .
[ kirchheimprop ] suppose that @xmath69 is an integer multiplicity rectifiable @xmath29-current of finite mass on @xmath16 , whose boundary is defined in the sense that the flux ( or average of the slice done via the distance function , in the terminology of @xcite ) @xmath457 we denote by @xmath458 the property that for all @xmath17 there exist a strictly decreasing sequence @xmath459 such that for all @xmath423 @xmath460 is well - defined and equal to zero .
[ propp ] let @xmath31 , @xmath461 , and let @xmath69 be the integral current as in the above proposition [ densitykessel ] .
suppose that @xmath6 minimizes the @xmath0-norm with constrained boundary trace , as in .
there exists then a constant @xmath462 which is independent of @xmath463 such that if @xmath464 then on the smaller ball @xmath465 the current @xmath69 has property @xmath458 . since the radius is @xmath29 and since we do not impose constraints on the size of @xmath16 , the above energy on @xmath466 can be considered to correspond to the correct scale - invariant version of energy , which in general has the form @xmath467 . in our case
( see @xcite ) the correct choice is @xmath468 . by integrating the above formula
we obtain that the rescaled energy @xmath472 is increasing in @xmath227 , where we use the notation @xmath473 .
we now use the small energy assumption on @xmath466 in order to obtain a rescaled energy bound on balls @xmath474 , with @xmath475 : @xmath476 where @xmath477 .
now suppose that @xmath458 were false ; we then have that for some @xmath74 there exists @xmath478 such that for almost all @xmath4790,r(x)[$ ] there holds @xmath480 , which implies @xmath481 we have used just the inequality @xmath482 and the jensen inequality on each @xmath483 .
the constants so far depend just on @xmath484 and on the dimension , so we can find @xmath462 independent of @xmath485 such that the two above inequalities are incompatible . thus the proof is finished .
we now describe a positive result suggesting that from proposition [ propp ] it could be possible to obtain a positive answer to the question [ questionereg ] above .
more precisely , we show that a stronger version of property @xmath458 of definition [ kirchheimprop ] is enough to obtain regularity .
this result was suggested to us by bernd kirchheim , but it will be evident that the proof does not extend to include assumptions that are as weak as in property @xmath458 .
[ kirchlemma ] let @xmath69 be an integral flat @xmath29-current of finite mass such that property @xmath458 holds , and the function @xmath486 of definition [ kirchheimprop ] satisfies a uniform lower bound @xmath487
. then @xmath488 .
the fact that @xmath489 for all test functions @xmath33 supported on some ball of radius @xmath490 , follows easily for @xmath33 of the form @xmath491 , because their superlevelsets are small enough balls .
+ we want now to prove that when @xmath492 is a general test function , @xmath493 . to this end
we start with a function @xmath494 where the radii of the @xmath495 are at most @xmath496 , and take a family of radial mollifiers @xmath497 , supported in balls of radius @xmath496 ( this can be achieved for @xmath377 small enough ) .
then since @xmath498 is radial and compactly supported in a ball of radius at most @xmath490 , we obtain @xmath499 now we claim that @xmath500 in @xmath501 implies @xmath502 in @xmath503-norm . indeed , @xmath504 and for @xmath505 we can estimate the absolute value of the above integral by @xmath506 which converges to zero as @xmath507 .
+ similarly we can prove @xmath508 : we can estimate @xmath509\rho_\epsilon(y)dy\right|\leq\omega(\epsilon)||\rho_\epsilon||_{l^1}\to 0,\ ] ] where @xmath510 is the modulus of continuity of @xmath511 .
now testing all the above convergences with @xmath69 , we obtain the wanted @xmath512 observe that for a doubling locally finite measure , to be zero on a fine covering ( as are the balls in the definition of property @xmath458 ) is equivalent to being zero , by the vitali covering theorem ( see chapter 2 of @xcite ) .
thus if we knew for example that @xmath75 were a measure , we would easily conclude .
on the other hand , in our case @xmath75 being a locally finite measure is equivalent to @xmath6 having just a locally finite number of singular points , which is the statement of the partial regularity @xcite ; we thus prefer not to use it as an assumption .
we formulate instead the following intriguing abstract question : in order not to mislead the reader , we observe that currents @xmath69 as in proposition [ densitykessel ] are _ more than just integer rectifiable of finite mass _ in ( our ) case @xmath96 , since they automatically conserve also some of the information on the higher integrability of @xmath6 .
this is the spirit of the following easy counterexample ( the idea is the same as for example 7.1 of @xcite ) . *
the slice @xmath514 exists and gives an integer when tested with the constant @xmath29 for all @xmath74 and all but at most countably many @xmath227 . * for no @xmath31 with @xmath96 is the current @xmath69 given by proposition [ densitykessel ] equal to @xmath513 .
we may suppose up to rescaling that @xmath515 where @xmath516 is a square of sidelenght @xmath15 ( this is more than needed for the rest of the proof , but makes the notations easier ) .
consider a sequence of positive numbers @xmath517 such that @xmath518 , then align a sequence of disjoint balls @xmath495 such that @xmath519 , along one of the axes of @xmath516 .
inside each @xmath495 two disjoint spheres @xmath520 of diameter @xmath517 can be packed ; then identify the oriented segment joining their centers with an integer @xmath29-current , and define @xmath69 as the sum of all these currents .
the @xmath69 has mass @xmath29 , but ( by applying jensen s inequality on smaller spheres concentric to the @xmath520 ) it is evident that @xmath6 as in the statement of the lemma must have @xmath521 .
this concludes the proof .
fabrice bethuel , ham brezis , jean - michel coron , _ relaxed energies for harmonic maps _ , variational methods ( paris , 1988 ) , progr .
nonlinear differential equations appl .
, * 4*,3752 , birkhuser boston,1990 .
thierry de pauw , robert hardt , _ application of scans and fractional power integrands _ , variational problems in riemannian geometry , progr .
nonlinear differential equations appl . , * 59 * , 1931 , birkhuser , basel , 2004 . | a slice distance for the class of weak abelian @xmath0-bundles in @xmath1 dimensions was introduced in @xcite , where it was used to prove the closure of such class of bundles for the weak @xmath0-convergence .
we further investigate this distance here , and we prove more properties of it , for example we show that it is hlder - continuous on the slices . using the same distance ,
we give here a notion of a boundary trace , giving a suitable setting for minimization problems on weak bundles .
we then state some conjectures and some open questions . |
the study of multiphoton excitation and ionization processes so far has concerned the interaction of matter with strong electromagnetic fields @xcite .
extending our previous works @xcite , we demonstrate in the present proposal that ideas and concepts developed in this field of research can also be applied for understanding the response of ultracold macroscopic matter waves in optical lattices to ac forcing with a slowly varying envelope .
such systems offer additional handles of control which are not available in experiments with atoms or molecules exposed to pulses of laser radiation .
therefore , they can give novel insights into multiphoton dynamics in general , and in particular may also allow one to systematically investigate the effects of interparticle interactions on such transitions .
the experimental investigation of bose - einstein condensates ( becs ) in time - periodically forced optical lattices has been pushed forward with remarkable vigor within the last years , addressing quite diverse topics such as parametric amplification of matter waves @xcite , dynamic localization @xcite , photon - assisted tunneling @xcite , coherent control of the superfluid - to - mott insulator transition @xcite , quantum ratchets @xcite , super bloch oscillations @xcite , quantum simulation of frustrated magnetism @xcite , controlled correlated tunneling @xcite , and even the realization of tunable gauge potentials @xcite .
thus , becs subjected to time - periodic forcing constitute an emerging branch of research @xcite .
here we outline how systematic use of the forcing s envelope will allow one to exctract crucial information about the systems dynamics .
the present theoretical study of multiphoton - like condensate dynamics refers to conditions recently realized by haller _
et al . _
@xcite : these authors have loaded becs of cs atoms into a vertically oriented 1d optical lattice @xmath0 .
here @xmath1 , where @xmath2 nm is the wavelength of the lattice - generating laser light , so that the lattice period is @xmath3 .
we choose a comparatively shallow lattice with depth @xmath4 , where @xmath5 is the single - photon recoil energy @xcite , with @xmath6 denoting the mass of the atoms .
the first excited band then is of the `` above - barrier '' type , so that atoms in this and all higher bands do not need to tunnel through the barriers when moving over the lattice .
using a magnetically induced feshbach resonance , the @xmath7-wave scattering length of the cs atoms is tuned to zero @xcite , so that one is dealing with a condensate of effectively noninteracting particles . by means of a time - periodic modulation of the levitation gradient employed for compensating gravity ,
an external oscillating force is introduced which acts on the atoms with maximum amplitude @xmath8 and frequency @xmath9 , such that their dynamics are governed by the hamiltonian @xmath10 with a dimensionless shape function @xmath11 determining the envelope of the pulses applied .
all calculations reported here are performed for the relatively low frequency @xmath12 hz , so that @xmath13 . as indicated in fig .
[ f_1 ] , this means that the gap @xmath14 between the lowest two bloch bands @xmath15 and
@xmath16 of the unperturbed optical lattice , which occurs at the brillouin zone edge @xmath17 , amounts to @xmath18 : exciting particles from the initially occupied lowest band to the first excited one requires the absorption of more than five `` photons '' . a first glimpse at the condensate dynamics under the action of a force @xmath19 is provided by bloch s acceleration theorem : the expectation value @xmath20 of an atomic wave packet in @xmath21-space evolves in time according to @xmath22 provided interband transitions remain negligible @xcite . assuming a sinusoidal force @xmath23 instantaneously switched on at time @xmath24 , this gives @xmath25 , where @xmath26 is a dimensionless measure of the driving amplitude .
thus , when considering a wave packet initially at rest in the lowest band , so that @xmath27 , the packet s center @xmath20 just reaches the brillouin zone edge when @xmath28 .
because the zone edge gives rise to zener - type transitions @xcite , one then expects pronounced excitation of higher bands .
therefore , one has @xmath29 as a rough order - of - magnitude estimate of the amplitude required for inducing multiphoton - like transitions .
this expectation is confirmed by the numerical calculations summarized in fig .
[ f_2 ] . here
we consider driving pulses @xmath30 as already incorporated into the hamiltonian ( [ eq : ham ] ) , with a smooth envelope @xmath31 the pulse length is fixed at 60 driving cycles , @xmath32 with @xmath33 . in order to model the dynamics of a noninteracting bec responding to such pulses , we start with an initial state @xmath34 made up from bloch waves @xmath35 of the lowest band , employing a gaussian momentum distribution @xmath36 ^ 2 } { 2\left(\delta k\right)^2}\right ) \label{eq : imd}\ ] ] centered around @xmath27 with width @xmath37 , and then solve the single - particle schrdinger equation by means of a crank - nicolson algorithm @xcite .
varying the maximum scaled amplitude @xmath38 from pulse to pulse , we plot the escape probability from the lattice at the end of each pulse , at @xmath39 . here
we assume that only atoms which finally still populate the lowest band remain in the lattice , since atoms which have been excited to the higher `` above - barrier '' bands tend to escape from the shallow lattice quite fast .
evidently , interband transitions start to make themselves felt at @xmath40 , and reach substantial strength when @xmath41 , confirming the above rough estimate .
thus , proof - of - principle experiments performed along these lines should establish the feasibility of using becs in driven optical lattices as novel probes for multiphoton - like transitions : take driving pulses with smooth envelopes , and measure the interband transition probability at the end of each pulse . in a series of such measurements with a constant pulse shape one
then should observe a pronounced onset of interband transitions when the maximum amplitude is successively increased .
in a second step , this approach can be employed for getting more detailed insight into multiphoton dynamics .
namely , the single - band acceleration theorem ( [ eq : act ] ) ignores an essential element : not only does the wave packet s center @xmath20 move within its band in response to the external forcing , but also the bands themselves are `` dressed '' by the drive , and therefore experience an ac stark shift @xcite . hence , initially nonresonant bands may be shifted such that their separation in energy approaches an integer multiple of @xmath42 for certain bloch wavenumbers @xmath21 , possibly leading to strong resonant interband coupling , that is , to a multiphoton resonance .
theoretically , such resonances are found by computing the quasienergy bands @xmath43 which emerge from the energy bands @xmath44 in the presence of a drive with constant amplitude .
these quasienergy bands reflect the ac - stark - shifted energy bands , projected into an interval of width @xmath45 , so that a multiphoton resonance translates into an avoided quasienergy crossing @xcite .
for example , fig .
[ f_3 ] shows the quasienergies @xmath46 with @xmath47 , which pertain to the pulses considered in fig .
[ f_2 ] , plotted versus the scaled amplitude @xmath48 .
the quasienergy originating from the ground - state energy @xmath49 of the optical lattice undergoes two well - resolved avoided crossings when @xmath50 , signaling the presence of two individual multiphoton resonances .
the observation that these `` large '' resonances begin to show up only for @xmath29 nicely relates the elaborate quasienergy approach to the previous elementary reasoning based on eq .
( [ eq : act ] ) . inspecting that elementary reasoning once again
, one expects multiphoton resonances to occur for smaller driving amplitudes when the initial packet is centered around a nonzero wavenumber , @xmath51 , since then smaller amplitudes are required for reaching the brillouin zone edge .
this expectation is confirmed by fig .
[ f_4 ] , which depicts quasienergies @xmath43 for @xmath52 , with the first resonance showing up already at @xmath53 .
experimentally , one can prepare an initial state with arbitrary @xmath54 by subjecting the condensate to a suitable `` kick '' @xcite .
thus , one should also be able to detect the resonances predicted by fig .
[ f_4 ] . to this end
, we propose a particular kind of `` avoided - quasienergy - crossing spectroscopy '' based on the use of _ asymmetric _ pulses @xmath11 . for illustration
, we assume that the rising part of such pulses be given by the first half of the envelope ( [ eq : env ] ) , with fixed switch - on time @xmath55 , while their decreasing part is described by the second half of a squared - sine envelope , but with a different switch - off time @xmath56 .
let the maximum scaled amplitude be @xmath57 . during the rising part of such a pulse
, a wave packet initially centered around @xmath58 then follows its quasienergy states adiabatically , until the instantaneous amplitude reaches the multiphoton resonance at @xmath53 visible in fig .
[ f_4 ] . then the packet undergoes a landau - zener transition to the anticrossing quasienergy state @xcite . due to the rapid switch - on of the pulse , and to the narrow quasienergy separation @xmath59 at the avoided crossing , that transition is almost complete .
thereafter , the packet again adiabatically follows the pulse envelope , until the resonance is encountered a second time when the amplitude decreases . if then @xmath60 , a major part of the wave function does not `` jump over '' the avoided crossing back to the initial state , but rather stays in the continuously connected quasienergy states
this implies that a major fraction of the condensate atoms is excited at the end of the pulse , escaping out of the lattice .
when such an experiment is performed repeatedly with fixed rise time @xmath61 while varying the switch - off duration @xmath56 , one should observe survival probabilities which drop exponentially with increasing @xmath62 , allowing one to extract the quasienergy separation @xmath63 at the avoided crossing from the drop rate by means of the known landau - zener formula for quasienergy states @xcite .
the results of a series of solutions to the schrdinger equation corresponding to this scenario are plotted in fig .
[ f_5 ] . from the slope of the numerical data we deduce a quasienergy gap @xmath64 , in agreement with the value @xmath65
read off from fig .
these findings clearly underline that this quasienergy gap @xmath66 is the actually relevant energy scale for the multiphoton transitions under scrutiny here , _ not _ the band separation @xmath67 indicated in fig .
[ f_1 ] .
there are further options offered by becs in driven optical lattices which have no match in laser - based multiphoton studies .
for example , one can switch off the driving force abruptly at any moment , and analyze the state of the wave packet at that particular instant .
specifically , we again utilize envelopes of the form ( [ eq : env ] ) , and let the amplitude rapidly drop to zero at some instant @xmath68 , with @xmath69 .
we set @xmath32 and take initial wave packets with @xmath27 , as for the previous calculations shown in fig .
[ f_2 ] , and plot the escape probabilities versus switch - off time @xmath68 in fig .
[ f_6 ] . for @xmath70
the wave packet simply follows its quasienergy states adiabatically , not encountering any of the resonances observed in fig .
[ f_3 ] . however , for @xmath71 interband transitions set in at @xmath72 , that is , when the instantaneous amplitude reaches @xmath73 ; this is due to a tiny resonance not resolved in fig .
when @xmath74 , pronounced interband transitions occur at @xmath75 , corresponding to the instantaneous amplitude @xmath76 : this is already in the regime of influence of the first of the two `` large '' resonances seen in fig .
thus , this `` pulse tracking '' strategy allows one to experimentally detect multiphoton resonances , that is , to find those values of the driving amplitude for which strong resonant interband transitions occur .
to summarize , we have argued that becs in shallow optical lattices exposed to ac forcing with frequencies of some 100 hz and smooth envelopes can be employed for mimicking multiphoton processes .
the enormous degree of controllability realizable with such set - ups enables one to obtain information not reachable with laser - irradiated crystalline solids @xcite ; in particular , we have suggested the use of asymmetric pulses for performing avoided - quasienergy - crossing spectroscopy .
moreover , we have shown how pulse tracking by abruptly switching off the driving amplitude allows one to monitor the dynamics at each moment during an individual pulse , and thus to locate multiphoton resonances between ac - stark - shifted bloch bands .
obviously , our approach also lends itself to systematic explorations of the effects of pulse shaping .
even more interestingly , one can activate interparticle interactions by suitably tuning the @xmath7-wave scattering length of the cs atoms @xcite .
hence , the detailed experimental investigation of the influence of such interactions on multiphoton transitions has come into immediate reach .
we are indebted to e. haller for detailed discussions , and for performing preliminary measurements confirming the viability of our suggestions .
we also acknowledge support from the deutsche forschungsgemeinschaft under grant no . ho 1771/6 . | we suggest that bose - einstein condensates in optical lattices subjected to ac forcing with a smooth envelope may provide detailed experimental access to multiphoton - like transitions between ac - stark - shifted bloch bands .
such transitions correspond to resonances described theoretically by avoided quasienergy crossings .
we show that the width of such anticrossings can be inferred from measurements involving asymmetric pulses .
we also introduce a pulse tracking strategy for locating the particular driving amplitudes for which resonances occur .
our numerical calculations refer to a currently existing experimental set - up [ haller _ et al .
_ , prl * 104 * , 200403 ( 2010 ) ] . |
elliptical galaxies are featureless , elliptical stellar systems , predominantly made with old stars .
they are a prime object in the currently popular hierarchical galaxy formation models under the @xmath3cdm universe , since in a cold - dark matter dominated universe , galaxies grow through merging with other galaxies , and it is well - known from simulations that merging between galaxies with comparable masses ( called major merger ) produce galaxies with the characteristics of ellipticals @xcite . therefore , studying elliptical galaxies is important to test our current theoretical understanding of the galaxy formation and the evolution mechanism .
many observational works have been performed for elliptical galaxies near and far , but our understanding on the evolution of elliptical galaxies is still far from complete . on one hand , there are pieces of observational evidence for a complex history in the evolution of elliptical galaxies in support of the hierarchical merging picture . for example , very deep imaging data reveal more than 70% of elliptical galaxies with tidally disrupted features that are indicative of their past merging history ( e.g. , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
small levels of recent star formation are found as blue cores in some ellipticals ( e.g. , * ? ? ?
* ; * ? ? ?
* ) and the uv excess light @xcite or as excessive mir fluxes @xcite .
these results break a simplistic view of the monolithic collapse model @xcite that ellipticals are old galaxies with little merging nor recent star formation activities . while these observational evidence point toward the hierarchical build - up of elliptical galaxies , other properties of ellipticals do not bode well with the merger scenario . tight scaling correlations relations
are observed for early - type galaxies , such as the fundamental plane relation ( e.g. , * ? ? ?
* ; * ? ? ?
* ) , and the color - magnitude relation @xcite .
also , the number density of early - type galaxies does not evolve as drastically as some of the simple hierarchical merging models have predicted @xcite .
the observational results that appear contradictory to each other can be reconciled by introducing dry merging in the evolution of elliptical galaxies .
dry ( or dissipation - less ) merging is a process where two gas - poor galaxies with old stellar population merge together to create a more massive , old galaxy .
although a dissipative merging between gas - rich galaxies ( wet merging ) inevitably introduces scatters in various scaling relations and too many blue ellipticals , the dry merging offers a way to avoid such a situation while allowing frequent merging activities in the hierarchical merging models .
this is possible since dry merging preserves the red colors of the already old pre - mergers , avoiding the introduction of additional scatters in the scaling relations arising from new star formation as is the case for wet merging .
dry merging is also gaining popularity as a possible mechanism to explain the observed rapid growth in sizes of early - type galaxies from high to low redshifts @xcite one of interesting consequences of dry merging can be found in the radial gradient of various properties , such as the metallicity , the age , and the colors of ellipticals .
it has been known that centers of low redshift elliptical galaxies are redder than their outskirts ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , and this radial change of color is called the color gradient .
the color gradient is in general attributed to a gradient in metallicity although a small amount of the age gradient is known to contribute to the color gradient , where low redshift ellipticals show the higher metallicities and the younger stellar ages at the center than in the outskirt . according to line indices analysis studies ,
the metallicity gradients have a mean value of about @xmath4 to @xmath5 per a decade change in radius while the slope of the age gradient is known to be of order of @xmath6 or less @xcite .
the metallicity and the age gradients can be derived from color profiles and such studies tend to implicate the metallicity gradients of order of @xmath7 @xcite , steeper than the spectroscopically determined metallicity gradients .
the origin of the discrepancy between the spectroscopic and the photometric studies is not clear , but could be related to a difference in the explored range of radius . in the spectroscopic studies , the gradients are determined within @xmath8 ( effective radius ) , while the photometric studies extend the radius coverage to a few @xmath9 @xmath8 .
when a monolithic collapse of the proto - galactic gas creates an elliptical galaxy , star formation can occur over several generations at center using the gas that is retained by a strong gravitational potential in the central part of the galaxy .
the gas is enriched by previous episodes of star formation , so stars in the central part are expected to have a high metallicity . on the other hand , the star formation truncates earlier in the outer region of an elliptical galaxy since the gas is lost easily due to a shallow gravitational potential preventing further episodes of star formation .
the net result is that an elliptical would have stars with higher metallicity and a bit younger mean stellar age at the inner region than the outer region , creating the metallicity , the age , and the color gradients that are in qualitative agreement with the observed results .
a similar situation is expected to occur after mergers of two gas - rich galaxies . however , according to the formation models of ellipticals through monolithic collapses of proto - galactic gas , the metallicity gradient is predicted to be quite steep , with @xmath10 to @xmath11 in the logarithmic value of metallicity per one dex change in radius @xcite .
simulations that implement merging activities predict that the metallicity gradient is shallower at @xmath4
@xmath5 ( e.g. , * ? ? ?
* ; * ? ? ?
dry merging modifies the above theoretical picture in the following way .
when two gas - poor galaxies with a comparable mass merge ( `` major dry merging '' ) , a re - distribution of stars dilute the metallicity gradient , therefore flattening it in the post - merger ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
an opposite trend is suggested to occur during minor dry merging , i.e. , merging between ellipticals with very different masses ( mass ratios of 510 to 1 or larger ) , where stars in a minor companion galaxy get deposited at a large distance from the center of the main pre - merging galaxy , avoiding stars to be mixed throughout the main galaxy , contrary to the case of major dry merging @xcite .
observations offer somewhat mixed views about the importance of dry merging , and the results from some of such studies are presented below .
@xcite find that color gradients of the brightest cluster galaxies ( bcgs ) are flatter that those of non - bcgs .
they conclude that a great number of dry merging events flattened c - m relation and individual bcg .
@xcite find color - gradients are flatter for elliptical galaxies in the cluster environment than in field , suggesting more frequent merging history in the high density environment .
@xcite derived the metallicity gradients of 51 nearby early - type galaxies using their spectroscopic data and find a flattening trend in the metallicity gradients toward a large galaxy mass but with a large scatter .
their interpretation of the result is that galaxies with a few @xmath9 @xmath12 are created through wet merging , but evolved via dry merging thereafter .
the increase in the scatter of the metallicity gradient is also found in other studies ( e.g. , * ? ? ?
* ) , and some works interpret that the increase in the scatter is suggested is caused by dry merging of objects with a variety of metallicity gradients ( e.g. , * ? ? ?
. however , such an interpretation requires galaxies with steep gradients to create merger products with a large scatter , and the number of such galaxies with steep gradients should decrease as galaxies experience successive dry merging events .
this will inevitably reduce the scatter in the mass - metallicity gradient relation at high mass end , which is inconsistent with the observed trend .
also , @xcite show that the large dispersion can be explained under the frame work of the monolithic collapse models where various star formation efficiencies produce the difference in gradients with the most efficient star formation creating the steepest gradient . in this paper
, we will examine the color / age / metallicity gradients and their connection to dry merging , analyzing the optical and the nir imaging data of a sample of 204 nearby elliptical galaxies that are divided into three different types ones showing no merging feature , ones with tidal - feature ( as a result of merging ) , and ones with dust - feature .
we will also present a simple model that predicts the evolution of the mass - metallicity gradient relation under major dry merging .
this model will be used to interpret our result .
the use of the optical and the nir data allows us to break the age - metallicity degeneracy that usually hampers the interpretation of colors of early - type galaxies with optical data alone @xcite . the large wavelength baseline between the optical and the nir are also advantageous in deriving color gradients which are known to be shallow in optical bands @xcite .
also , nir data , especially @xmath1-band , of an elliptical galaxy is known to be a good indicator of stellar mass , since mass - to - light ratios ( @xmath13 ) in nir are less sensitive to galaxy types or ages than in the optical @xcite . here , we will use the @xmath1-band light as a mass indicator of elliptical galaxies .
our sample will be drawn from a sample of early - type galaxies in @xcite which classifies early type galaxies based on merging features .
they are galaxies with very deep imaging data reaching down to the surface brightness ( sb ) limit of @xmath14 , enabling us to subdivide and examine elliptical galaxies into three different types that reflect different stages of merging .
the relaxed types are objects without strong merging feature . if they had undergone merging in the past , and they should have done so a long time ago ( @xmath15 several gyrs ) . presumably , such objects should have well - established gradient properties with tighter scatters .
the tidal feature types are objects that underwent merging relatively recently ( @xmath16 several gyrs ) .
they may or may not share the gradient properties similar to the relaxed types , but we can expect the gradient properties to be less established than the relaxed - types with a somewhat larger scatter . finally , the dust - feature types are objects with dust and gas , and they are suggested to be involved in a merger with a gas - rich companion ( e.g. * ? ? ? * ) .
therefore , such objects need to be considered separately when examining dry merging . throughout this paper ,
we adopt a concordance cosmology of @xmath17 , @xmath18 , and @xmath19 km sec@xmath20 mpc@xmath20 ( e.g. * ? ? ?
* ) , and ab system for magnitudes .
the elliptical galaxies that we study in this paper are selected from a catalog of nearby ( @xmath21 ) , luminous ( @xmath22 ) early - type galaxies in the sdss stripe 82 imaging area which is located along the celestial equator in the southern galactic cap ( @xmath23,@xmath24 ; @xcite ; hereafter k10 ) .
the reason why we chose the catalog is that it accompanies images that are @xmath252 magnitudes deeper than sdss single scan images , enabling us to identify faint merging features like tidal tails for sub - classification of early - type galaxies . in k10 , they classified galaxies into various morphological classes ( early - type galaxies and late - type galaxies and sa - like systems which are bulge - dominated galaxies with faint spiral features ) through a visual inspection of both monochromatic stripe 82 and multi - color sdss standard - depth image . in k10 ,
early - type galaxies are divided further into 5 types : early - types without any merging features , early - types with tidal features , dust features , both the dust and the tidal features , and early - types that are interacting . for our study
, we regrouped the early - type galaxies into three types 238 objects without merging features ( relaxed type ) , 38 objects with tidal features ( tidal - feature type ) and 22 objects that exhibit dust features including cases showing both tidal and dust features ( dust - feature type ) .
we excluded the interacting types in k10 since the number of such object is small and they in general pose a challenge during the surface brightness fitting process .
overall , the `` relaxed '' early - types are considered as objects that experienced mergers long time ago so that their merging features disappeared . the `` tidal - feature '' early - types are considered as objects that experienced a dry merger or a wet merger relatively recently ( @xmath26 a few gyrs ago ) and the early - types with dust - lanes are considered as objects that experienced a wet merging recently . figures 1
3 show example images of the three types of early - type galaxies . in the next step ,
we cross - matched the k10 early - type galaxies with the nir data from ukidss las . during this process
, we also excluded 18 elliptical galaxies that have bright objects in the vicinity ( to an extent that they can not be masked out ) since a reliable fitting of their surface brightness profiles becomes difficult under the presence of the bright neighbors .
these neighbors are foreground stars or galaxies .
example images of the early - types with bright neighbors are presented in figure 4 ( the top two panels ) . a visual inspection of
the k10 early - type galaxies reveal objects with disks ( 23% of the sample ) .
the presence of the disk makes our color gradient analysis complicated since the stellar populations in disks may significantly differ from those in bulges .
also , objects that underwent major merging events are likely to have a non - significant disk component .
therefore , we excluded early - type galaxies with a significant disk component ( which we suspect mostly are s0 s or s0a s ) .
this is done by matching the k10 early - type galaxies with those in a catalog created by @xcite which provides the bulge to total light ratio ( @xmath27 ) of sdss galaxies from a 2-dimensional surface brightness fitting using gim2d @xcite . in general ,
elliptical galaxies have @xmath28 in the optical wavelengths longer than 0.4 micron ( e.g. * ? ? ?
* ; * ? ? ?
* ) , therefore we select early - types with @xmath28 , and call them `` elliptical galaxies '' .
most of the k10 early - type galaxies are found in the catalog of @xcite , but some were not , because they are too bright and extended for the automated gim2d pipeline to handle them properly . for such early - types that are not in the catalog of @xcite , we performed a 2-dimensional surface brightness fitting using galfit @xcite , and rejected those with @xmath29 ( 5 such cases ; see 3.1 )
. see figure [ disky ] for disky early - types rejected by this procedure .
after the rejection of the disky early - type galaxies , 206 `` elliptical '' galaxies remain in the k10-las matched sample .
additionally 2 elliptical galaxies are excluded from the sample since they are too small ( the outer boundary for the color gradient range is smaller than @xmath30 arcsec ) for deriving reliable color gradients .
images of these small early - types are given in figure [ rej ] ( the bottom two panels ) . finally , we have 162 relaxed , 32 tidal - feature and 10 dust - feature type elliptical galaxies in the final sample for our analysis as indicated in table [ tbl1 ] .
we used the sdss stripe 82 imaging data for the study of optical properties .
the stripe 82 data are taken from the sdss data release 7 @xcite .
a co - addition of about 50 sdss single scans provides @xmath31 images @xmath252 mags deeper ( sb limits of @xmath14 ) than the main sdss images of individual scan .
the @xmath32-band and @xmath33-band images were downloaded from the sdss data archive server ( das ) , and they were used to derive surface brightness profiles .
we find that the optical band seeing is not uniform with the range of @xmath34 @xmath35 .
this point will become important when deriving optical - nir color gradient as we shall describe later .
the pixel scale of the sdss images is 0.3961 for the nir images , we used the data release 8 of ukidss las @xcite .
most but not all of the sdss stripe 82 region was covered ( @xmath36 ) in @xmath37-bands by ukidss las .
typical seeing of an las image is 1 fwhm in @xmath1-band .
the depth of the ukidss las data is shallower than the sdss stripe 82 data ( sb limit is @xmath38 ) , so the fitting range for an optical - nir color gradient is limited by the nir data .
we limited ourselves to the use of only @xmath1-band data , since @xmath1-band provides the maximal wavelength baseline in conjunction with the optical bands , and the use of @xmath39 or @xmath40-bands appear redundant for the stellar population analysis ( e.g. * ? ? ? * ) .
although inclusion of more passbands , especially those in uv , is likely to improve the analysis , we limit ourselves to the use of @xmath0 , and @xmath1-band data for simplicity here . the @xmath1-band images of our sample galaxies are downloaded from the wfcam science archive .
the pixel size of an las image is 0.4 , comparable to but a bit larger than the pixel size of the sdss images .
in order to investigate color / age / metallicity gradients , we derived sb profiles of each elliptical in @xmath0 , and @xmath1 ( see figure [ rel ] [ disky ] ) . the first step in the sb profile
fitting is to estimate a proper background , and to subtract it from the images .
this is because , in the outer parts of galaxies where the sky noise is dominant , sb can change substantially depending on the value of the sky background .
therefore , the background subtraction is very important in the analysis of color gradients and the derivation of sb at the outer region of galaxies .
both the sdss stripe 82 and the ukidss las data offer sky background estimates in each image , however we find that these values are only approximately accurate .
hence , we re - estimated sky background of each image with a growth curve analysis .
we carried out the growth curve analysis in the following way .
the initial guess of the background is obtained from sky value indicated in the fits image header .
the initial guess is subtracted from the image , and the growth curve is constructed using the iraf ellipse task .
we examined the growth curve out to @xmath41 the isophotal radius of the object ( typically , 15 40 arcsec in radius , with isophote at 1.5-@xmath42 of the sky noise ) , and a small amount of background value was added or subtracted manually until the background become flat at such a radius . through this process , we determine the sky background value to an accuracy of @xmath250.03% . after subtracting the background
, we used the iraf ellipse task to fit the sb profile .
we fitted the @xmath33-band images first in order to use the @xmath33-band results as a reference to compare with the other band data , because the photometric sensitivity of the @xmath33-band is highest among the sdss filters @xcite .
we fixed the center position as the position that is computed by ellipse , and the position angle and the ellipticity are set free . in some cases
, elliptical galaxies exhibit a twist of isophotes in position angle and ellipticity as a function of radius , therefore fixing these two parameters to single values often result in a poor fitting of the sb profile with the ellipse model .
we set a step size for the radial increase to be 1.5 @xmath9 the previous radius of an isophote with the starting radius of 2 pixels ( 0.8 ) .
the non - linear increase of the isophotal radius allows us to match s / n of the outer isophotes to the inner ones .
once the fitting of the @xmath33-band images is done , we fitted the @xmath32-band and the @xmath1-band images using the best - fit ellipse parameters from the @xmath33-band image .
since the @xmath33-band images have a slightly different pixel scale and a position angle in comparison to the @xmath1-band images ( in some cases ) , we modified the radial steps ( in pixel ) and the position angle values for the @xmath1-band ellipse parameters before fitting the @xmath1-band images to match those of the @xmath33-band fits .
objects in the vicinity of an elliptical galaxy were masked out by using the positional information from sextractor @xcite .
some objects that are not separated from the elliptical galaxy were masked out manually during the ellipse fitting process .
also masked out were dust features in dust - feature type ellipticals .
figure 6 shows examples of the masking of the neighbor objects .
some of the dust - feature type ellipticals have more than 50% of pixels masked out when the masking was done conservatively ( i.e. , many pixels outside the dust features were also masked ) . to see if this affected the sb - fitting significantly , we varied the fraction of pixels to be masked out by decreasing it to 30% level and performing sb fittings to such images .
we did not find a strong change in the derived sb profile through this exercise ( see figure [ mask_prof1 ] ) .
finally , sb profiles were constructed along the major axis of each ellipse .
figures [ rel ] [ disky ] show examples of the sb fitting , together with @xmath33- and @xmath1-band images of elliptical galaxies .
the derived total @xmath1-band magnitudes and the major - axis effective radii @xmath8 in @xmath1 and @xmath33 from the fitting are presented in table [ tab_sp ] .
color profiles were obtained by subtracting a sb profile in one band from a sb profile in another band . as the sdss and the ukidss / las images have different pixel scales , we linearly interpolated isophotal radii of the @xmath1-band data in @xmath43 scale , because elliptical galaxies are known to have nearly linear sb profiles in logarithmic scale as a function of @xmath43 ( de vaucouleurs law ) .
color gradient slopes were obtained by a least square fitting of a 1st order polynomial to the color profiles divided by the logarithmic radius scale in units of @xmath8 as given in eqs .
( [ eqn_cg_gr ] ) ( [ eqn_cg_gk ] ) .
the mpfit of interactive data language ( idl ) is used for the linear fit @xcite .
@xmath44 @xmath45 @xmath46 the effective radius in @xmath1-band , @xmath47 , was used for @xmath48 and @xmath49 and the @xmath50 was used for @xmath51 .
magnitude errors of color profiles at each radius were taken as sb profile errors in two bands added in quadrature .
the inner boundary of the fitting range was set to be 1.5 times the seeing fwhm of the image with the worse seeing , and the outer boundary is set to be the location where sb starts to drop below 1.5 times the background noise of the image with the shallower depth ( see 4.1 ) . note that we used sdss images with regular depths ( the depth comparable to ukidss las images ) to check if the shallower depth of the ukidss las image affect the color gradient values .
through such an exercise , we confirm that limiting the fitting range to the depth of the shallower image does not introduce systematic bias in the color gradient measurements although uncertainties in the gradient values increase as a result of the loss of s / n .
the error in the color gradient is the formal fitting error returned at the end of the least - squares fitting process .
figures [ col_prof_lm ] [ col_prof_hm ] show examples of the @xmath52 , @xmath53 , and @xmath54 color profiles in three different mass ( luminosity ) bins , together with the color gradient values . also , we present in figure [ col_prof_comb ] the median @xmath52 , @xmath53 , and @xmath54 color profiles in three different mass ( luminosity ) bins , overlaid on the color gradient profiles of all the ellipticals belonging to each mass bin .
we derived age and metallicity gradients by using a @xmath52 and @xmath53 color - color diagram .
figure [ ssp ] demonstrates the procedure used for deriving the age and the metallicity gradients , showing three example cases .
we constructed an age - metallicity grid on the @xmath52 and @xmath53 diagram using a simple stellar population ( ssp ) model of ( * ? ? ?
* hereafter bc03 ) , the salpeter initial mass function ( imf ) and the padova 1994 stellar evolutionary track were used for the ssp model . to derive the age and the metallicity gradients , locations of the points on a color profile
are plotted on the color - color diagram . by doing a bilinear interpolation on the age - metallicity grid
, we derived the age and the metallicity for each radial point .
then , the age and the metallicity gradients were calculated in a similar way to the color gradients using the following equations .
the errors for the age and the metallicity measurements were determined by a drawing 1-@xmath42 error box over each data point on the grid , and taking the largest / smallest age / metallicity values at corners of the error box to be the upper and lower 1-@xmath42 limits .
@xmath55 @xmath56 finally , we discuss how the derived age / metallicity gradients can be affected by the choice of the star formation history and the imf .
for example , a small amount of star formation may make the underlying stellar population look to be younger and more metal than they really are , if they are analyzed by an ssp model ( li et al .
figure [ csp ] shows the age - metallicity grid in @xmath52 versus @xmath53 color - color space for three models incorporating different star formation histories ( exponentially decaying star formation history and a constant burst with a burst duration of 1 gyr ) and a different imf .
as shown in figure [ csp ] , extending the duration of the star formation activity ( e.g. by adopting a longer burst period in a burst model , or @xmath57 , the exponential time scale of an exponentially decaying star formation model ) would broaden the age intervals in the age - metallicity grid in the @xmath52 direction due to an enhanced contribution of younger stars to the bluer part of the spectrum .
the age - metallicity grid expands in the @xmath53 direction too , but to a lesser extent than in @xmath52 .
on the other hand , the extended star formation activity has little effect in the metallicity grid in the color - color space .
the net effect is that the extended star formation can make the age gradient substantially shallower ( by a factor of a few or more in some cases ) than in the analysis with the ssp model , while keeping the metallicity gradient results nearly intact .
this point is made clear in figure [ grad_t_z_comp ] , where we compare the age / metallicity gradients of our ellipticals calculated from different sf history models against those from the ssp model .
we find that the average of the age gradient flattens by 50% going from the ssp model to the exponentially decaying sf model with @xmath58 gyr ( @xmath59 to @xmath60 ) , much less so in the average @xmath61 values .
the effect of the imf is mild in the color - color grid as discussed in @xcite .
a bottom - heavy imf @xcite would deform the color - color grid with respect to the salpeter imf prediction , but by a very small amount ( @xmath62 mag in color ) since low mass stars have little contribution to the integrated light of elliptical galaxies .
an example is given in figure [ grad_t_z_comp ] for the chabrier imf @xcite , showing a non - significant change in the gradient values .
therefore , we consider that the effect of the imf is negligible in the analysis of the age / metallicity gradient
. figures [ age_z_prof_lm ] [ age_z_prof_hm ] show examples of age and metallicity profiles in three different mass ( luminosity ) bins , together with the gradient values .
figure [ hist_col ] shows the marginal distributions of color gradients of different types of elliptical galaxies . also , table [ tab_col_grad ] summarizes the median color gradients of each type of elliptical galaxies .
the overall median color gradient steepens from @xmath63 to @xmath64 as we increase the wavelength base line as we expected .
the overall color gradients are in good agreement with previous measurements made by @xcite . among different types of ellipticals , the color gradient median values are very similar between the relaxed ellipticals and the tidal - feature ellipticals .
a k - s test of the color gradient distribution confirms that the relaxed types and the tidal feature types share a nearly identical distribution of color gradients in all the three colors ( table [ tab_ks ] ) .
this is similar to the result of @xcite who find no strong distinctions between the fundamental planes of relaxed and tidal feature ellipticals . on the other hand ,
the median color gradient of the dust - feature type ellipticals is markedly steeper than the other two types by twice to three times ( table [ tab_col_grad ] ) .
a k - s test gives less than 5 % probability that the color gradient distributions of the dust - feature ellipticals and the relaxed ellipticals are identical ( table [ tab_ks ] ) .
note that we masked out the dust - features when fitting the sb profiles of the dust - feature ellipticals , in order to minimize the dust extinction affecting color gradient measurements .
therefore , we suggest that a difference in the formation and the evolution process is the origin for the discrepancy , rather than the dust extinction .
the fact that the relaxed type ellipticals and the tidal - feature type ellipticals share a similar color gradient distribution means that they share a similar merging history .
the steep color gradient in the dust - feature types is in agreement with the expected steep color gradient from a monolithic collapse or a dissipative major merger with a high star formation efficiency .
therefore , we suspect that the dust - feature type ellipticals are the products of merging between gas - rich galaxies where the remaining gas producing the dust features . in figure [ kmag_col ]
, we present the @xmath1-band absolute magnitude ( @xmath65 ) versus the color gradients in three different colors .
since @xmath65 is closely related to the stellar mass ( @xmath66 ) , we also mark the stellar mass on the x - axis , where the stellar mass is computed assuming a mass - to - light ratio of 1 . for a @xmath1-band mass - to - light ratio of 1 ,
the absolute magnitudes of @xmath67 and @xmath68 mags correspond to the stellar mass of @xmath69 and @xmath70 respectively @xcite .
different types of elliptical galaxies are plotted with different symbols , and the median , the error of the median , and the rms dispersion of the color gradients are indicated at each absolute magnitude bin , allowing us to understand how the color gradient changes as a function of the absolute magnitudes ( or the stellar masses ) . at the bottom of each figure ,
we indicate median errors in the color gradient measurements .
if dry merging mixes up stellar population in each galaxy , a naive expectation is that the color gradients become shallower and the dispersion becomes smaller toward the brighter magnitude .
however , figure [ kmag_col ] shows that the median values of the color gradients are nearly constant as a function of absolute magnitude , somewhat in contrary to the naive expectation ( possibly with an exception at the brightest end ) . on the other hand
, the dispersion of the color gradients appears to decrease at the brighter end , consistent with the expectation .
a similar trend is evident in figure [ col_prof_comb ] too .
however , the steady decrease in the scatter is illusionary and reflects the larger measurement errors as discussed below . to examine this trend more closely
, we plot the intrinsic scatter of the color gradients as a function of @xmath1-band absolute magnitude in figure [ scat_col ] .
note that the measurement errors add noises to the intrinsic scatter , therefore we subtracted the contribution of the measurement errors from the observed scatter to obtain the intrinsic scatter as described below .
note that the subtraction of the measurement errors is more important at the faint end than at the bright end since faint objects have large color gradient errors due to the lack of data points to fit for their sb .
we assumed that the color gradients are distributed in gaussian with its mean value equal to the measured median value of the color gradient with a certain intrinsic dispersion .
then , we constructed the measurement error distribution for actual measurement errors .
we took a median value of the measurement errors ( at each absolute magnitude bin ) , and subtracted it from the observed rms scatter to obtain an initial guess for the intrinsic scatter . in the next step , we randomly picked color gradients from the normal distribution of the color gradients , assigning measurement errors to each value by randomly picking a measurement error from the measurement error distribution that we constructed earlier .
the median value and the rms scatter are computed from this simulated set of color gradients .
the rms scatter was compared with the observed rms scatter , and the value of the intrinsic scatter was adjusted if the rms scatter from the simulation does not match the observed rms scatter . finally , after iterations of the above process , we determined the intrinsic scatter to be the intrinsic scatter in the simulation where the rms scatter of the simulation is identical to the observed rms scatter .
error of the intrinsic scatter of color gradients , @xmath71 , was estimated by propagating the color gradient measurement errors in the following way . here
, @xmath72 is the intrinsic scatter , @xmath73 is the number of galaxies in a certain magnitude bin , @xmath74 is the color gradient value of the ith galaxy , @xmath75 is the measurement error of the color gradient of the ith galaxy , and @xmath76 is the average of the color gradients .
the values of the scatters are presented in table [ tab_scat ] and shown in figure [ scat_col ] .
although the estimates of the error values are mathematically sound , we caution that the the error estimates do not include potential systematic effects discussed in section 3.3 , and may be under- or over - estimated due to small number statistics in some of the bins .
@xmath77 @xmath78 @xmath79 in figure [ scat_col ] , we find that the steady decrease of the rms scatter toward the massive end disappears for the optical - nir color gradients when the intrinsic scatter is considered .
the decrease in the intrinsic scatter is only visible at the most massive end at a few @xmath42 level .
the implication of the result will be discussed later , in relation to the merging activity ( 5 ) .
our analysis ( 3.3 ) allows us to derive age and metallicity gradients .
the results of the age and the metallicity gradient are described in this section .
figure [ hist_tz ] shows the marginal distribution of the age and the metallicity gradients . as was the case for the color gradients ,
the relaxed type ellipticals and the tidal - feature type ellipticals share a nearly identical distribution , with a positive age gradient ( @xmath80 ) and a negative metallicity gradient ( @xmath81 ) .
the dust - feature type ellipticals have small age gradients ( except for two outliers ) , and significantly steep metallicity gradients of @xmath82 .
the very steep metallicity gradient of dust - feature type ellipticals supports the idea that they are the products of wet merging events .
the derived mean @xmath61 is steeper than @xmath61 values derived from studies that are based on the line indices ( @xmath83 @xmath5 ; @xcite ; @xcite ; @xcite ; @xcite ; citealtani07 ) , but is in agreement with the values in @xcite and @xcite that are based on color profile analysis like our study .
the origin of the discrepancy is unclear although a slightly positive age gradient in the color gradient analysis steepens the metallicity gradient but not to the extent that it resolves the difference .
the line indices analysis that derive the flatter @xmath61 assume @xmath84 or find null @xmath85 .
while the line indices studies are limited at @xmath86 , the color gradient studies are done out to a few @xmath87 , so this may be the cause of the discrepancy .
the absolute values of the gradients are quantities that are subject to the analysis methods ( e.g. , the adopted stellar population synthesis models , see li et al .
2007 ) so that we caution readers not to pay too much attention to the absolute values but concentrating on the relative changes in the gradients . in figure [ kmag_az ]
, we show the @xmath65 versus the age and the metallicity gradients . with the thick points ,
we indicate the median @xmath85 and @xmath61 values and their errors . on these points ,
the thin error bars show the rms dispersion .
the thick error bars represent the errors of the median , which are derived by dividing the rms dispersion by @xmath88 where @xmath73 is the number of objects in the corresponding bin . like the color gradients , the age and the metallicity gradients
do not show a significant change as a function of @xmath65 .
one exception is the brightest bin ( corresponding to the stellar mass of @xmath89 ) where both @xmath85 and @xmath61 start to flatten .
however this trend at the brightest bin should be taken with a caution since only six objects are included in the bin .
just like the color gradients , this result is consistent with a similar analysis performed by @xcite . on the other hand , @xcite do not show a result at the very massive end of @xmath89 , so the flattening of the age and the metallicity gradients at the brightest bin is not noticed in their work .
the flattening of the age and the metallicity gradients at the brightest end is interesting since this is exactly what is expected if these ellipticals underwent frequent dry merging .
obviously , more number statistics at the bright end is desired to make a stronger case for this .
similarly to the color gradients , the scatters in the age and the color gradients appear to decrease at the brightest magnitude bin . as a closer examination of this trend
, we show in figure [ scat_az ] and table [ tab_scat ] the intrinsic scatter of the age and the metallicity gradients as a function of the absolute magnitude . the intrinsic scatter in these gradients
are obtained in an iterative process identical to the method we employed for the intrinsic scatter of color gradients ( see the previous subsection ) .
errors of the intrinsic scatters are estimated using eqs .
( [ eqn_scat_err1 ] ) ( [ eqn_scat_err3 ] ) , but here replacing the color gradients with @xmath85 or @xmath61 .
the derived intrinsic rms scatter is about 0.3 at a lower mass range of @xmath90 , which is roughly in agreement with the scatter found in @xcite at a similar mass range . like the scatters of color gradients , the scatters of these quantities stay nearly constant out to @xmath91 or @xmath92 , but decreases at the most massive end by nearly a factor of two per a factor of two change in mass ( or luminosity ) .
the reduction in the scatter is dominated by the relaxed and the tidal - feature type ellipticals .
on the other hand , the reduction in the scatter is inversed for the tidal - feature type ellipticals , although the larger errors in the scatter ( due to the smaller number of galaxies available ) for these types of objects make it difficult to draw a firm conclusion .
overall , the two evidences the possible flattening of the age and the metallicity gradients and the decrease in the scatter of these gradients at the most massive end are in agreement with the expectation of what would happen under dry merging , and this point will be discussed in more detail below .
in the previous section , we saw that the scatters in the color / age / metallicity gradients decrease at the brightest @xmath65 . here , we present a simple model that calculates how the metallicity gradient would change as a function of absolute magnitude and mass if major dry merging is the dominant mechanism for the growth of elliptical galaxies at the massive end .
to do so , we use a simulation result from @xcite where they perform simulations of major dry merging of elliptical galaxies with various initial conditions , and present the metallicity gradients of pre - mergers and merger products .
borrowing their results , we establish a relation between the average metallicity gradients of two galaxies before merging , and the metallicity gradient of the merger product ( figure [ dmerg ] ) .
we find that there is a good correlation between the average of the metallicity gradients of pre - merging galaxies ( @xmath93/2 $ ] ) and the metallicity gradient of the merger product ( @xmath94 ) as @xmath95 the relation shows that the metallicity gradient flattens on average when two galaxies merge , with the resulting @xmath61 being about 60 % of the average value of the @xmath61 s of the two pre - merging galaxies .
the correlation is reasonably tight , with the largest rms dispersion of 0.03 in @xmath94 .
this dispersion is sufficiently smaller than the reduction in the metallicity gradient value ( which is order of 0.1 ) , therefore we ignore the dispersion in eq .
( [ eq_dmerg ] ) in the following discussion .
two consequences are expected from eq .
( [ eq_dmerg ] ) .
one is that the metallicity gradient flattens gradually as the masses of ellipticals increase in a dry merging dominated evolution . roughly speaking , there should be a decrease in the metallicity gradient by a factor of 0.55 , over a mass doubling scale ( 0.3 dex in log(m ) ) .
another consequence is the decrease of the scatter in the color and the stellar population gradients at a high mass end .
suppose that we have a variety of the stellar population gradients for pre - merging ellipticals .
if we randomly merge galaxies with different metallicity gradients , the merging averages the metallicity gradient values and remove galaxies with extreme metallicity gradients . since the merger product will be more massive than each of its predecessors , we expect that the major dry merging would reduce the scatter in the color gradients at high mass end . to see how the mean @xmath61 and its intrinsic scatter would evolve under dry merging , we performed a simple simulation . here
, we assumed an initial distribution of @xmath61 in gaussian with its mean value at @xmath96 and the intrinsic scatter at @xmath97 .
these values correspond roughly to the observed mean @xmath61 and its scatter at @xmath98 .
we assume that a certain fraction , @xmath99 , of galaxies in a mass bin underwent dry merging .
this free parameter @xmath99 will control the amount of change in the mean metallicity gradient and its scatter .
figure [ dm_frac ] shows the result of this simulation .
the metallicity gradient and its scatter change the most when @xmath99 is the largest . at @xmath100 , the mean metallicity gradient and its scatter changes by a factor of 0.6 and 0.4 respectively , per each mass doubling .
we find that this trend is independent of the initial mean @xmath61 and @xmath101 values .
based on the simulation , we derive the following relations at @xmath100 : @xmath102 and @xmath103 here , @xmath104 , @xmath105 , and @xmath106 are the logarithmic changes in the mean metallicity gradient , the intrinsic scatter of the metallicity gradient , and the stellar mass , respectively .
the above relations show that the change in @xmath101 is more prominent than the change in @xmath107 as a result of dry merging by a factor of two . with the theoretical expectations from major dry merging at hand , we now proceed to interpret our observational results about @xmath61 .
figure [ scat_sim ] shows the comparison of the observed change in @xmath61 and @xmath101 with a model prediction . in the figure ,
the error in @xmath101 is estimated in the same way as the errors in the scatter of the color gradient values as described in section 4.2 and using eq .
the model plot assumes no major dry merging at @xmath108 , @xmath109 at @xmath110 , and @xmath111 at @xmath112 .
as we have shown earlier , the reduction in @xmath101 is observed at the massive end ( @xmath113 ) .
the major dry merging can provide an explanation for this result .
the observed reduction in the scatter is about 0.1 over the mass increase of 0.4 dex , and according to figure [ dm_frac ] , this can be achieved by assuming that 70 100 % of ellipticals at the higher mass bin have experienced major dry merging while the rest went through the formation mechanisms similar to those at the lower mass bins ( i.e. , they possess similar @xmath61 and @xmath85 distributions to those at the lower mass bin ) .
the possible formation and the evolution mechanisms at the lower mass bin include the wet merging , the minor merging , and the monolithic collapse ( see discussions below ) . according to our model , this amount of major dry merging needs to flatten @xmath61 by 0.15 over a mass doubling scale .
this value is rather small compared to the change in the scatter , but should be detectable with a sufficient number of objects . at the moment
, we find such a trend , but with a rather large error bar . in comparison to other works ,
our results are in agreement with @xcite where they find a reduced scatter in color gradients of bcgs . also note that some other studies report a flattening of the metallicity gradient at the massive end ( e.g. , * ? ? ?
* ) . while we find a supporting evidence for major dry merging at the high mass end , major dry merging is probably not a main mechanism driving the evolution of lower mass ellipticals with @xmath114 . between @xmath115 and @xmath110 ,
the @xmath101 values stay nearly constant .
the same trend is true for the color and the age gradients .
clearly , the trend is not consistent with the expectation from the major dry merging model we presented in this paper . in order to explain the null change in the metallicity gradients
, @xmath99 should be less than @xmath116 .
the effect of major dry merging seems negligible in the evolution of ellipticals at this mass range .
the origin of the large scatter in @xmath61 at this mass range is unclear , although there are plenty of possibilities responsible for the dispersion .
minor dry merging could steepen the metallicity gradient , rather then flattening it , by distributing the merged stellar remnants in the outer region not the inner region @xcite .
dissipative collapse could create a variety of gradients depending on the star formation efficiencies how the gravitational potential is shaped for each object @xcite .
minor merging with a gas - rich galaxy could place younger generation stars at central or outer regions , making galaxies blue at certain areas and complicating the analysis of the result ( e.g. , * ? ? ?
we examined the color , the age , and the metallicity gradients of of 204 nearby ( @xmath21 ) , massive ( @xmath117 mag , or @xmath118 ) elliptical galaxies ( early - type galaxies with @xmath119 ) , divided into three morphological types the relaxed , the tidal - feature , and the dust - feature types .
the subdivision of elliptical galaxies was possible by using the sdss stripe 82 images which goes about 2 mag deeper than regular sdss images , and the derivation of the age and the metallicity gradients was done by adding @xmath1-band data of the ukidss las to the sdss data which enabled us to break the age - metallicity degeneracy .
no significant difference is found between relaxed and tidal - feature types in the gradient properties .
however , the dust - feature type ellipticals have steeper color gradients which seem to originate from wet merging rather than the dust extinction effect .
the gradient values of elliptical galaxies are found to be nearly constant over the explored mass range , but we find that the intrinsic scatter of @xmath61 decreases at @xmath2 by about 0.1 .
to interpret our results , we constructed a simple model that describes changes in @xmath61 and @xmath101 as a function of mass under major dry merging .
the model predicts that the mean @xmath61 flattens , and @xmath101 decreases as mass of an elliptical galaxy doubles as a result of major dry merging . from the analysis of the change in @xmath61 and its scatter , we draw the following conclusions , with a caveat that the result is based on a rather small number of objects at the massive end ( @xmath120 ) .
first , the observed reduction in the intrinsic scatter in the @xmath121 versus @xmath61 relation supports the idea that major dry merging has acted on a significant number of ellipticals at the very massive end ( @xmath122 ) .
second , if the reduction of the scatter in @xmath61 is indeed caused by major dry mergers , @xmath61 should flatten at the higher mass .
the result from our analysis of the sdss / ukidss - las ellipticals is consistent with this scenario at the massive end , but not conclusive .
third , at the lower mass regime ( @xmath114 ) , the major dry merging does not appear to be the dominant process in the evolution of ellipticals .
at such a mass range , the mean @xmath61 is nearly constant , and the scatter in the @xmath123 versus @xmath61 relation is large and nearly constant ( @xmath124 ) too , implying that various physical mechanisms are responsible for the large scatter .
this work was supported by the creative initiative program , no .
2010 - 0000712 , of the national research foundation of korea ( nrfk ) funded by the korean government ( mest ) .
we thank our ceou / snu colleagues for useful discussion .
this paper uses the data taken with the united kingdom infrared telescope ( ukirt ) which is operated by the joint astronomy centre on behalf of the science and technology facilities and from the sloan digital sky survey project .
abazajian , k. n. , adelman - mccarthy , j. k. , ageros , m. a. , et al .
2009 , , 182 , 543 annibali , f. , bressan , a. , rampazzo , r. , zeilinger , w. w. , & danese , l. 2007 , , 463 , 455 barnes , j. e. , & hernquist , l. 1992 , , 30 , 705 baum , w. a. 1959 , , 71 , 106 bell , e. f. , & de jong , r. s. 2000 , , 312 , 497 bell , e. f. , & de jong , r. s. 2001 , , 550 , 212 bernardi , m. , roche , n. , shankar , f. , & sheth , r. k. 2011 , , 412 , l6 bertin , e. , & arnouts , s. 1996 , , 117 , 393 bruzual , g. , & charlot , s. 2003 , , 344 , 1000 carlberg , r. g. 1984 , , 286 , 416 chabrier , g. 2003 , , 115 , 763 daddi , e. , renzini , a. , pirzkal , n. , et al .
2005 , , 626 , 680 di matteo , p. , pipino , a. , lehnert , m. d. , combes , f. , & semelin , b. 2009 , , 499 , 427 djorgovski , s. , & davis , m. 1987 , , 313 , 59 drory , n. , bender , r. , feulner , g. , et al .
2004 , , 608 , 742 eggen , o. j. , lynden - bell , d. , & sandage , a. r. 1962 , , 136 , 748 franx , m. , & illingworth , g. 1990 , , 359 , l41 gil de paz , a. , & madore , b. f. 2002 , , 123 , 1864 gunn , j. e. , carr , m. , rockosi , c. , et al .
1998 , , 116 , 3040 hansson , k. s. a. , lisker , t. , & grebel , e. k. 2012 , , 427 , 2376 hinkley , s. , & i m , m. 2001 , , 560 , l41 hopkins , p. f. , cox , t. j. , dutta , s. n. , et al .
2009 , , 181 , 135 hwang , h. s. , geller , m. j. , diaferio , a. , & rines , k. j. 2012 , , 752 , 64 i m , m. , griffiths , r. e. , & ratnatunga , k. u. 1997 , , 475 , 457 i m , m. , griffiths , r. e. , naim , a. , et al . 1999 , , 510 , 82 i m , m. , faber , s. m. , gebhardt , k. , et al .
2001 , , 122 , 750 i m , m. , simard , l. , faber , s. m. , et al . 2002 , , 571 ,
136 jun , h. d. , & i m , m. 2008 , , 678 , l97 kaviraj , s. 2010 , , 406 , 382 kaviraj , s. , ting , y .- s . , bureau , m. , et al .
2012 , , 423 , 49 kawata , d. , & gibson , b. k. 2003 , , 340 , 908 kim , t. , sheth , k. , hinz , j. l. , et al .
2012 , , 753 , 43 ko , j. , & i m , m. 2005 , journal of korean astronomical society , 38 , 149 ko , j. , i m , m. , lee , h. m. , et al . 2009 , , 695 , l198 ko , j. , i m , m. , lee , h. m. , et al .
2012 , , 745 , 181 kobayashi , c. 2004 , , 347 , 740 kuntschner , h. , emsellem , e. , bacon , r. , et al .
2010 , , 408 , 97 la barbera , f. , de carvalho , r. r. , de la rosa , i. g. , et al .
2010 , , 140 , 1528 larson , r. b. 1974 , , 166 , 585 lawrence , a. , warren , s. j. , almaini , o. , et al .
2007 , , 379 , 1599 lee , j. h. , lee , m. g. , & hwang , h. s. 2006 , , 650 , 148 li , z. , han , z. , & zhang , f. 2007 , , 464 , 853 mancone , c. l. , & gonzalez , a. h. 2012 , , 124 , 606 markwardt , c. b. 2009 , astronomical data analysis software and systems xviii , 411 , 251 naab , t. , & trujillo , i. 2006 , , 369 , 625 ogando , r. l. c. , maia , m. a. g. , chiappini , c. , et al .
2005 , , 632 , l61 pahre , m. a. , djorgovski , s. g. , & de carvalho , r. r. 1995 , , 453 , l17 peletier , r. f. , davies , r. l. , illingworth , g. d. , davis , l. e. , & cawson , m. 1990 , , 100 , 1091 peng , c. y. , ho , l. c. , impey , c. d. , & rix , h .- w .
2010 , , 139 , 2097 pipino , a. , dercole , a. , chiappini , c. , & matteucci , f. 2010 , , 407 , 1347 rawle , t. d. , smith , r. j. , & lucey , j. r. 2010 , , 401 , 852 renzini , a. 2006 , , 44 , 141 roche , n. , bernardi , m. , & hyde , j. 2010 , , 407 , 1231 saglia , r. p. , maraston , c. , greggio , l. , bender , r. , & ziegler , b. 2000 , , 360 , 911 snchez - blzquez , p. , forbes , d. a. , strader , j. , brodie , j. , & proctor , r. 2007 , , 377 , 759 scarlata , c. , carollo , c. m. , lilly , s. j. , et al .
2007 , , 172 , 494 sheen , y .- k .
, yi , s. k. , ree , c. h. , & lee , j. 2012 , , 202 , 8 shim , h. , i m , m. , lee , h. m. , et al .
2011 , , 727 , 14 simard , l. , koo , d. c. , faber , s. m. , et al . 1999 , , 519 , 563 simard , l. , mendel , j. t. , patton , d. r. , ellison , s. l. , & mcconnachie , a. w. 2011 , , 196 , 11 simien , f. , & de vaucouleurs , g. 1986 , , 302 , 564 spolaor , m. , kobayashi , c. , forbes , d. a. , couch , w. j. , & hau , g. k. t. 2010 , , 408 , 272 tal , t. , van dokkum , p. g. , nelan , j. , & bezanson , r. 2009 , , 138 , 1417 tamura , n. , & ohta , k. 2003 , , 126 , 596 toomre , a. 1977 , evolution of galaxies and stellar populations , 401 tortora , c. , & napolitano , n. r. 2012 , , 421 , 2478 trujillo , i. , frster schreiber , n. m. , rudnick , g. , et al .
2006 , , 650 , 18 van dokkum , p. g. 2005 , , 130 , 2647 van dokkum , p. g. , franx , m. , kriek , m. , et al .
2008 , , 677 , l5 van dokkum , p. g. , & conroy , c. 2010 , , 468 , 940 van dokkum , p. g. , & conroy , c. 2011 , , 735 , l13 visvanathan , n. , & sandage , a. 1977 , , 216 , 214 white , s. d. m. 1980 , , 191 , 1p worthey , g. 1994 , , 95 , 107 yi , s. k. , yoon , s .- j . , kaviraj , s. , et al .
2005 , , 619 , l111 -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] + -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] + -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] + -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] -band ( left ) , in @xmath1-band ( middle ) and their surface brightness profiles in @xmath32 , @xmath33 and @xmath1-bands ( right ) .
the scale bar and @xmath27 in the images represent 10 arcsec and bulge to total light ratios . the vertical solid lines and
the dashed lines in the plot are the inner and the outer radii for deriving @xmath53 and @xmath54 color gradients .
the vertical dotted line means the outer boundary for the @xmath52 color gradient fit .
the ellipses in the @xmath1-band images indicate the outer boundary for the analysis of color profiles for the @xmath1-band .
the i d on top of the sb profile figures indicate the i d of the objects in @xcite as indicated in table [ tab_sp].,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] + but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] + but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] + but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for tidal - feature ellipticals .
note the faint tidal features around each elliptical , which are indicative of their past merging activities.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals . note the dust - lanes in the @xmath33-band images.,title="fig : " ] + but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] + but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] + but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for dust - feature ellipticals .
note the dust - lanes in the @xmath33-band images.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of the presence of close neighbors ( the top 2 rows ) and small sizes ( the bottom 2 rows).,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] + but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] but for the k10 early - types that are excluded from our analysis because of their low ( @xmath125 ) @xmath27 ratios.,title="fig : " ] and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] + and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] + and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] and @xmath53 diagram .
the age and the metallicity grids are created with bc03 ssp model in the manuscript .
the arrow head corresponds to the outermost radius .
right : the corresponding age and metallicity profiles .
also indicated are the gradient values.,title="fig : " ] + but for bc03 composite stellar population model of exponential decreasing star formation rate ( @xmath126 1 gyr)(upper left ) , a single burst expanding 1 gyr ( upper right ) and same ssp but with chabrier imf ( lower right).,title="fig : " ] but for bc03 composite stellar population model of exponential decreasing star formation rate ( @xmath126 1 gyr)(upper left ) , a single burst expanding 1 gyr ( upper right ) and same ssp but with chabrier imf ( lower right).,title="fig : " ] + but for bc03 composite stellar population model of exponential decreasing star formation rate ( @xmath126 1 gyr)(upper left ) , a single burst expanding 1 gyr ( upper right ) and same ssp but with chabrier imf ( lower right).,title="fig : " ] but for the age and the metallicity gradients . like the intrinsic scatters of the color gradients , the relaxed and the tidal - feature types show a reduced scatter at higher mass which can be interpreted as a consequence due major dry merging.,title="fig : " ] + but for the age and the metallicity gradients . like the intrinsic scatters of the color gradients ,
the relaxed and the tidal - feature types show a reduced scatter at higher mass which can be interpreted as a consequence due major dry merging.,title="fig : " ] and its scatter @xmath101 changes due to major dry merging , as a function of galaxy mass .
the fraction , @xmath99 indicates the faction of objects that underwent major dry merging .
the solid line indicates the change in @xmath101 , while the dashed line show how @xmath107 evolves as a result of the dry merging .
major dry merging reduces the scatter and the mean metallicity gradient value . ]
ccccccc relaxed & @xmath129 & @xmath130 & @xmath131 & @xmath132 & @xmath133 + tidal - feature & @xmath134 & @xmath135 & @xmath136 & @xmath137 & @xmath138 + dust - feature & @xmath139 & @xmath140 & @xmath141 & @xmath142 & @xmath143 + + total & @xmath144 & @xmath145 & @xmath146 & @xmath147 & @xmath148 + [ tab_col_grad ] @xmath159 & all & @xmath160 & @xmath161 & @xmath162 + & relaxed & @xmath163 & @xmath164 & @xmath165 + & tidal - feature & @xmath166 & @xmath167 & @xmath168 + & relaxed & tidal - feature & @xmath163 & @xmath169 & @xmath170 + & dust - feature & @xmath171 & @xmath172 & + @xmath173 & all & @xmath174 & @xmath175 & @xmath176 + & relaxed & @xmath177 & @xmath178 & @xmath179 + & tidal - feature & @xmath180 & @xmath181 & @xmath182 + & relaxed & tidal - feature & @xmath183 & @xmath184 & @xmath185 + & dust - feature & @xmath186 & @xmath187 & + @xmath188 & all & @xmath189 & @xmath190 & @xmath191 + & relaxed & @xmath192 & @xmath193 & @xmath194 + & tidal - feature & @xmath195 & @xmath196 & @xmath197 + & relaxed & tidal - feature & @xmath198 & @xmath199 & @xmath200 + & dust - feature & @xmath201 & @xmath202 & + @xmath203 & all & @xmath204 & @xmath205 & @xmath206 + & relaxed & @xmath207 & @xmath208 & @xmath209 + & tidal - feature & @xmath210 & @xmath211 & @xmath212 + & relaxed & tidal - feature & @xmath213 & @xmath214 & @xmath215 + & dust - feature & @xmath216 & @xmath217 & + @xmath101 & all & @xmath218 & @xmath219 & @xmath220 + & relaxed & @xmath221 & @xmath222 & @xmath223 + & tidal - feature & @xmath224 & @xmath225 & @xmath226 + & relaxed & tidal - feature & @xmath227 & @xmath228 & @xmath229 + & dust - feature & @xmath230 & @xmath231 & + [ tab_scat ] ccccccccccccc 24 & 0.0350 & 5.28 & 3.21 & @xmath232 & @xmath233 & @xmath234 & @xmath235 & @xmath236 & @xmath237 & @xmath238 & @xmath239 & r + 76 & 0.0431 & 2.35 & 3.21 & @xmath232 & @xmath240 & @xmath241 & @xmath242 & @xmath243 & @xmath244 & @xmath245 & @xmath246 & r + 89 & 0.0275 & 7.92 & 6.42 & @xmath247 & @xmath248 & @xmath249 & @xmath250 & @xmath251 & @xmath252 & @xmath253 & @xmath254 & t + 112 & 0.0313 & 11.88 & 8.03 & @xmath255 & @xmath256 & @xmath257 & @xmath258 & @xmath259 & @xmath260 & @xmath261 & @xmath262 & r + 135 & 0.0302 & 7.92 & 4.82 & @xmath263 & @xmath264 & @xmath265 & @xmath266 & @xmath267 & @xmath268 & @xmath269 & @xmath270 & t + 192 & 0.0267 & 5.28 & 4.82 & @xmath271 & @xmath272 & @xmath273 & @xmath274 & @xmath275 & @xmath276 & @xmath277 & @xmath270 & r + 221 & 0.0306 & 5.28 & 3.21 & @xmath278 & @xmath279 & @xmath280 & @xmath281 & @xmath282 & @xmath283 & @xmath284 & @xmath285 & r + 261 & 0.0422 & 5.28 & 3.21 & @xmath286 & @xmath287 & @xmath288 & @xmath289 & @xmath290 & @xmath291 & @xmath292 & @xmath293 & t + 269 & 0.0444 & 5.28 & 3.21 & @xmath294 & @xmath295 & @xmath296 & @xmath297 & @xmath298 & @xmath299 & @xmath300 & @xmath301 & t + 272 & 0.0443 & 3.52 & 3.21 & @xmath302 & @xmath303 & @xmath304 & @xmath305 & @xmath306 & @xmath307 & @xmath308 & @xmath309 & r + 280 & 0.0448 & 11.88 & 6.42 & @xmath271 & @xmath310 & @xmath311 & @xmath312 & @xmath313 & @xmath314 & @xmath315 & @xmath270 & t + 284 & 0.0500 & 3.52 & 3.21 & @xmath316 & @xmath317 & @xmath318 & @xmath319 & @xmath320 & @xmath321 & @xmath322 & @xmath239 & r + 326 & 0.0395 & 5.28 & 3.21 & @xmath323 & @xmath324 & @xmath325 & @xmath326 & @xmath327 & @xmath328 & @xmath329 & @xmath330 & r + 340 & 0.0483 & 5.28 & 3.21 & @xmath331 & @xmath332 & @xmath333 & @xmath334 & @xmath335 & @xmath336 & @xmath337 & @xmath338 & r + 346 & 0.0497 & 2.81 & 3.21 & @xmath339 & @xmath340 & @xmath341 & @xmath342 & @xmath343 & @xmath344 & @xmath345 & @xmath346 & r + 354 & 0.0380 & 1.68 & 1.61 & @xmath347 & @xmath348 & @xmath349 & @xmath350 & @xmath351 & @xmath268 & @xmath269 & @xmath330 & t + 359 & 0.0334 & 7.92 & 4.82 & @xmath263 & @xmath352 & @xmath353 & @xmath354 & @xmath355 & @xmath252 & @xmath356 & @xmath270 & r + 415 & 0.0453 & 17.82 & 16.05 & @xmath357 & @xmath358 & @xmath359 & @xmath360 & @xmath361 & @xmath362 & @xmath363 & @xmath364 & r + 448 & 0.0274 & 7.92 & 4.82 & @xmath365 & @xmath366 & @xmath367 & @xmath368 & @xmath369 & @xmath370 & @xmath371 & @xmath372 & r + 477 & 0.0475 & 4.22 & 3.21 & @xmath373 & @xmath374 & @xmath375 & @xmath376 & @xmath377 & @xmath378 & @xmath379 & @xmath380 & r + 481 & 0.0177 & 17.82 & 14.45 & @xmath381 & @xmath382 & @xmath248 & @xmath383 & @xmath384 & @xmath385 & @xmath386 & @xmath293 & t + 483 & 0.0353 & 3.52 & 3.21 & @xmath387 & @xmath388 & @xmath389 & @xmath376 & @xmath390 & @xmath391 & @xmath392 & @xmath393 & r + 506 & 0.0476 & 7.92 & 4.82 & @xmath286 & @xmath394 & @xmath395 & @xmath396 & @xmath397 & @xmath398 & @xmath399 & @xmath393 & r + 518 & 0.0407 & 7.92 & 6.02 & @xmath381 & @xmath400 & @xmath401 & @xmath402 & @xmath403 & @xmath404 & @xmath405 & @xmath406 & d + 528 & 0.0412 & 11.88 & 4.41 & @xmath407 & @xmath408 & @xmath409 & @xmath410 & @xmath411 & @xmath412 & @xmath413 & @xmath414 & d + 536 & 0.0387 & 5.28 & 4.82 & @xmath415 & @xmath416 & @xmath417 & @xmath418 & @xmath419 & @xmath420 & @xmath421 & @xmath330 & d + 537 & 0.0369 & 7.92 & 4.82 & @xmath271 & @xmath422 & @xmath423 & @xmath424 & @xmath425 & @xmath426 & @xmath413 & @xmath393 & r + 556 & 0.0326 & 5.28 & 3.21 & @xmath427 & @xmath295 & @xmath295 & @xmath428 & @xmath429 & @xmath430 & @xmath431 & @xmath432 & r + 560 & 0.0235 & 5.28 & 4.82 & @xmath433 & @xmath434 & @xmath435 & @xmath436 & @xmath437 & @xmath438 & @xmath439 & @xmath440 & r + 561 & 0.0494 & 3.52 & 3.21 & @xmath441 & @xmath442 & @xmath443 & @xmath444 & @xmath445 & @xmath446 & @xmath447 & @xmath448 & t + 597 & 0.0460 & 3.52 & 3.21 & @xmath449 & @xmath450 & @xmath451 & @xmath452 & @xmath453 & @xmath454 & @xmath455 & @xmath456 & r + 601 & 0.0474 & 5.28 & 3.21 & @xmath457 & @xmath458 & @xmath459 & @xmath460 & @xmath461 & @xmath462 & @xmath463 & @xmath464 & r + 606 & 0.0468 & 5.28 & 4.82 & @xmath316 & @xmath465 & @xmath466 & @xmath467 & @xmath468 & @xmath469 & @xmath470 & @xmath262 & r + 624 & 0.0426 & 3.52 & 3.21 & @xmath471 & @xmath472 & @xmath473 & @xmath474 & @xmath475 & @xmath476 & @xmath308 & @xmath477 & r + 639 & 0.0337 & 5.28 & 3.21 & @xmath478 & @xmath479 & @xmath480 & @xmath481 & @xmath482 & @xmath483 & @xmath484 & @xmath485 & r + 640 & 0.0418 & 5.50 & 8.03 & @xmath263 & @xmath435 & @xmath486 & @xmath487 & @xmath488 & @xmath489 & @xmath490 & @xmath477 & r + 675 & 0.0366 & 5.28 & 4.82 & @xmath316 & @xmath491 & @xmath492 & @xmath493 & @xmath494 & @xmath495 & @xmath496 & @xmath464 & r + 681 & 0.0427 & 7.92 & 4.82 & @xmath497 & @xmath498 & @xmath499 & @xmath500 & @xmath501 & @xmath502 & @xmath503 & @xmath504 & r + 723 & 0.0473 & 3.52 & 3.21 & @xmath505 & @xmath506 & @xmath506 & @xmath507 & @xmath508 & @xmath509 & @xmath510 & @xmath432 & t + 731 & 0.0456 & 2.99 & 3.21 & @xmath511 & @xmath512 & @xmath513 & @xmath514 & @xmath515 & @xmath516 & @xmath484 & @xmath406 & r + 739 & 0.0297 & 7.92 & 4.82 & @xmath517 & @xmath518 & @xmath519 & @xmath520 & @xmath521 & @xmath522 & @xmath455 & @xmath477 & r + 745 & 0.0475 & 5.28 & 4.82 & @xmath523 & @xmath524 & @xmath525 & @xmath526 & @xmath527 & @xmath321 & @xmath322 & @xmath301 & r + 776 & 0.0467 & 4.24 & 3.21 & @xmath232 & @xmath528 & @xmath529 & @xmath530 & @xmath531 & @xmath532 & @xmath329 & @xmath239 & r + 789 & 0.0405 & 5.28 & 4.82 & @xmath533 & @xmath534 & @xmath535 & @xmath536 & @xmath537 & @xmath538 & @xmath539 & @xmath540 & r + 822 & 0.0479 & 11.88 & 6.42 & @xmath541 & @xmath542 & @xmath543 & @xmath544 & @xmath545 & @xmath546 & @xmath547 & @xmath406 & r + 825 & 0.0236 & 17.82 & 6.02 & @xmath548 & @xmath549 & @xmath550 & @xmath551 & @xmath552 & @xmath553 & @xmath253 & @xmath554 & d + 829 & 0.0485 & 11.88 & 6.42 & @xmath555 & @xmath556 & @xmath557 & @xmath558 & @xmath559 & @xmath560 & @xmath561 & @xmath338 & t + 831 & 0.0490 & 6.50 & 4.82 & @xmath357 & @xmath562 & @xmath563 & @xmath564 & @xmath353 & @xmath565 & @xmath566 & @xmath567 & r + 844 & 0.0346 & 5.28 & 4.82 & @xmath568 & @xmath569 & @xmath570 & @xmath571 & @xmath572 & @xmath573 & @xmath574 & @xmath575 & r + 900 & 0.0440 & 5.28 & 3.21 & @xmath576 & @xmath577 & @xmath578 & @xmath579 & @xmath580 & @xmath581 & @xmath582 & @xmath583 & t + 947 & 0.0443 & 5.28 & 3.21 & @xmath286 & @xmath584 & @xmath491 & @xmath585 & @xmath586 & @xmath587 & @xmath588 & @xmath372 & r + 983 & 0.0127 & 11.88 & 14.45 & @xmath294 & @xmath358 & @xmath359 & @xmath589 & @xmath590 & @xmath591 & @xmath592 & @xmath593 & r + 1006 & 0.0500 & 2.88 & 3.21 & @xmath594 & @xmath595 & @xmath596 & @xmath597 & @xmath598 & @xmath307 & @xmath599 & @xmath600 & r + 1025 & 0.0480 & 2.35 & 1.61 & @xmath601 & @xmath602 & @xmath603 & @xmath376 & @xmath604 & @xmath605 & @xmath606 & @xmath607 & r + 1029 & 0.0498 & 3.58 & 3.21 & @xmath608 & @xmath609 & @xmath610 & @xmath611 & @xmath612 & @xmath613 & @xmath614 & @xmath615 & r + 1098 & 0.0472 & 3.01 & 3.21 & @xmath616 & @xmath617 & @xmath618 & @xmath242 & @xmath619 & @xmath620 & @xmath621 & @xmath622 & r + 1103 & 0.0392 & 7.92 & 3.21 & @xmath623 & @xmath624 & @xmath625 & @xmath626 & @xmath627 & @xmath628 & @xmath629 & @xmath593 & r + 1108 & 0.0468 & 4.18 & 3.21 & @xmath630 & @xmath631 & @xmath632 & @xmath633 & @xmath634 & @xmath635 & @xmath636 & @xmath637 & r + 1114 & 0.0440 & 5.28 & 4.82 & @xmath638 & @xmath639 & @xmath640 & @xmath641 & @xmath642 & @xmath643 & @xmath644 & @xmath246 & r + 1115 & 0.0446 & 2.92 & 3.21 & @xmath645 & @xmath646 & @xmath647 & @xmath648 & @xmath649 & @xmath650 & @xmath651 & @xmath554 & r + 1116 & 0.0443 & 2.35 & 3.21 & @xmath652 & @xmath653 & @xmath654 & @xmath655 & @xmath656 & @xmath657 & @xmath421 & @xmath540 & r + 1177 & 0.0441 & 3.52 & 4.82 & @xmath658 & @xmath659 & @xmath660 & @xmath661 & @xmath479 & @xmath662 & @xmath629 & @xmath380 & r + 1186 & 0.0421 & 5.28 & 4.82 & @xmath523 & @xmath663 & @xmath664 & @xmath665 & @xmath666 & @xmath667 & @xmath668 & @xmath600 & r + 1191 & 0.0431 & 3.52 & 3.21 & @xmath376 & @xmath669 & @xmath670 & @xmath671 & @xmath672 & @xmath673 & @xmath674 & @xmath239 & r + 1203 & 0.0435 & 4.18 & 3.21 & @xmath232 & @xmath675 & @xmath676 & @xmath633 & @xmath677 & @xmath678 & @xmath679 & @xmath262 & r + 1217 & 0.0232 & 11.88 & 6.02 & @xmath680 & @xmath681 & @xmath682 & @xmath683 & @xmath684 & @xmath685 & @xmath686 & @xmath687 & d + 1221 & 0.0432 & 3.52 & 4.82 & @xmath688 & @xmath537 & @xmath689 & @xmath690 & @xmath691 & @xmath692 & @xmath356 & @xmath338 & r + 1228 & 0.0214 & 11.88 & 8.03 & @xmath357 & @xmath693 & @xmath694 & @xmath695 & @xmath696 & @xmath697 & @xmath698 & @xmath270 & r + 1238 & 0.0233 & 7.92 & 4.82 & @xmath699 & @xmath700 & @xmath701 & @xmath702 & @xmath703 & @xmath605 & @xmath606 & @xmath338 & r + 1254 & 0.0495 & 5.28 & 3.21 & @xmath278 & @xmath704 & @xmath705 & @xmath706 & @xmath707 & @xmath708 & @xmath709 & @xmath372 & r + 1280 & 0.0275 & 5.28 & 3.21 & @xmath497 & @xmath710 & @xmath711 & @xmath712 & @xmath713 & @xmath714 & @xmath715 & @xmath716 & r + 1287 & 0.0446 & 5.28 & 3.21 & @xmath717 & @xmath718 & @xmath719 & @xmath720 & @xmath721 & @xmath722 & @xmath723 & @xmath637 & r + 1322 & 0.0437 & 11.88 & 9.63 & @xmath357 & @xmath724 & @xmath725 & @xmath726 & @xmath727 & @xmath728 & @xmath729 & @xmath456 & r + 1330 & 0.0419 & 3.52 & 2.81 & @xmath541 & @xmath730 & @xmath731 & @xmath732 & @xmath733 & @xmath268 & @xmath269 & @xmath734 & r + 1335 & 0.0454 & 7.92 & 4.82 & @xmath735 & @xmath435 & @xmath736 & @xmath737 & @xmath738 & @xmath538 & @xmath539 & @xmath380 & r + 1339 & 0.0443 & 5.28 & 4.82 & @xmath331 & @xmath295 & @xmath739 & @xmath740 & @xmath741 & @xmath495 & @xmath496 & @xmath742 & r + 1352 & 0.0361 & 5.28 & 4.82 & @xmath743 & @xmath744 & @xmath745 & @xmath746 & @xmath747 & @xmath748 & @xmath749 & @xmath262 & r + 1367 & 0.0359 & 5.28 & 3.21 & @xmath750 & @xmath751 & @xmath752 & @xmath753 & @xmath754 & @xmath755 & @xmath756 & @xmath254 & r + 1377 & 0.0421 & 4.24 & 3.21 & @xmath757 & @xmath758 & @xmath759 & @xmath676 & @xmath760 & @xmath761 & @xmath277 & @xmath330 & r + 1406 & 0.0493 & 5.28 & 6.42 & @xmath762 & @xmath763 & @xmath764 & @xmath765 & @xmath766 & @xmath767 & @xmath768 & @xmath464 & t + 1421 & 0.0248 & 5.28 & 4.82 & @xmath769 & @xmath551 & @xmath770 & @xmath771 & @xmath772 & @xmath773 & @xmath674 & @xmath477 & r + 1436 & 0.0364 & 7.92 & 6.42 & @xmath541 & @xmath774 & @xmath774 & @xmath775 & @xmath776 & @xmath777 & @xmath778 & @xmath338 & t + 1440 & 0.0489 & 5.28 & 3.21 & @xmath779 & @xmath780 & @xmath781 & @xmath782 & @xmath783 & @xmath784 & @xmath785 & @xmath364 & r + 1451 & 0.0466 & 3.52 & 3.21 & @xmath471 & @xmath786 & @xmath787 & @xmath788 & @xmath789 & @xmath790 & @xmath379 & @xmath380 & r + 1461 & 0.0419 & 5.28 & 4.82 & @xmath769 & @xmath436 & @xmath791 & @xmath792 & @xmath793 & @xmath794 & @xmath668 & @xmath456 & r + 1462 & 0.0451 & 7.92 & 4.82 & @xmath795 & @xmath796 & @xmath797 & @xmath798 & @xmath799 & @xmath800 & @xmath801 & @xmath802 & r + 1470 & 0.0472 & 3.94 & 3.21 & @xmath803 & @xmath739 & @xmath804 & @xmath805 & @xmath806 & @xmath807 & @xmath808 & @xmath338 & r + 1477 & 0.0443 & 5.28 & 4.82 & @xmath271 & @xmath809 & @xmath810 & @xmath811 & @xmath812 & @xmath813 & @xmath470 & @xmath254 & r + 1526 & 0.0370 & 5.18 & 4.82 & @xmath814 & @xmath815 & @xmath816 & @xmath817 & @xmath818 & @xmath819 & @xmath820 & @xmath802 & r + 1551 & 0.0439 & 3.17 & 3.21 & @xmath821 & @xmath822 & @xmath823 & @xmath824 & @xmath825 & @xmath826 & @xmath827 & @xmath254 & r + 1670 & 0.0431 & 5.28 & 4.82 & @xmath505 & @xmath828 & @xmath829 & @xmath830 & @xmath831 & @xmath832 & @xmath833 & @xmath802 & r + 1694 & 0.0370 & 3.38 & 3.21 & @xmath407 & @xmath834 & @xmath835 & @xmath836 & @xmath837 & @xmath838 & @xmath606 & @xmath839 & r + 1698 & 0.0393 & 5.28 & 4.82 & @xmath407 & @xmath840 & @xmath841 & @xmath842 & @xmath843 & @xmath844 & @xmath845 & @xmath687 & r + 1737 & 0.0462 & 5.28 & 4.82 & @xmath743 & @xmath846 & @xmath847 & @xmath848 & @xmath849 & @xmath850 & @xmath356 & @xmath477 & r + 1775 & 0.0411 & 3.52 & 3.21 & @xmath851 & @xmath852 & @xmath853 & @xmath854 & @xmath855 & @xmath856 & @xmath857 & @xmath239 & r + 1777 & 0.0407 & 11.88 & 6.42 & @xmath858 & @xmath859 & @xmath860 & @xmath861 & @xmath862 & @xmath502 & @xmath863 & @xmath637 & r + 1779 & 0.0407 & 5.28 & 4.82 & @xmath278 & @xmath864 & @xmath865 & @xmath866 & @xmath867 & @xmath868 & @xmath588 & @xmath372 & r + 1790 & 0.0424 & 4.23 & 3.21 & @xmath869 & @xmath870 & @xmath871 & @xmath633 & @xmath872 & @xmath873 & @xmath636 & @xmath440 & r + 1791 & 0.0414 & 3.52 & 3.21 & @xmath874 & @xmath875 & @xmath876 & @xmath877 & @xmath878 & @xmath879 & @xmath880 & @xmath802 & t + 1792 & 0.0419 & 5.28 & 4.82 & @xmath881 & @xmath882 & @xmath883 & @xmath884 & @xmath885 & @xmath886 & @xmath651 & @xmath262 & r + 1808 & 0.0418 & 7.92 & 6.42 & @xmath294 & @xmath887 & @xmath888 & @xmath889 & @xmath890 & @xmath891 & @xmath582 & @xmath393 & r + 1818 & 0.0450 & 5.28 & 2.81 & @xmath323 & @xmath306 & @xmath892 & @xmath893 & @xmath894 & @xmath807 & @xmath808 & @xmath615 & d + 1823 & 0.0387 & 11.88 & 6.42 & @xmath533 & @xmath895 & @xmath896 & @xmath897 & @xmath898 & @xmath899 & @xmath900 & @xmath372 & t + 1862 & 0.0224 & 5.28 & 3.21 & @xmath901 & @xmath395 & @xmath902 & @xmath903 & @xmath904 & @xmath438 & @xmath439 & @xmath607 & r + 1864 & 0.0473 & 2.74 & 3.21 & @xmath645 & @xmath905 & @xmath906 & @xmath907 & @xmath908 & @xmath909 & @xmath910 & @xmath504 & r + 1865 & 0.0499 & 5.28 & 3.21 & @xmath743 & @xmath911 & @xmath911 & @xmath912 & @xmath913 & @xmath914 & @xmath915 & @xmath687 & r + 1874 & 0.0419 & 4.20 & 3.21 & @xmath916 & @xmath917 & @xmath918 & @xmath919 & @xmath920 & @xmath283 & @xmath284 & @xmath540 & r + 1880 & 0.0465 & 7.92 & 3.21 & @xmath568 & @xmath921 & @xmath922 & @xmath923 & @xmath924 & @xmath925 & @xmath833 & @xmath239 & t + 1890 & 0.0254 & 5.28 & 4.82 & @xmath795 & @xmath926 & @xmath927 & @xmath928 & @xmath929 & @xmath930 & @xmath880 & @xmath456 & r + 1968 & 0.0376 & 7.92 & 6.42 & @xmath931 & @xmath932 & @xmath933 & @xmath934 & @xmath935 & @xmath936 & @xmath937 & @xmath938 & r + 1977 & 0.0381 & 3.52 & 3.21 & @xmath939 & @xmath940 & @xmath941 & @xmath942 & @xmath943 & @xmath944 & @xmath245 & @xmath432 & t + 1988 & 0.0359 & 3.52 & 3.21 & @xmath688 & @xmath945 & @xmath946 & @xmath947 & @xmath948 & @xmath949 & @xmath950 & @xmath734 & r + 2003 & 0.0220 & 17.82 & 17.66 & @xmath951 & @xmath895 & @xmath952 & @xmath953 & @xmath954 & @xmath260 & @xmath768 & @xmath456 & r + 2027 & 0.0430 & 11.88 & 8.03 & @xmath955 & @xmath956 & @xmath957 & @xmath958 & @xmath959 & @xmath960 & @xmath961 & @xmath301 & r + 2040 & 0.0418 & 5.28 & 3.21 & @xmath331 & @xmath962 & @xmath963 & @xmath964 & @xmath965 & @xmath553 & @xmath966 & @xmath414 & r + 2064 & 0.0407 & 2.95 & 3.21 & @xmath967 & @xmath968 & @xmath969 & @xmath970 & @xmath971 & @xmath972 & @xmath973 & @xmath974 & r + 2068 & 0.0418 & 11.88 & 8.03 & @xmath555 & @xmath975 & @xmath976 & @xmath977 & @xmath978 & @xmath502 & @xmath503 & @xmath338 & t + 2069 & 0.0432 & 5.28 & 4.82 & @xmath979 & @xmath980 & @xmath981 & @xmath982 & @xmath983 & @xmath879 & @xmath880 & @xmath270 & r + 2070 & 0.0419 & 3.52 & 3.21 & @xmath232 & @xmath984 & @xmath882 & @xmath985 & @xmath986 & @xmath987 & @xmath973 & @xmath485 & r + 2099 & 0.0249 & 13.86 & 6.42 & @xmath988 & @xmath989 & @xmath249 & @xmath990 & @xmath991 & @xmath972 & @xmath992 & @xmath440 & r + 2102 & 0.0254 & 5.28 & 4.82 & @xmath993 & @xmath333 & @xmath453 & @xmath994 & @xmath442 & @xmath995 & @xmath996 & @xmath330 & r + 2116 & 0.0416 & 11.88 & 6.02 & @xmath286 & @xmath997 & @xmath998 & @xmath999 & @xmath1000 & @xmath314 & @xmath1001 & @xmath607 & d + 2136 & 0.0371 & 7.92 & 6.42 & @xmath523 & @xmath1002 & @xmath1003 & @xmath1004 & @xmath1005 & @xmath930 & @xmath880 & @xmath406 & r + 2146 & 0.0378 & 26.73 & 9.63 & @xmath357 & @xmath1006 & @xmath1007 & @xmath1008 & @xmath862 & @xmath1009 & @xmath1010 & @xmath293 & r + 2165 & 0.0317 & 5.28 & 4.82 & @xmath1011 & @xmath1012 & @xmath1013 & @xmath1014 & @xmath1015 & @xmath1016 & @xmath463 & @xmath575 & r + 2176 & 0.0470 & 4.50 & 3.21 & @xmath232 & @xmath1017 & @xmath1018 & @xmath1019 & @xmath1020 & @xmath1021 & @xmath910 & @xmath239 & r + 2196 & 0.0391 & 11.88 & 8.03 & @xmath263 & @xmath324 & @xmath1022 & @xmath1023 & @xmath1024 & @xmath1025 & @xmath863 & @xmath393 & r + 2201 & 0.0496 & 5.28 & 4.82 & @xmath1026 & @xmath1027 & @xmath1028 & @xmath1029 & @xmath1030 & @xmath1031 & @xmath644 & @xmath406 & r + 2298 & 0.0336 & 5.28 & 3.21 & @xmath630 & @xmath1032 & @xmath660 & @xmath1033 & @xmath1034 & @xmath1035 & @xmath1036 & @xmath464 & r + 2331 & 0.0418 & 7.92 & 4.82 & @xmath1026 & @xmath1037 & @xmath1038 & @xmath1039 & @xmath1040 & @xmath1041 & @xmath1042 & @xmath637 & r + 2345 & 0.0458 & 4.14 & 3.21 & @xmath263 & @xmath1043 & @xmath1044 & @xmath1045 & @xmath1046 & @xmath1047 & @xmath651 & @xmath239 & r + 2346 & 0.0422 & 3.52 & 3.21 & @xmath1026 & @xmath1048 & @xmath453 & @xmath1049 & @xmath1050 & @xmath1051 & @xmath329 & @xmath938 & r + 2361 & 0.0423 & 2.85 & 3.21 & @xmath511 & @xmath1052 & @xmath1053 & @xmath1054 & @xmath1055 & @xmath283 & @xmath284 & @xmath615 & r + 2362 & 0.0283 & 5.28 & 3.21 & @xmath568 & @xmath1056 & @xmath1057 & @xmath1058 & @xmath1059 & @xmath1060 & @xmath1061 & @xmath239 & r + 2366 & 0.0380 & 3.52 & 3.21 & @xmath478 & @xmath1062 & @xmath1063 & @xmath1064 & @xmath1065 & @xmath1066 & @xmath996 & @xmath607 & r + 2368 & 0.0294 & 5.28 & 4.82 & @xmath1067 & @xmath1068 & @xmath1069 & @xmath1070 & @xmath1071 & @xmath1072 & @xmath1073 & @xmath270 & r + 2371 & 0.0135 & 17.82 & 12.44 & @xmath1074 & @xmath1075 & @xmath1076 & @xmath1077 & @xmath1078 & @xmath1079 & @xmath1080 & @xmath448 & d + 2377 & 0.0391 & 4.06 & 3.21 & @xmath1081 & @xmath1082 & @xmath1083 & @xmath676 & @xmath304 & @xmath1021 & @xmath245 & @xmath575 & r + 2381 & 0.0374 & 5.94 & 3.21 & @xmath1084 & @xmath1085 & @xmath1086 & @xmath1087 & @xmath1088 & @xmath1089 & @xmath973 & @xmath246 & t + 2389 & 0.0407 & 3.52 & 3.21 & @xmath457 & @xmath1090 & @xmath1091 & @xmath1092 & @xmath320 & @xmath1093 & @xmath827 & @xmath262 & r + 2422 & 0.0439 & 4.17 & 3.21 & @xmath638 & @xmath1094 & @xmath1095 & @xmath376 & @xmath1096 & @xmath1097 & @xmath284 & @xmath687 & r + 2424 & 0.0488 & 5.28 & 4.82 & @xmath1098 & @xmath1099 & @xmath1099 & @xmath1100 & @xmath1101 & @xmath1102 & @xmath300 & @xmath607 & r + 2431 & 0.0419 & 3.52 & 3.21 & @xmath294 & @xmath1103 & @xmath772 & @xmath866 & @xmath1104 & @xmath1105 & @xmath1106 & @xmath600 & r + 2438 & 0.0440 & 7.92 & 4.82 & @xmath533 & @xmath1107 & @xmath1108 & @xmath1109 & @xmath812 & @xmath1102 & @xmath300 & @xmath293 & r + 2440 & 0.0437 & 4.24 & 3.21 & @xmath1110 & @xmath1111 & @xmath1112 & @xmath1113 & @xmath1114 & @xmath1115 & @xmath413 & @xmath364 & r + 2454 & 0.0442 & 5.28 & 4.41 & @xmath638 & @xmath1116 & @xmath1117 & @xmath1118 & @xmath1119 & @xmath1120 & @xmath496 & @xmath1121 & d + 2467 & 0.0452 & 4.06 & 3.21 & @xmath1122 & @xmath1123 & @xmath1124 & @xmath676 & @xmath1125 & @xmath1126 & @xmath392 & @xmath338 & r + 2494 & 0.0402 & 5.28 & 3.21 & @xmath316 & @xmath1127 & @xmath1128 & @xmath1129 & @xmath1130 & @xmath1131 & @xmath1132 & @xmath406 & r + 2495 & 0.0242 & 7.92 & 6.42 & @xmath901 & @xmath1133 & @xmath1134 & @xmath1135 & @xmath1136 & @xmath1137 & @xmath801 & @xmath615 & t + 2528 & 0.0443 & 5.28 & 4.82 & @xmath1138 & @xmath1139 & @xmath1140 & @xmath1141 & @xmath1142 & @xmath321 & @xmath322 & @xmath622 & r + 2548 & 0.0281 & 11.88 & 6.42 & @xmath955 & @xmath932 & @xmath1143 & @xmath1144 & @xmath1145 & @xmath777 & @xmath1146 & @xmath716 & t + 2585 & 0.0437 & 2.15 & 1.61 & @xmath1147 & @xmath1148 & @xmath1149 & @xmath1150 & @xmath1151 & @xmath1152 & @xmath1153 & @xmath687 & r + 2587 & 0.0437 & 6.60 & 4.82 & @xmath541 & @xmath1154 & @xmath1155 & @xmath1156 & @xmath1157 & @xmath1158 & @xmath1159 & @xmath448 & r + 2648 & 0.0466 & 5.28 & 3.21 & @xmath533 & @xmath1160 & @xmath1161 & @xmath1162 & @xmath1163 & @xmath1164 & @xmath1165 & @xmath464 & r + 2664 & 0.0415 & 11.88 & 6.42 & @xmath294 & @xmath1166 & @xmath1167 & @xmath1168 & @xmath1169 & @xmath398 & @xmath399 & @xmath309 & t + 2676 & 0.0313 & 5.28 & 4.82 & @xmath286 & @xmath498 & @xmath696 & @xmath1170 & @xmath1171 & @xmath1172 & @xmath1173 & @xmath1121 & r + 2684 & 0.0308 & 5.28 & 3.21 & @xmath1174 & @xmath1175 & @xmath1176 & @xmath1177 & @xmath1178 & @xmath1072 & @xmath1073 & @xmath246 & r + 2696 & 0.0400 & 5.28 & 3.21 & @xmath541 & @xmath1082 & @xmath1179 & @xmath1180 & @xmath1181 & @xmath1182 & @xmath1183 & @xmath406 & r + 2712 & 0.0421 & 11.88 & 9.63 & @xmath568 & @xmath1134 & @xmath1184 & @xmath829 & @xmath1185 & @xmath936 & @xmath937 & @xmath414 & t + 2756 & 0.0398 & 3.52 & 3.21 & @xmath549 & @xmath1186 & @xmath1187 & @xmath1188 & @xmath1189 & @xmath692 & @xmath1190 & @xmath456 & r + 2760 & 0.0371 & 5.28 & 3.21 & @xmath717 & @xmath727 & @xmath1191 & @xmath1192 & @xmath1193 & @xmath697 & @xmath698 & @xmath485 & r + 2783 & 0.0430 & 7.92 & 4.82 & @xmath263 & @xmath1194 & @xmath1195 & @xmath842 & @xmath1196 & @xmath1197 & @xmath1198 & @xmath567 & r + 2820 & 0.0270 & 7.92 & 4.82 & @xmath1199 & @xmath1200 & @xmath724 & @xmath1201 & @xmath1202 & @xmath1203 & @xmath808 & @xmath1204 & t + 2829 & 0.0231 & 7.92 & 3.21 & @xmath1205 & @xmath1206 & @xmath1207 & @xmath1208 & @xmath1209 & @xmath1210 & @xmath1073 & @xmath1211 & r + 2836 & 0.0341 & 3.52 & 3.21 & @xmath407 & @xmath1212 & @xmath1212 & @xmath1213 & @xmath1214 & @xmath1215 & @xmath950 & @xmath485 & r + 2838 & 0.0458 & 5.28 & 3.21 & @xmath541 & @xmath332 & @xmath1216 & @xmath1217 & @xmath1218 & @xmath1219 & @xmath1220 & @xmath637 & t + 2877 & 0.0386 & 3.52 & 3.21 & @xmath1221 & @xmath831 & @xmath1222 & @xmath1223 & @xmath1224 & @xmath237 & @xmath238 & @xmath393 & r + 2931 & 0.0180 & 17.82 & 9.63 & @xmath1225 & @xmath1226 & @xmath1227 & @xmath1228 & @xmath1206 & @xmath1152 & @xmath1153 & @xmath393 & r + 2971 & 0.0360 & 5.28 & 3.21 & @xmath1229 & @xmath332 & @xmath1230 & @xmath1231 & @xmath1232 & @xmath1233 & @xmath1234 & @xmath239 & r + 2982 & 0.0425 & 7.92 & 4.82 & @xmath568 & @xmath1235 & @xmath1236 & @xmath1237 & @xmath1238 & @xmath1197 & @xmath1198 & @xmath406 & r + 2983 & 0.0446 & 5.28 & 3.21 & @xmath1239 & @xmath1240 & @xmath1241 & @xmath676 & @xmath1242 & @xmath1243 & @xmath1244 & @xmath575 & r + 3049 & 0.0449 & 4.17 & 3.21 & @xmath1229 & @xmath1245 & @xmath1246 & @xmath633 & @xmath1247 & @xmath1248 & @xmath674 & @xmath440 & r + 3073 & 0.0229 & 11.88 & 6.42 & @xmath1249 & @xmath1250 & @xmath1251 & @xmath1252 & @xmath1253 & @xmath807 & @xmath1254 & @xmath1121 & r + 3097 & 0.0438 & 3.52 & 3.21 & @xmath1255 & @xmath1256 & @xmath1257 & @xmath1258 & @xmath1259 & @xmath1260 & @xmath238 & @xmath414 & r + 3122 & 0.0487 & 3.22 & 3.21 & @xmath1261 & @xmath1262 & @xmath1263 & @xmath1264 & @xmath1265 & @xmath1266 & @xmath606 & @xmath270 & r + 3124 & 0.0415 & 7.92 & 8.03 & @xmath1225 & @xmath449 & @xmath1267 & @xmath376 & @xmath1268 & @xmath1269 & @xmath1270 & @xmath393 & t + 3131 & 0.0437 & 7.92 & 4.82 & @xmath533 & @xmath1271 & @xmath1272 & @xmath1273 & @xmath793 & @xmath1274 & @xmath1220 & @xmath414 & r + 3148 & 0.0291 & 11.88 & 8.03 & @xmath357 & @xmath471 & @xmath1275 & @xmath1276 & @xmath569 & @xmath1277 & @xmath455 & @xmath456 & r + 3191 & 0.0442 & 5.28 & 4.82 & @xmath1229 & @xmath1278 & @xmath1279 & @xmath1280 & @xmath1281 & @xmath1282 & @xmath421 & @xmath477 & t + 3218 & 0.0431 & 17.82 & 11.24 & @xmath357 & @xmath1283 & @xmath535 & @xmath1284 & @xmath1285 & @xmath767 & @xmath768 & @xmath372 & r + 3247 & 0.0450 & 5.28 & 3.21 & @xmath1026 & @xmath1286 & @xmath1287 & @xmath1288 & @xmath1289 & @xmath1290 & @xmath496 & @xmath239 & r + 3308 & 0.0345 & 3.52 & 3.21 & @xmath967 & @xmath1291 & @xmath1292 & @xmath1293 & @xmath1294 & @xmath1295 & @xmath1296 & @xmath432 & t + 3343 & 0.0456 & 5.28 & 3.21 & @xmath1098 & @xmath945 & @xmath1297 & @xmath1298 & @xmath1299 & @xmath1300 & @xmath1301 & @xmath440 & r + 3374 & 0.0268 & 11.88 & 4.82 & @xmath795 & @xmath796 & @xmath1302 & @xmath1303 & @xmath1304 & @xmath995 & @xmath996 & @xmath338 & t + 3392 & 0.0388 & 11.88 & 6.42 & @xmath1225 & @xmath840 & @xmath1305 & @xmath1306 & @xmath1307 & @xmath1308 & @xmath1309 & @xmath687 & r + 3402 & 0.0437 & 4.00 & 3.21 & @xmath316 & @xmath1310 & @xmath1167 & @xmath572 & @xmath1311 & @xmath1312 & @xmath1301 & @xmath1313 & r + 3418 & 0.0448 & 17.82 & 12.84 & @xmath1314 & @xmath1315 & @xmath1316 & @xmath1317 & @xmath1318 & @xmath1319 & @xmath1320 & @xmath742 & t + 3438 & 0.0250 & 3.52 & 3.21 & @xmath407 & @xmath1321 & @xmath1322 & @xmath1323 & @xmath1324 & @xmath1325 & @xmath1326 & @xmath1313 & r + 3440 & 0.0293 & 11.88 & 8.03 & @xmath1314 & @xmath1327 & @xmath1328 & @xmath1329 & @xmath1226 & @xmath502 & @xmath503 & @xmath583 & r + 3442 & 0.0182 & 7.92 & 6.42 & @xmath735 & @xmath1330 & @xmath1331 & @xmath1332 & @xmath1333 & @xmath1334 & @xmath1335 & @xmath393 & r + 3457 & 0.0493 & 3.10 & 3.21 & @xmath1336 & @xmath1169 & @xmath1337 & @xmath1338 & @xmath1339 & @xmath1340 & @xmath674 & @xmath293 & r + 3468 & 0.0188 & 7.92 & 6.42 & @xmath278 & @xmath1341 & @xmath1342 & @xmath676 & @xmath549 & @xmath1343 & @xmath1344 & @xmath346 & r + 3471 & 0.0176 & 11.88 & 4.82 & @xmath1225 & @xmath1007 & @xmath1345 & @xmath1346 & @xmath1347 & @xmath1348 & @xmath1349 & @xmath448 & t + 3499 & 0.0283 & 11.88 & 6.42 & @xmath735 & @xmath498 & @xmath491 & @xmath1350 & @xmath1351 & @xmath1352 & @xmath644 & @xmath448 & r + 3523 & 0.0498 & 2.35 & 3.21 & @xmath373 & @xmath1353 & @xmath1216 & @xmath1354 & @xmath1355 & @xmath692 & @xmath1190 & @xmath477 & r + 3551 & 0.0478 & 5.35 & 3.21 & @xmath762 & @xmath1356 & @xmath1357 & @xmath1358 & @xmath1359 & @xmath1360 & @xmath1061 & @xmath456 & r + 3582 & 0.0182 & 7.92 & 4.82 & @xmath955 & @xmath1361 & @xmath1362 & @xmath1363 & @xmath1364 & @xmath773 & @xmath1365 & @xmath432 & r + 3620 & 0.0431 & 7.92 & 4.82 & @xmath568 & @xmath1366 & @xmath1367 & @xmath1368 & @xmath1369 & @xmath1164 & @xmath399 & @xmath485 & r + 3644 & 0.0445 & 5.28 & 3.21 & @xmath247 & @xmath1370 & @xmath1371 & @xmath1372 & @xmath1373 & @xmath1021 & @xmath910 & @xmath239 & r + 3657 & 0.0280 & 17.82 & 7.62 & @xmath1374 & @xmath1375 & @xmath1376 & @xmath1377 & @xmath1378 & @xmath1379 & @xmath1198 & @xmath716 & d + 3676 & 0.0365 & 3.52 & 3.21 & @xmath1098 & @xmath1166 & @xmath332 & @xmath1380 & @xmath1381 & @xmath1382 & @xmath1383 & @xmath504 & r + 3689 & 0.0419 & 7.92 & 6.42 & @xmath735 & @xmath1384 & @xmath700 & @xmath1385 & @xmath1386 & @xmath1387 & @xmath961 & @xmath716 & t + 3710 & 0.0179 & 7.92 & 6.42 & @xmath1225 & @xmath331 & @xmath851 & @xmath1388 & @xmath1389 & @xmath1390 & @xmath1391 & @xmath593 & r + 3725 & 0.0499 & 11.88 & 8.03 & @xmath533 & @xmath1392 & @xmath256 & @xmath1393 & @xmath1366 & @xmath1394 & @xmath778 & @xmath456 & r + 3726 & 0.0494 & 5.10 & 3.21 & @xmath1067 & @xmath1395 & @xmath1396 & @xmath1397 & @xmath1398 & @xmath1399 & @xmath1400 & @xmath440 & r + [ tab_sp ] | it has been suggested that merging plays an important role in the formation and the evolution of elliptical galaxies . while gas dissipation by star formation is believed to steepen metallicity and color gradients of the merger products , mixing of stars through dissipation - less merging ( dry merging ) is believed to flatten them . in order to understand the past merging history of elliptical galaxies
, we studied the optical - near infrared ( nir ) color gradients of 204 elliptical galaxies .
these galaxies are selected from the overlap region of the sloan digital sky survey ( sdss ) stripe 82 and the ukirt infrared deep sky survey ( ukidss ) large area survey ( las ) .
the use of the optical and the nir data ( @xmath0 , and @xmath1 ) provides large wavelength baselines , and breaks the age - metallicity degeneracy , allowing us to derive age and metallicity gradients .
the use of the deep sdss stripe 82 images make it possible for us to examine how the color / age / metallicity gradients are related to merging features .
we find that the optical - nir color and the age / metallicity gradients of elliptical galaxies with tidal features are consistent with those of relaxed ellipticals suggesting that the two populations underwent a similar merging history on average and that mixing of stars was more or less completed before the tidal features disappear .
elliptical galaxies with dust features have steeper color gradients than the other two types , even after masking out dust features during the analysis , which can be due to a process involving wet merging .
more importantly , we find that the scatter in the color / age / metallicity gradients of the relaxed and merging feature types , decreases as their luminosities ( or masses ) increase at @xmath2 but stays to be large at lower luminosities .
mean metallicity gradients appear nearly constant over the explored mass range , but a possible flattening is observed at the massive end . according to our toy model that predicts how the distribution of metallicity gradients changes as a result of major dry merging , the mean metallicity gradient should flatten by 40% and its scatter become smaller by 80% per a mass doubling scale if ellipticals evolve only through major dry merger .
our result , although limited by a number statistics at the massive end , is consistent with the picture that major dry merging is an important mechanism for the evolution for ellipticals at @xmath2 , but less important at the lower mass range . |
the @xmath0 annihilation at low energies , below 3 gev , is the main source of information on the properties of the light vector mesons ( @xmath1 , @xmath2 , @xmath3 ) and their excited states
. other important application of the @xmath0 annihilation is calculation of the hadronic contribution into the anomalous magnetic moment of muon @xmath4 and the running electromagnetic constant .
the difference between the measured value of @xmath5 @xcite and recent standard model calculations @xcite varies from 3.2 to 3.4 sigma .
currently the hadronic leading - order contribution into @xmath5 is dominant source of uncertainty of the standard model @xmath5 prediction .
it is evaluated using experimental data on @xmath0 annihilation via dispersion integral @xmath6 where @xmath7 is the ratio of the total cross section of the @xmath0 annihilation into hadrons to the @xmath8 cross section , and @xmath9 is the kernel function decreasing monotonically as approximately @xmath10 .
due to such @xmath11-dependence of @xmath9 the main contribution into @xmath4 comes from the low energy region .
table [ tab1 ] shows the contributions of the different @xmath0 center - of - mass ( c.m . )
energy regions and the different processes into the @xmath4 @xcite . [ cols="<,^ " , ] table [ tab2 ] also contains the kloe preliminary result based on statistics collected in 2002 @xcite .
the new software trigger , improvement in the offline background filter , and more accurate calculation of the luminosity were introduced , leading to decrease of systematic error .
a bias in the evaluation on the trigger correction for the published analysis of 2001 data was found which affects at the low @xmath12 mass region .
the 0.7% decrease of the predicted cross section for bhabha scattering used for normalization , also was found . the updated 2001 @xmath13 cross section and the preliminary 2002 cross section
are compared in fig .
[ kloe_2pi_01 ] . in spite of nonstatistical difference between the two measurements seen near the maximum of @xmath1-peak ,
the values of @xmath14 calculated over full mass region , @xmath15 for 2001 data and @xmath16 for 2002 data , are in good agreement .
the kloe also presents the measurement of @xmath13 cross section with the isr photon detected at large angle ( @xmath17 ) . in this case
the threshold mass region becomes accessible .
comparison of the small angle and large angle measurements are demonstrated in fig .
[ kloe_2pi_02 ] .
the systematic uncertainty for the large angle measurement is shown by the shaded band .
the main source of the uncertainty is subtraction of the final state radiation background .
the large angle and small angle measurements agree within the systematic errors .
the good cross check for the isr measurement of @xmath13 is simultaneous selection of @xmath18 events .
this allows to test isr method using the qed process with known cross section .
taking @xmath19 ratio reduces the systematic uncertainty , in particular , due to radiative effects .
kloe plans to measure @xmath19 ratio for the final result .
other kloe result is the measurement of the @xmath20 cross section in the vicinity of @xmath3-resonance @xcite . in this region two process , nonresonant @xmath21 production and @xmath22 decay , determine the total @xmath21 cross section .
the decay reveals itself as interference pattern in the cross section energy dependence .
the process was studied in the two final states : @xmath23 and @xmath24 . in both modes
the real and imaginary parts of the relative interference amplitude were measured .
the obtained interference parameters and @xmath22 branching fraction agree with the results of the previous snd measurement @xcite but have better accuracy . from the ratio of the @xmath20 cross sections measured in the two different final states , the ratio of the @xmath2 branching fractions was obtained : @xmath25 .
the kloe result has accuracy better than the world average value ( @xmath26 ) and differs from it by 2 standard deviations .
the inclusion of the kloe result into the global fit of the @xmath2 parameters will lead to growth of the @xmath2 leptonic width .
the babar isr program includes the study of all significant hadronic processes @xmath27 from threshold to c.m .
energy about 4.5 gev .
one of the purposes of this program is measurement of @xmath28 in the 12 gev energy range with improved accuracy .
the current status of babar measurements in this mass range is demonstrated in fig .
[ babar ] . to obtain the total hadronic cross section the @xmath12 , @xmath29 , @xmath30 , @xmath31 , @xmath32 , @xmath33 ,
@xmath34 processes must be measured additionally .
the final states @xmath12 , @xmath29 , @xmath31 , and @xmath34 are currently under study . here
we discuss the latest babar results on @xmath35 , @xmath36 , @xmath37 , @xmath38 , @xmath39 , @xmath40 modes and measurement of strange baryon form factors .
* @xmath41 .
* this channel is very important for calculation of @xmath4 and spectroscopy of the excited @xmath1-like states .
the babar preliminary results on @xmath41 cross section measurement is presented in fig .
[ 4pi_1 ] .
the current systematic error of the measurement varies from 8% in the cross section peak to 14% at 4.5 gev .
the comparison with existing data is shown in fig .
[ 4pi_2 ] . in the energy range below 1.4 gev
the the babar results are in good agreement with snd data . for higher energies the accuracy of the cross section
is significantly improved .
the babar data exceed the cross section measured by dm2 by about 25% . from the analysis of @xmath42 and @xmath43 mass distributions
it was found that the @xmath21 , @xmath44 and @xmath45 are dominant intermediate states for @xmath41 process .
the @xmath46 contribution is also seen .
the @xmath47 process was observed for the first time , its contribution was found be surprisingly large , about 30% of the total @xmath48 cross section at @xmath49 gev .
* @xmath50 @xcite . *
the measured cross section for @xmath50 is shown in fig .
[ babar ] .
the systematic error of the measurement is about 7% . in @xmath51 mass spectrum the @xmath2 and @xmath52 peaks
were observed corresponding the @xmath53 and @xmath54 intermediate states . non-@xmath53 and non-@xmath54 events have @xmath55 structure , where @xmath56 may be @xmath57 or @xmath58 .
the cross sections for all three selected components were measured .
the @xmath54 and @xmath53 cross sections agree with previous experiments but have significantly better accuracy .
the @xmath59 cross section is shown in fig .
[ babar ] . a part of @xmath53 events contain @xmath60 meson .
the cross sections for the @xmath61 and non-@xmath61 components of the @xmath53 final state is shown in figs .
[ babar_omf0 ] and [ babar_om2pi ] .
the latter cross section was fitted by the sum of the contributions of @xmath2-like resonances . from the fit , the parameters @xmath62 meson , @xmath63 gev and @xmath64 ,
were determined .
* @xmath65 @xcite .
* the @xmath37 final state was studied for the first time .
three intermediate states , @xmath66 , @xmath67 , and @xmath68 , were found to contribute to @xmath37 .
the @xmath65 total cross section , and the cross sections for @xmath67 and @xmath68 final states are shown in fig .
[ babar_4pieta ] .
the two latter cross sections are described by the single breit - wigner function with parameters : @xmath69 gev , @xmath70 gev for @xmath67 and @xmath71 gev , @xmath72 gev for @xmath68 .
both sets of parameters are close to those listed in the pdg tables @xcite for the @xmath73 .
* @xmath74 and @xmath75 @xcite . * the @xmath74 and @xmath76 cross sections measured for the first time are shown in fig .
[ babar_2k3pi ] . in @xmath42 mass spectrum for @xmath40 channel
the @xmath52 and @xmath2 peaks are observed corresponding the @xmath77 and @xmath78 intermediate states .
the cross section for @xmath79 is shown in fig .
[ babar_2k3pi ] . the @xmath80 process will be discussed in the next subsection .
* @xmath81 , @xmath82 , @xmath83 , @xmath84 . *
the @xmath85 is the dominant decay mode of the @xmath86 meson .
therefore the study of @xmath87 process is very important for determination of its parameters .
this process was studied at babar in the two final states , @xmath38 and @xmath82 .
the measured cross sections for both channels are shown in fig .
[ babar_kkpi ] .
the babar results are in reasonable agreement with the results of the previous measurements by dm1 and dm2 but have much smaller statistical and systematic errors .
the systematic uncertainty of the babar measurement is about 5% .
[ babar_kkpi](right ) shows the dalitz plot for @xmath82 final state .
it is seen that the @xmath88 and @xmath89 intermediate state give the main contributions to the @xmath85 production .
for @xmath82 final state both neutral @xmath90 and charged @xmath91 intermediate states are possible .
since the @xmath92 and @xmath91 amplitudes are the sum and the difference of the isovector and isoscalar amplitudes , the dalitz plot population strongly depends on isospin composition . from the dalitz plot analysis the moduli and relative phase of the isoscalar and isovector amplitudes both for @xmath88 and @xmath89 final states were determined .
the obtained isoscalar and isovector @xmath93 cross sections is shown in fig .
[ babar_phieta ] .
the @xmath94 process was studied by babar in the @xmath95 and @xmath96 final states .
the resulted cross section is shown in fig .
[ babar_phieta](right ) .
the @xmath94 is the best channel for study of excited @xmath3 state .
the contributions of @xmath2-like states are suppressed by ozi rule .
the global fit of the @xmath97 and @xmath81 cross sections , isovector and isoscalar @xmath88 amplitudes , and their relative phase was performed to determinate parameters of the @xmath3 and @xmath1 excitations decaying into these final states .
the fit results are shown in fig .
[ babar_phieta ] .
the isovector @xmath88 component is described by a single broad resonance which mass is compatible with that for @xmath98 @xcite .
the obtained @xmath99 parameters are following : @xmath100 mev , @xmath101 mev , @xmath102 ev , @xmath103 . in the @xmath84
cross section the peak with @xmath104 mev and @xmath105 mev is seen with 2 sigma significance .
these parameters are close to the mass and width of the state recently observed by babar in @xmath106 mode @xcite .
the cross section for @xmath107 measured by babar in the @xmath38 final state is shown in fig .
[ babar_phipi0 ] .
this reaction is suitable for search of exotic isovector resonances . for the ordinary isovector states ,
@xmath108 decay is suppressed by ozi rule . with available limited statistics the @xmath107 cross section
is described by the single resonance with @xmath109 mev and @xmath110 mev .
these parameters are compatible with those for @xmath111 @xcite
. * baryon form factors @xcite .
* babar continued the study of the baryon time - like electromagnetic form factors started by the work @xcite on the @xmath112 reaction .
the cross section for @xmath113 , where @xmath114 is a spin-1/2 baryon depends on two form factors , magnetic ( @xmath115 ) and electric ( @xmath116 ) . from the measurements of the total cross section for @xmath117 , @xmath118 , and @xmath119 reactions ,
the effective @xmath120 , @xmath120-@xmath121 , and @xmath121 form factors were extracted .
these form factors together with the previously measured @xcite proton form factor are shown in fig .
[ babar_baryons ] . for @xmath122 channel
the @xmath120 angular distributions were studied and the ratio of the form factors was determined for the two energy regions : @xmath123 for 2.232.40 gev and @xmath124 for 2.42.8 gev .
the obtained ratios are consistent both with @xmath125 and with the results for @xmath112 , where this ratio was found to be significantly greater than unity near threshold .
the measurement of @xmath120 polarization via its decay to @xmath126 gives possibility to extract the relative phase between the electric and magnetic form factors .
the limited experimental statistics allowed to set only very weak limits on this phase : @xmath127 .
the hypothesis of the vector current conservation ( cvc ) connects the isovector @xmath0 cross section with the vector spectral function in @xmath128 decay : @xmath129 this relation holds in the limit of exact isospin invariance . therefore the @xmath128 spectral function must be corrected for isospin - breaking effects before comparison with the corresponding @xmath0 cross section .
such corrections were considered in ref .
the pion form factor and @xmath130 cross sections calculated from the isospin - breaking corrected @xmath128 spectral functions are compared with the direct @xmath0 measurements in figs .
[ tau2pi],[tau4pi ] . fig .
[ tau2pi ] was taken from ref .
the @xmath128 @xmath43 spectral function is obtained by averaging the aleph , cleo , and opal data .
the @xmath128 @xmath131 spectral functions were taken from ref .
for all three reactions a significant discrepancy between @xmath0 and @xmath128 data is observed .
these disagreements are difficult to explain by unaccounted isospin breaking effects .
new precise @xmath128 data are needed to confirm the observed discrepancies .
the significant progress has been reached in the measurements of the low energy hadronic @xmath0 cross sections in last few years , leading to improvement of the accuracy of the total hadronic @xmath0 cross section . in particular , using the recent cmd-2 , snd , kloe , and babar data reduces the error of the @xmath4 from about @xmath132 in 2003 to about @xmath133 in 2007 .
this error is now comparable with the error of the light - by - light scattering contribution , also @xmath134 @xcite .
there are , however , the discrepancy between the energy dependencies of the pion form factors measured directly in @xmath0 collisions at vepp-2 m and via isr at kloe .
the new isr data on @xmath135 are expected from the kloe and babar .
both experiments plan to measure @xmath136 ratio .
this will reduce the systematic error , in particular , due to radiative effects .
the new precise data on multihadron cross sections for the energy region from 1 to 3 gev obtained at vepp-2 m and in babar experiment allows to improve our knowledge of the properties of the @xmath1 , @xmath2 , and @xmath3 excitations .
all major isoscalar channels ( @xmath42 , @xmath137 , @xmath138 , @xmath139 , @xmath84 ) have been measured .
some theoretical input is needed to perform a global fit to the measured cross sections and extract a complete set of parameters of the isoscalar resonances .
the many isovector cross sections ( @xmath140 , @xmath141 , @xmath142 , @xmath139 ) are also measured . the final babar result on @xmath143
is expected soon .
the discrepancy between the @xmath0 and isospin - breaking corrected data @xmath128 data is observed for @xmath43 and @xmath131 final states .
this discrepancy is difficult to explain by unaccounted isospin breaking effects .
new data on @xmath144 and @xmath145 decays from @xmath114 factories , and isr data on @xmath43 channel from kloe and babar should provide a test the existing @xmath0 and @xmath128 measurements .
99 g. w. bennett _ et al . _ [ muon g-2 collaboration ] , phys .
d * 73 * , 072003 ( 2006 ) [ arxiv : hep - ex/0602035 ] .
m. davier , nucl .
suppl . * 169 * , 288 ( 2007 ) [ arxiv : hep - ph/0701163 ] .
f. jegerlehner , arxiv : hep - ph/0703125 .
k. hagiwara , a. d. martin , d. nomura and t. teubner , phys .
b * 649 * , 173 ( 2007 ) [ arxiv : hep - ph/0611102 ] .
yao _ et al .
_ [ particle data group ] , j. phys .
g * 33 * , 1 ( 2006 ) .
j. z. bai _ et al . _
[ bes collaboration ] , phys .
lett . * 88 * , 101802 ( 2002 ) [ arxiv : hep - ex/0102003 ] .
g. bonneau and f. martin , nucl .
b * 27 * , 381 ( 1971 ) .
m. n. achasov _ et al . _
[ snd collaboration ] , arxiv:0707.2279 [ hep - ex ] .
m. n. achasov _ et al .
_ [ snd collaboration ] , arxiv:0709.1007 [ hep - ex ] . r. r. akhmetshin _ et al . _ [ cmd-2 collaboration ] , phys .
b * 578 * , 285 ( 2004 ) [ arxiv : hep - ex/0308008 ] .
r. r. akhmetshin _ et al .
_ [ cmd-2 collaboration ] , phys .
b * 648 * , 28 ( 2007 ) [ arxiv : hep - ex/0610021 ] .
r. r. akhmetshin _ et al . _
[ cmd-2 collaboration ] , jetp lett . * 84 * , 413 ( 2006 ) [ pisma zh .
* 84 * , 491 ( 2006 ) ] [ arxiv : hep - ex/0610016 ] . v.
m. aulchenko _ et al . _
[ cmd-2 collaboration ] , jetp lett . * 82 * , 743 ( 2005 ) [ pisma zh .
teor . fiz .
* 82 * , 841 ( 2005 ) ] [ arxiv : hep - ex/0603021 ] .
m. n. achasov _ et al . _
[ snd collaboration ] , j. exp .
* 103 * , 380 ( 2006 ) [ zh
. eksp .
fiz . * 130 * , 437 ( 2006 ) ] [ arxiv : hep - ex/0605013 ] .
m. n. achasov _ et al . _
[ snd collaboration ] , phys .
b * 474 * , 188 ( 2000 ) [ arxiv : hep - ex/0001048 ] .
a. aloisio _ et al .
_ [ kloe collaboration ] , phys .
b * 606 * , 12 ( 2005 ) [ arxiv : hep - ex/0407048 ] .
f. ambrosino _ et al .
_ [ kloe collaboration ] , arxiv:0707.4078 [ hep - ex ] .
f. ambrosino _ et al . _
[ kloe collaboration ] , arxiv:0707.4130 [ hep - ex ] . v.
m. aulchenko _ et al .
_ , j. exp .
* 90 * ( 2000 ) 927 [ zh .
* 90 * ( 2000 ) 1067 ] .
b. aubert _ et al . _ [ babar collaboration ] , arxiv:0708.2461 [ hep - ex ] .
b. aubert _ et al . _ [ babar collaboration ] , phys .
d * 74 * , 091103 ( 2006 ) [ arxiv : hep - ex/0610018 ] .
b. aubert _ et al .
_ [ babar collaboration ] , arxiv:0709.1988 [ hep - ex ] .
b. aubert _ et al .
_ [ babar collaboration ] , phys .
d * 73 * , 012005 ( 2006 ) [ arxiv : hep - ex/0512023 ] .
v. cirigliano , g. ecker and h. neufeld , jhep * 0208 * , 002 ( 2002 ) [ arxiv : hep - ph/0207310 ]
. s. schael _ et al . _
[ aleph collaboration ] , phys .
* 421 * , 191 ( 2005 ) [ arxiv : hep - ex/0506072 ] .
j. bijnens and j. prades , mod .
a * 22 * , 767 ( 2007 ) [ arxiv : hep - ph/0702170 ] . | the recent results of the cmd-2 , snd , kloe , and babar experiments on @xmath0 annihilation into hadrons at low energies are reviewed . |
i have tried in this paper to separate mathematics from physical interpretation . in this closing passage
however , i will bring the two together again .
there is no limitation in bell s theorem on the space in which the hidden variables live .
the `` measurement functions '' could be imagined to `` perform '' or `` implement '' calculations in any suitable algebraic or other mathematical framework , and hidden variables can include elements of any exotic mathematical space .
moreover there is no objection whatever within bell s theorem to allow the outcomes , though of necessity encoded as @xmath15 and @xmath1 , to be thought of being members of a larger mathematical space than the set @xmath66 . in the context of chsh
, all hidden variables can be reduced to , or subsumed in , the outcomes of the other measurements to the two which were actually done .
`` realism '' , however it is defined , comes down , effectively , to the mathematical existence of the outcomes of unperformed experiments , alongside of those which were actually performed .
`` locality '' refers to the attempt to `` locate '' those counterfactual outcomes in the `` obvious '' region of space and time . alongside of the assumptions of realism and locality ( the second only being meaningful
given the first ) we need an assumption of freedom : the freedom of the experimenter to perform either measurement .
this does not need to involve metaphysical assumptions either of free will or of existence of true randomness .
it just involves the assumption that the physical processes going on at one measurement location ca nt have access to the measurement choice made at the other location , till after the measurement outcome has been committed to . and
of course , bell s theorem applies to ordinary correlations computed in the ordinary way on christian s outcomes @xmath29 and @xmath30 , which as we have seen are actually always perfectly anti correlated , whatever the measurement settings .
no violation of bell s theorem ( chsh inequality version ) . in real experiments ,
ordinary correlations are computed in the ordinary way on binary outcomes and _ do _ violate the chsh inequality .
so even if christian s algebra had been correct , what relevance does it have to the real world ?
as we have seen , christian s work applies to correlations obtained by dividing the raw correlation between measurement outcomes by the pure bivectors @xmath67 and @xmath24 , and within his own model would lead to standardized correlations which are not even real numbers .
this simply has got nothing at all to do with bell s own programme , as far as it is usually interpreted .
some of those writing critical evaluation of christian s work have expressed the hope that it might at least provide a mathematical framework for the theoretical side of quantum mechanics , in which the usual structure of hilbert spaces , projection operators , and so on , could be entirely replaced by a mathematical structure having a much closer connection with , for instance , the geometry of the real world .
it seems to this author that that is indeed a legitimate quest .
however in view of the failure of this particular attempt , those wanting to do this job are going to have to look elsewhere .
there remains a psychological question , why so strong a need is felt by so many researchers to `` disprove bell '' in one way or another ? at a rough guess ,
at least one new proposal comes up per year .
many pass by unnoticed , but from time to time one of them attracts some interest and even media attention .
having studied a number of these proposals in depth , i see two main strategies of would - be bell - deniers .
but please notice , i do not mean to imply that these strategies are deliberate : i believe they are found `` accidentally '' ; i have no doubt of the sincerity of the proposers .
the first strategy ( the strategy , i would guess , in the case in question ) is to build elaborate mathematical models of such complexity and exotic nature that the author him or herself is probably the only person who ever worked through all the details . somewhere in the midst of the complexity
a simple mistake is made , usually resulting from suppression of an important index or variable .
there is a hidden and non - local hidden variable .
the second strategy is to simply build elaborate versions of detection loophole models .
sometimes the same proposal can be interpreted in both ways at the same time . interpreting the proposal as the result of a hidden mistake or as a detection loophole model
are both interpretations of the reader , not of the writer . according to the anna karenina principle of evolutionary biology , in order for things to succeed , everything has to go exactly right , while for failure , it suffices if any one of a myriad factors is wrong . since errors
are typically accidental and not recognized , an apparently logical deduction which leads to a manifestly incorrect conclusion does not need to allow a unique diagnosis .
if every apparently logical step had been taken with explicit citation of the mathematical rule which was being used , and in a specified context , one could say where the first misstep was taken .
but mathematics is almost never written like that , and for good reasons . the writer and the reader , coming from the same scientific community , share a host of `` hidden assumptions '' which can safely be taken for granted , as long as no self - contradiction occurs .
saying that the error actually occurred in such - and - such an equation at such - and - such a substitution depends on various assumptions .
the author who still sincerely believes in his result will therefore claim that the diagnosis is wrong because the wrong context has been assumed .
we can be grateful for christian that he has had the generosity to write his one page paper with a more or less complete derivation of his key result in a more or less completely explicit context , without distraction from the author s intended physical interpretation of the mathematics .
the mathematics should stand on its own , the interpretation is `` free '' .
my finding is that in this case , the mathematics does not stand on its own .
after posting this paper on arxiv.org i became aware that others , not surprisingly , had already published interesting critiques on christian s work .
in particular , florin moldoveanu has published a comprehensive review , moldoveanu ( 2011 ) , which also cites many other works .
his `` error 1 '' is the same as the error on which i have focussed .
i had the advantage of the availability of christian s `` one page paper '' , and restrict myself strictly to that particular presentation .
moldoveanu has carried out a review of the whole corpus of works , which is complicated by the fact that different notations and different definitions are used in different papers .
i also found it useful to strictly separate mathematical consistency from physical and metaphysical interpretation .
if the mathematics is fatally flawed at the outset , then we need not spend energy on tracking down further errors , or on debating the soundness of the ideas .
after my own preprint was posted on arxiv.org , christian has published a refutation of it , see christian ( 2012a ) .
also , his magum opus , the book christian ( 2012b ) , has now been published .
the first twenty - five pages ( which consists of 14 pages of front - matter and 11 numbered pages ) are freely available as a pdf file .
the reader will find the identical error as in the one - page paper on page number 10 of the book , in the transition from formula ( 1.23 ) to ( 1.24 ) , where christian appeals to the multiplication table ( 1.8 ) .
christian argues that he is not making an error at this point , but introducing a new postulate . in that case
it is curious that he does not draw attention to the fact that he is introducing a daring new ingredient into his model , especially in view of the fact that the new postulate contradicts the earlier made ( and used ) postulates .
christian s work has stimulated a number of interesting discoveries and inventions .
i would especially like to draw attention to sascha vongehr s `` quantum randi challenge '' on science20.com , see vongehr ( 2011 ) .
the idea is to insist that those who believe bell got it all wrong , to deliver by providing computer programs which simulate their local realistic violation of bell s inequality .
a successful simulation will get the attention of science journalists and science communicators and educators , and thereby of the whole scientific community , without having to pass the barrier of hostile peer review .
christian , j. , ( 2011 ) , _ disproof of bell s theorem _ ,
arxiv:1103.1879 , + http://arxiv.org/abs/1103.1879 christian , j. , ( 2012a ) , _ refutation of richard gill s argument against my disproof of bell s theorem _ , arxiv:1203.2529 , http://arxiv.org/abs/1203.2529 christian , j. , ( 2012b ) , _ on the origins of quantum correlations _ , arxiv:1201.0775 , + http://arxiv.org/abs/1201.0775 .
this preprint reproduces the first chapter of the book `` disproof of bell s theorem . illuminating the illusion of entanglement '' by j.j .
christian , published by brown walker , + http://www.brownwalker.com/book.php?method=isbn&book=1599425645 moldoveanu , f. ( 2011 ) , _ disproof of joy christian s `` disproof of bell s theorem '' _ ,
arxiv:1109.0535 , http://arxiv.org/abs/1109.0535 vongehr , s. ( 2011 ) , the official quantum randi challenge , + http://www.science20.com/alpha_meme/official_quantum_randi_challenge-80168
* the body of the one page paper arxiv:1103.1879v1 * | i point out a simple algebraic error in joy christian s refutation of bell s theorem . in substituting the result of multiplying some derived bivectors with one another by consultation of their multiplication table
, he confuses the generic vectors which he used to define the table , with other specific vectors having a special role in the paper , which had been introduced earlier .
the result should be expressed in terms of the derived bivectors which indeed do follow this multiplication table . when correcting this calculation , the result is not the singlet correlation any more . moreover , curiously , his normalized correlations are independent of the number of measurements and certainly do not require letting @xmath0 converge to infinity . at the same time ,
his unnormalized or raw correlations are identically equal to @xmath1 , independently of the number of measurements ! correctly computed , his standardized correlations are the bivectors @xmath2 , and they find their origin entirely in his normalization or standardization factors .
i conclude that his research program has been set up around an elaborately hidden but trivial mistake
. in at least 11 papers on quant - ph author joy christian proposes a local hidden variables model for quantum correlations which disproves bell s theorem `` by counterexample '' in a number of different settings , including the famous chsh and ghz versions .
fortunately one of these papers is just one page long and concentrates on the mathematical heart of his work .
unfortunately for his grand project , this version enables us to clearly see a rather shorter derivation of the desired correlations , which exposes an error in his own derivation .
the error is connected with an unfortunate notational ambiguity at the very start of the paper . in a nutshell : the same symbols are used to denote both a certain fixed basis @xmath3 in terms of which two other bases are defined , @xmath4 and @xmath5 , as well as to express the generic algebraic multiplication rules which these latter two bases satisfy .
this tiny ambiguity , though harmless locally , is probably the reason why later on , when apparently using the multiplication tables for the new bases , he silently shifts from the derived bases to the original special basis .
the reader will need a copy of the one page paper christian ( 2011 ) , arxiv:1103.1879 ( the body of this paper is reproduced in the appendix ) .
the context in which he works is so - called geometric algebra , which in this case means that we are working within the so - called even sub - algebra of standard clifford algebra , or if you prefer , with quaternions . at the start of the paper
the author fixes a _
bivector basis _ @xmath6 satisfying ( usual kronecker and levi - civita symbols @xmath7 , @xmath8 ) the multiplication rules @xmath9 he writes that @xmath10 , @xmath11 , @xmath12 are _ defined _ by this multiplication table , but that is not exactly true .
one can say that the _ algebra _ of bivectors is defined by the multiplication table , but the bivectors themselves are clearly not , since different bases can have the same multiplication table .
the _ bivector algebra _ is the algebra of formal real linear combinations of real numbers and @xmath10 , @xmath11 , @xmath12 .
it is therefore a four dimension real vector space , with on top of the vector space structure a multiplication operation or _
bivector product_. this is defined by combining ordinary real multiplication with the @xmath13 multiplication table of the bivector basis into the obvious @xmath14 multiplication table for the four vector - space basis elements @xmath15 , @xmath10 , @xmath11 , and @xmath12 .
the bivector product is associative but not commutative .
non - zero elements have multiplicative inverses , left and right inverses coincide .
vector - space scalar multiplication of elements of the algebra by real numbers is identical with algebraic multiplication of elements of the algebra , either from the left or the right , by elements in the one dimensional subspace generated by @xmath15 .
all this is no more than a standard definition of the quaternionic number system , which contains a unique copy of the real number system as well as many overlapping copies of the complex number system .
every quaternion has a real part and a quaternionic part .
if the latter part is zero we call the quaternion real ; if the former part is zero we call it purely quaternionic .
if we prefer to talk about elements of the bivector algebra one can correspondingly identify within the algebra two special kinds of elements which we call _ real _ and _ purely bivectorial _ respectively .
i will use the word `` bivector '' as synonym for `` element of the bivector algebra '' . according to my terminology
, each _ bivector _ can be uniquely decomposed into the sum of a _ real number _ and a _
pure bivector_. christian next defines new sets of bivectors @xmath16 , where @xmath17 or @xmath1 .
these can also each be considered to form a bivector basis , but now have multiplication tables which depend on @xmath18 . to be precise :
@xmath19 he somewhat dangerously writes that these new bivectors are `` defined '' by the new algebraic rules @xmath20 indeed there is a sense in which this is true , but within the context of the paper , it is clear that the two new `` @xmath18 bivector bases '' are defined in relation to the initially fixed ( even if arbitrary ) basis , and then just `` happen '' to satisfy the new multiplication tables .
of course , the @xmath15 bivector basis is the same as the original bivector basis , the @xmath1 bivector basis is just @xmath1 times the first . at the start of the paper christian defines `` measurement functions '' @xmath21 and @xmath22 , where @xmath18 is a binary hidden variable ( a fair coin toss , outcomes coded @xmath15 and @xmath1 ) , and @xmath23 and @xmath24 are two unit vectors in ordinary real three dimensional space .
the measurement functions are defined as two bivectors and the definition appears complicated but christian claims , and that claim can easily be checked , that @xmath25 and @xmath26 , which of course are reals .
i will give the definition and verify this claim in a moment , using a notation which will make life more easy .
but first i want to point out an important consequence . since @xmath27
, it follows that @xmath28 independently of @xmath23 , @xmath24 and @xmath18 ( each within their respective domain ) .
this striking relation is however not observed by christian . for completeness
, here are christian s definitions of @xmath29 and @xmath30 , expressed in a convenient notation which i will return to again later .
use the symbols @xmath23 and @xmath24 also to denote the pure bivectors @xmath31 and @xmath32 . the reader may check that they are both square roots of @xmath1 . define in the same way @xmath33 and @xmath34 .
clearly , @xmath35 then we have @xmath36 because @xmath37 and @xmath38 , it follows that @xmath39 in view of this algebraic fact , it is curious that computation of the correlation between many independent @xmath29 and @xmath30 measurements should proceed so laboriously , and indeed , according to a curious definition , but let us accept the definition ( his equation ( 5 ) , and coincidentally mine too ) which christian elsewhere gives reasons for : @xmath40 there is slight ambiguity in this expression concerning the division by two terms : is this division by a product , or is this two successive divisions ?
division stands for multiplication by the inverse , but is this supposed to take place on the left or on the right ?
the answer is given by inspecting the calculations in christian s equations ( 5 ) to ( 7 ) : the two terms in the denominator of this fraction are supposed to divide left and right hand side of the numerator respectively , corresponding to their order in the denominator .
those are the crucial calculations which , end of his ( 7 ) , lead to his desired result @xmath41 the fact we have previously observed concerning the product of the @xmath29 and @xmath30 measurements enables us to make a grand short cut through christian s derivation of the right hand side of his ( 7 ) from the left hand side of his ( 5 ) . with again the convention that @xmath23 and @xmath24 also denote the pure bivectors @xmath31 and @xmath32 , on substituting the value @xmath1 of all the @xmath0 products , we obtain @xmath42 .
both of @xmath23 and @xmath24 are square roots of @xmath1 . from @xmath43
we find @xmath44 , hence @xmath45 and @xmath46 .
thus @xmath47 .
this bivector product must be evaluated as @xmath48 . the second term does not vanish , unless @xmath49 .
it is conventionally denoted @xmath50 and it is a pure bivector whose coefficients are the three coefficients of the usual euclidean vector cross product @xmath51 . using the notational identification between pure bivectors and real vectors , we can even define the pure bivector wedge product by writing @xmath52 .
we also define the pure bivector dot product @xmath53 ( left hand side a bivector , right hand side a real number ) .
conclusion : christian s correlation is not @xmath54 but @xmath55 however christian prefers a more complicated derivation .
naturally , if he too follows accurately his own algebraic rules , he can not obtain a different answer .
however his derivation appears to make use of the law of large numbers . in the last complicated term before the end of his ( 7 ) , there appears a sum over @xmath0 terms each involving its own @xmath56 , and the argument is clearly that this cancels in the limit as @xmath57 , by the law of large numbers .
indeed , if this expression is correct , and unless @xmath49 , it is _ only _ in the limit that it vanishes , when we obtain @xmath58 .
something has gone wrong here .
the mistake is in the transition from his formulas ( 6 ) to ( 7 ) where christian is using the @xmath18-multiplication tables to simplify linear combinations of products of the @xmath59 .
we have already written down the correct multiplication table , which expresses these products in terms of the same basis vectors @xmath60 .
the last @xmath61 appearing in ( 7 ) should actually be @xmath62 ! the other @xmath56 in the same expression is the @xmath56 appearing on the right hand side of the @xmath56-multiplication table : altogether , @xmath63 . by definition @xmath64 .
making this correct substitution gives us a factor @xmath65 .
finally we obtain the same result as i earlier got from a shorter route .
sanity has been restored . |
the extended fkm for the spinless fermions on a @xmath14-dimensional hypercubic lattice is @xmath15 where @xmath16 @xmath17 and @xmath18 @xmath19 is the creation ( annihilation ) operator of heavy ( @xmath0 ) and light ( @xmath2 ) electron at lattice site @xmath20 .
the first two terms of ( 1 ) are the kinetic energies corresponding to quantum - mechanical hopping of @xmath2 and @xmath0 electrons between the nearest neighbor sites @xmath20 and @xmath21 with hopping probabilities @xmath22 and @xmath23 , respectively .
the third term represents the on - site coulomb interaction between the @xmath2 electrons with density @xmath24 and the @xmath0 electrons with density @xmath25 , where @xmath26 is the number of lattice sites .
usually , the hopping integral of the @xmath2 electrons is taken to be the unit of energy @xmath27 and the @xmath0-electron hopping integral is considered in the limit @xmath28 .
this is a reason why the @xmath2 electrons are called light and the @xmath0 electrons heavy . in our hf study of the extended fkm
we go beyond the usual hf approach @xcite in which only homogeneous solutions are postulated . in accordance with @xcite
we consider here also inhomogeneous solutions modeled by a periodic modulation of the order parameters : @xmath29 where @xmath30 and @xmath31 is the order parameter of the cdw state for the @xmath2- and @xmath0-electrons and @xmath32 is the excitonic average .
the nesting vector @xmath33 for @xmath11 and @xmath34 for @xmath35 . using expressions for @xmath36,@xmath37 and @xmath38 the hf hamiltonian of the extended fkm can be written as @xmath39 following the work of brydon et al .
@xcite the effective hf hamiltonian is diagonalized by canonical transformation @xmath40 where @xmath41 are solutions of the associated bogoliubov - de gennes ( bdg ) eigenequations : @xmath42 with @xmath43 and @xmath44 the corresponding energy dispersions @xmath45 and @xmath46 can be obtained directly by the fourier transform of the @xmath2- and @xmath0-electron hopping amplitudes and for the case of hypercubic lattice they are given by ( @xmath47 ) : @xmath48 the hf parameters @xmath49 can be written directly in terms of the bogoliubov - de gennes eigenvectors : @xmath50 @xmath51 @xmath52 @xmath53 @xmath54 @xmath55 where the prime denotes summation over half the brillouin zone and @xmath56\}$ ] is the fermi distribution function .
the same approach has been used recently by brydon et al .
@xcite to study the interplay between excitonic effects and the cdw instability in the fkm with on - site as well as non - local hybridization .
here we use the zero temperature variant of this procedure to describe ground - state phase diagram of the spinless fkm with @xmath0-@xmath0 hopping .
to determine the ground - state phase diagram of the extended fkm in the @xmath57-@xmath23 plane ( corresponding to selected @xmath1 ) the hf equations are solved self - consistently for each pair of @xmath58 values .
we use an exact diagonalization method to solve the bogoliubov - de gennes equation .
we start with an initial set of order parameters . by solving eq .
( 7 ) , the new order parameters are computed via eqs .
( 12 ) to ( 17 ) and are substituted back into eq .
the iteration is repeated until a desired accuracy is achieved .
first we have examined the two - dimensional extended fkm model in the intermediate coupling regime and @xmath23 negative . for this case there exists the comprehensive phase diagram of the model obtained by a cpmc technique @xcite for @xmath0-electron hopping integrals @xmath59 . according these monte - carlo studies
the phase diagram of the extended fkm consists of only three main phases , and namely , ( i ) the integer - valent state ( @xmath60 ) , ( ii ) the mixed - valent cdw state ( @xmath7 ) , and ( iii ) the mixed - valent ferroelectric state that is stable for remaining values of @xmath61 ( @xmath62 ) . in fig . 1 we have displayed typical examples of our hf solutions obtained for @xmath49 in the intermediate coupling regime @xmath63
it is seen that the extended fkm in the hf approximation with the cdw instability exhibits non - vanishing excitonic @xmath64-expectation value for all @xmath0-electron densities except the case when @xmath65 and 1 .
thus in accordance with the quantum monte - carlo studies @xcite we have found that the ferroelectric ground state with the spontaneous polarization is stabilized when the system is in the mixed valence regime and the sign of the @xmath0-electron hopping integral is opposite to the sign of the @xmath2-electron one . the fact that hf solutions can describe the existence of ferroelectric ground - state with spontaneous polarization is not surprising , since this state has been found already in the homogeneous hf solution of the conventional fkm ( @xmath66 ) in the limit of vanishing hybridization @xmath67 @xcite , even for all @xmath0-electron concentrations ( for all values of @xmath57 from the @xmath2-electron band ) .
what is however surprising is that the hf solutions with the cdw instability reproduce perfectly the ground - state phase diagram obtained by cpmc method for all examined values of @xmath0-electron hopping ( @xmath68 ) .
this is clearly demonstrated in fig . 2 , where both phase diagrams are compared .
the fact that the hf approximation with the cdw instability can describe qualitatively as well as quantitatively ground - state properties of the fkm with @xmath0-electron hopping motivated us to extend our hf study on cases that have been not described by quantum monte - carlo simulations . at first
this is the case of small @xmath0-electron hopping integrals ( @xmath69 ) that has been not considered in the original work of batista et al .
@xcite because numerical difficulties which appear in the quantum monte - carlo simulations for small @xmath23 ( the limitations in the numerical accuracy ) .
the second interesting case that we would like to study here within the hf theory is the three - dimensional case for which the numerical results are very rare due to numerical limitations on the size of clusters .
let us first discuss our two - dimensional results obtained in the limit of small values of @xmath0-electron hopping integrals . in fig .
3 we present results of detailed hf analysis performed in this limit for @xmath70 and @xmath71 .
it is seen that the non - vanishing excitonic @xmath13 expectation value persists also for small values of @xmath72 but now the inhomogeneous solution @xmath73 ( with ab - sublattice oscillations in the excitonic and charge order parameters ) is stabilized against the homogeneous one ( @xmath74 ) .
the effect is especially strong when we approach the @xmath66 limit .
this is clearly demonstrated in fig . 4 , where the complete intermediate - coupling phase diagram of the fkm with @xmath0-@xmath0-hopping is displayed .
five different phases depicted in fig . 4 as @xmath75 ( the full @xmath0 band ) , @xmath76 ( the excitonic phases ) , @xmath77 ( the cdw phase ) and @xmath78 ( the full @xmath2 band ) correspond to following hf solutions : @xmath79 the stability of different hf solutions was also checked numerically by calculating the total energy and it was found that all phases presented in the ground - state phase diagram represent the most stable hf solutions . to determine the type of transitions between different phases we have performed an exhaustive numerical study of the @xmath57 dependence of the hf order parameters ( the typical examples are shown in fig . 1 and fig . 3 ) .
at a first glance it seems that there are both first - order ( @xmath23 large ) and second - order ( @xmath23 small ) phase transitions in the extended fkm with @xmath0-@xmath0 hopping .
however , a more detailed analysis of numerical data ( with much higher resolution than used in fig . 1 and fig .
3 ) showed that the @xmath80 phase persists also for large @xmath23 , although its stability region is now considerably reduced ( see insets in fig . 1 ) .
thus there is no difference between the case of small and large values of @xmath23 . in both cases the hf order parameters change continuously indicating that the phase transitions between different phases presented in the ( @xmath57-@xmath23 ) ground - state phase diagram are of the second order .
the same calculations we have performed also in the weak coupling limit ( @xmath81 ) .
we have found that the phase diagrams obtained in the weak and intermediate coupling regime have qualitatively the same form and only one difference between them is that the ferroelectric domain ( @xmath82 ) is stabilized against remaining phases with decreasing coulomb interaction ( see fig .
of course , this fact does not imply automatically that the excitonic @xmath13 expectation value persists also for vanishing @xmath1 and that the coulomb interaction @xmath1 is not necessary for a stabilization of the ferroelectric state , what should be in a contradiction with conclusions based on the cpmc simulations .
indeed , calculations that we have performed for different values of @xmath23 at the selected @xmath0-electron density @xmath83 showed ( see fig .
5 ) that the excitonic @xmath84 expectation value is zero for @xmath85 , rapidly increases with increasing @xmath1 and tends to the saturated state for @xmath1 sufficiently large .
this confirms independently the third postulate of batista et al .
@xcite and namely , that the local coulomb interaction between the different orbitals is required in order to stabilize the ferroelectric state with the spontaneous polarization . before discussing the case of positive @xmath23
let us show explicitly the hf solution for the limit of the conventional fkm ( @xmath66 ) .
for this case we have found that @xmath86 in the @xmath87 and @xmath88 phase , while @xmath89 in the @xmath80 phase .
the last solution implies that the extitonic @xmath90-expectation value is equal to @xmath91 on the @xmath92 sublattice of the hypercubic lattice , while @xmath93 on the @xmath94 sublattice . for the symmetric case @xmath95 our solutions are fully consistent with the czycholl s ones obtained in the limit of infinite dimensions @xcite .
on the other hand both these inhomogeneous solutions fully differ from the homogeneous one @xcite that predicts a non - zero excitonic @xmath96-expectation value for all @xmath57 from the mixed valence regime with maximum of @xmath96 at @xmath95 .
similar calculations as for @xmath97 we have performed also for @xmath98 .
we have found that the ground - state phase diagram for @xmath98 has exactly the same form as for @xmath97 , however five different phases @xmath99 and @xmath88 are now characterized by : @xmath100 thus the main difference between the phase diagrams obtained for negative and positive @xmath23 is that the ferroelectric domain @xmath82 at @xmath97 is replaced by the antiferroelectric one at @xmath98 .
these two large domains are separated by a relatively narrow @xmath80 domain within which the sublattice excitonic averages ( @xmath101 ) change continuously ( see fig . 6 ) from the ferroelectric case ( @xmath102 ) to the antiferroelectric case ( @xmath103 ) .
qualitatively the same picture we have observed also for the three - dimensional case .
this is illustrated in fig .
7 , where the ground - state phase diagrams of the extended fkm are plotted for two different values of coulomb interaction ( @xmath63 and @xmath104 ) .
these results indicate that ferroelectricity remains robust with respect to the increasing dimension of the system , what should be important for an application of hf solutions on a description of real three dimensional systems . in conclusion
, we have calculated the ground - state phase diagram of the spinless fkm with @xmath0-@xmath0 hopping in the hf approximation with the cdw instability .
we have found that the hf solutions with the cdw instability reproduce perfectly the two - dimensional intermediate coupling phase diagram of the extended fkm calculated by cpmc method .
using this fact we have extended our hf study on cases that have been not described by cpmc and namely , the case of small values of @xmath0-electron hopping integrals , the case of weak coulomb interactions and the three - dimensional case .
we have found that the ferroelectric ground state with the spontaneous polarization remains stable in all examined cases .
this work was supported by the slovak grant agency for science under grant no .
2/7057/27 and the slovak apvv grant agency under grant lpp-0047 - 06 .
i would also like to acknowledge h. enarikov for a technical help during the preparation of manuscript .
note added.- _ after submitting this work we came to know about the work of schneider and czycholl @xcite who studied the extended fkm in the limit of infinite dimensions and obtained results similar to ours .
. dependence of the hf parameters @xmath105 and @xmath106 on the @xmath0-level energy @xmath57 calculated ( with step @xmath107 ) for three different values of @xmath23 ( @xmath108 ) and @xmath63 .
insets show the @xmath109 case at much higher resolution ( the numerical data have been obtained with step @xmath110 ) .
the case of @xmath111 is analogous to @xmath109 . | the hartree - fock ( hf ) approximation with the charge - density - wave ( cdw ) instability is used to study the ground - state phase diagram of the spinless falicov - kimball model ( fkm ) extended by @xmath0-@xmath0 hopping in two and three dimensions .
it is shown that the hf solutions with the cdw instability reproduce perfectly the two - dimensional intermediate coupling phase diagram of the fkm model with @xmath0-@xmath0 hopping calculated recently by constrained path monte carlo ( cpmc ) method .
using this fact we have extended our hf study on cases that have been not described by cpmc , and namely , ( i ) the case of small values of @xmath0-electron hopping integrals , ( ii ) the case of weak coulomb interactions and ( iii ) the three - dimensional case .
we have found that ferroelectricity remains robust with respect to the reducing strength of coupling ( @xmath0-electron hopping ) as well as with respect to the increasing dimension of the system .
= -1pt the falicov - kimball model ( fkm ) is a paradigmatic example of simple model to study correlation effects in interacting fermion systems on a lattice @xcite .
the model was originally proposed to describe metal - insulator transitions and has since been investigated in connection with a variety of problems such as binary alloys @xcite , the formation of ionic crystals @xcite , and ordering in mixed - valence systems @xcite . in the last few years the fkm
was extensively studied in connection with the exciting idea of electronic ferroelectricity @xcite .
the motivation for these studies comes from the pioneering work of portengen at al .
@xcite who studied the fkm with a k - dependent hybridization in the hartree - fock ( hf ) approximation and found that the coulomb interaction @xmath1 between the itinerant @xmath2-electrons and the localized @xmath0-electrons gives rise a non - vanishing excitonic @xmath3-expectation value even in the limit of vanishing hybridization @xmath4 . as an applied ( optical ) electrical field provides for excitations between d- and f - states and thus for a polarization expectation value @xmath5 , the finding of a spontaneous @xmath6 ( without hybridization or electric field ) has been interpreted as evidence for electronic ferroelectricity .
this result stimulated further theoretical studies of the model .
analytical calculations within well controlled approximation ( for @xmath1 small ) performed by czycholl @xcite in infinite dimensions did not confirm the existence of electronic ferroelectricity .
in contrast to results obtained by portengen et al .
@xcite he found that the fkm in the symmetric case ( @xmath7 ) does not allow for a ferroelectric ground state with a spontaneous polarization , i.e. , there is no nonvanishing @xmath3-expectation value in the limit of vanishing hybridization .
the same conclusion has been also obtained independently by extrapolation of small - cluster exact - diagonalization and density matrix renormalization group ( dmrg ) calculations in the one dimension for both intermediate and strong interactions @xcite . in these regions
the finite - size effects are negligible and thus the results can be satisfactory extrapolated to the thermodynamic limit .
hybridization between the itinerant @xmath2 and localized @xmath0 states , however , is not the only way to develop @xmath2-@xmath0 coherence .
recent theoretical works of batista et al .
@xcite showed that the ground state with a spontaneous electric polarization can also be induced by @xmath0-@xmath0 hopping for dimensions @xmath8 . in the strong coupling limit
this result has been proven by mapping the extended fkm into the @xmath9 spin 1/2 model with a magnetic field along the @xmath10-direction , while in the intermediate coupling regime the ferroelectric state has been identified numerically by constrained path monte carlo ( cpmc ) technique .
on the base of these results the authors postulated the following conditions that favor the formation of the electronically driven ferroelectric state : ( a ) the system must be in a mixed - valence regime and the two bands involved must have different parity .
( b ) it is best , though not necessary , if both bands have similar bandwidths .
( c ) a local coulomb repulsion ( @xmath1 ) between the different orbitals is required . in the present paper
we study the extended fkm ( the spinless fkm with @xmath0-@xmath0 hopping ) in the hf approximation with the charge - density - wave ( cdw ) instability . for reasons mentioned above
we restrict our studies on dimensions @xmath8 .
first we show that the hf solutions with the cdw instability reproduce perfectly the ground - state phase diagram obtained by the cpmc method ( @xmath11 ) for intermediate coulomb interactions @xcite .
this `` calibration '' allows us to extend calculations to the case of small values of the @xmath0-electron hopping integral @xmath12 , that has been omitted in the cpmc phase diagram for numerical problems . just in this region
we have found a new phase that corresponds to the inhomogeneous solution for @xmath13-expectation value .
this result completes the ground - state phase diagram of the two - dimensional fkm extended by @xmath0-@xmath0 hopping for intermediate couplings .
the same calculations we have performed also in the weak coupling limit ( for @xmath11 ) as well as in three dimensions .
we have found that the ferroelectricity remains robust with respect to the reducing strength of the coupling as well as with respect to the increasing dimension of the system . |
in wavelet analysis , we often use translation , dilation , and modulation of functions . for a function @xmath0 , throughout the paper we shall use the following notation @xmath1 where @xmath2 denotes the imaginary unit . in this paper we shall use @xmath3 as a dilation factor . in applications , @xmath4 is often taken to be a positive integer greater than one , in particular , the simplest case @xmath5 is often used .
classical wavelets are often defined and studied in the time / space domain with the generating wavelet functions belonging to the square integrable function space @xmath6 . for @xmath7 and for a subset @xmath8 of square integrable functions in @xmath6 , linked to discretization of a continuous wavelet transform ( see @xcite ) , the following homogeneous wavelet system @xmath9 is generated by the translation and dilation of the wavelet functions in @xmath8 and has been extensively studied in the function space @xmath6 in the literature of wavelet analysis . to mention only a few references here ,
see @xcite@xcite . in this paper , however , we shall see that it is more natural to study a nonhomogeneous wavelet system in the frequency domain .
it is important to point out here that the elements in a set @xmath10 of this paper are not necessarily distinct and @xmath11 in a summation means that @xmath12 visits every element ( with multiplicity ) in @xmath10 once and only once .
for example , for @xmath13 , all the functions @xmath14 are not necessarily distinct and @xmath15 in means @xmath16 .
most known classical homogeneous wavelet systems @xmath17 in the literature are often derived from scalar refinable functions @xmath18 or from refinable function vectors @xmath19 in @xmath6 ( @xcite ) .
let us recall the definition of a scalar refinable function here .
a function or distribution @xmath18 on @xmath20 is said to be _ refinable _ ( or @xmath4-refinable ) if there exists a sequence @xmath21 of complex numbers , called the _ refinement mask _ or the _ low - pass filter _ for the scalar refinable function @xmath18 , such that @xmath22 with the above series converging in a proper sense , e.g. , in @xmath6 .
wavelet functions in the generating set @xmath8 of a homogeneous wavelet system @xmath17 are often derived from the refinable function @xmath18 by @xmath23 where @xmath24 are sequences on @xmath25 , called _ wavelet masks _ or _ high - pass filters_. for the infinite series in and to make sense , one often imposes some decay condition on the refinable function @xmath18 and wavelet filters @xmath26 so that all the infinite series in and are well - defined in a proper sense .
nevertheless , even for the simplest case of a compactly supported scalar refinable function ( or distribution ) @xmath18 with a finitely supported mask @xmath27 , the associated refinable function @xmath18 with mask @xmath27 does not always belong to @xmath6 .
in fact , it is far from trivial to check whether @xmath28 in terms of its mask @xmath27 , see @xcite and references therein for detail .
one of the motivations of this paper is to study wavelets and framelets without such stringent conditions on either the generating wavelet functions @xmath29 or their wavelet filters @xmath26 for @xmath15 . for @xmath30
, the fourier transform used in this paper is defined to be @xmath31 , @xmath32 and can be naturally extended to square integrable functions and tempered distributions . under certain assumptions , taking fourier
transform on both sides of and , one can easily rewrite and in the frequency domain as follows : @xmath33 and @xmath34 provided that all the @xmath35-periodic ( lebesgue ) measurable functions @xmath36 and similarly @xmath37 are properly defined . in the following , we shall see that it is often more convenient to work with and in the frequency domain rather than and in the time / space domain .
if there exist positive real numbers @xmath38 and @xmath39 such that the @xmath35-periodic measurable function @xmath40 satisfies @xmath41 for almost every @xmath42 $ ] ( this condition is automatically satisfied with @xmath43 if @xmath40 is a @xmath35-periodic trigonometric polynomial with @xmath44 ) , for a dilation factor @xmath4 such that @xmath45 , then it is easy to see ( also c.f .
section 3 ) that one can define a measurable function @xmath46 such that @xmath47 regardless of whether @xmath46 in is a square integrable function or not , @xmath46 is a well - defined measurable function obviously satisfying the frequency - based refinement equation with @xmath48 being replaced by @xmath46 .
the function @xmath46 is called the ( frequency - based ) _ standard refinable function _ with mask @xmath40 and dilation @xmath4 .
all the wavelet functions @xmath49 in with @xmath50 are also well - defined measurable functions provided that all @xmath51 are measurable .
this motivates us to study refinable functions and wavelets in the frequency domain using and so that we can avoid some technical issues such as the convergence of the infinite series in and as well as membership in @xmath6 of the generating refinable function @xmath18 and the generating wavelet functions @xmath52 . for wavelets derived from refinable functions or refinable function vectors ,
one of the most important key features of wavelets and framelets is its associated fast wavelet transform , which is based on the following nonhomogeneous wavelet system : @xmath53 where @xmath54 is an integer , representing the coarsest decomposition ( or scale ) level of its fast wavelet transform .
in fact , a one - level fast wavelet transform is just a transform between two sets of wavelet coefficients of a given function represented under two nonhomogeneous wavelet systems at two consecutive scale levels .
naturally , for a multi - level wavelet transform , there is a underlying sequence of nonhomogeneous wavelet systems at all scale levels , instead of just one single wavelet system .
for a given @xmath55 at some scale @xmath56 , we shall see in this paper that via the dilation operation it naturally produces a sequence of nonhomogeneous wavelet systems @xmath57 for all integers @xmath54 with almost all properties preserved .
consequently , it often suffices to study only one nonhomogeneous wavelet system instead of a sequence of them .
this desirable property of nonhomogeneous wavelet systems is not shared by homogeneous wavelet systems .
furthermore , as @xmath58 , the limit of the sequence @xmath59 will naturally lead to a homogeneous wavelet system @xmath17 .
hence , in certain sense , a homogeneous wavelet system could be regarded as the limit system of a sequence of nonhomogeneous wavelet systems .
see section 3 for more detail . for a homogeneous wavelet system @xmath17 that is derived from a refinable function or a refinable function vector , due to the absence of a refinable function @xmath18 in the system
, the homogeneous wavelet system @xmath17 does not automatically correspond to a fast wavelet transform without ambiguity .
in fact , the wavelet functions in @xmath8 of @xmath17 could be derived from many other ( equivalent ) refinable functions , which correspond to different fast wavelet transforms with different sets of wavelet filters .
more precisely , for a @xmath35-periodic measurable function @xmath60 such that @xmath61 for almost every @xmath32 , define @xmath62 , then it is evident that @xmath63 is also refinable and satisfies @xmath64 \hat { \mathring{\phi}}(\xi ) \quad \hbox{and}\quad \hat{\psi}(\df\xi)=[\wh{b^\psi}(\xi)/\theta(\xi ) ] \hat{\mathring{\phi}}(\xi ) , \qquad a.e.\ ; \xi\in \r , \ ; \psi\in \psi.\ ] ] such a change of generators from a refinable function @xmath18 to another equivalent refinable function @xmath65 is in fact the key idea in the oblique extension principle ( oep ) in @xcite ( also see @xcite ) to construct compactly supported homogeneous wavelet frames @xmath17 in @xmath6 with high vanishing moments derived from refinable functions and refinable function vectors .
see @xcite for a detailed discussion on a fast wavelet transform based on a homogeneous wavelet system obtained via oep from refinable function vectors .
the effect of the change of generators on its fast wavelet transform is addressed in @xcite .
as we shall see in sections 3 and 4 , nonhomogeneous wavelet systems are closely related to nonstationary wavelets ( see @xcite ) and are naturally employed in a pair of nonhomogeneous dual wavelet frames in a pair of dual sobolev spaces introduced in @xcite . due to these and other considerations
, it seems more natural and more important for us to study nonhomogeneous wavelet systems @xmath57 rather than the extensively studied homogeneous wavelet systems @xmath17 .
this allows us to understand better many aspects of wavelet theory such as a wavelet filter bank induced by oep and its associated wavelets in the function setting without a priori condition on the generating wavelet functions . following the standard notation , we denote by @xmath66 the linear space of all compactly supported @xmath67 ( test ) functions with the usual topology , and @xmath68 denotes the linear space of all distributions ,
that is , @xmath68 is the dual space of @xmath66 . by duality , the definition in for translation
, dilation and modulation can be easily generalized from functions to distributions .
moreover , when @xmath69 in is a square integrable function or more generally a tempered distribution , for @xmath70 , we have @xmath71 in this paper , we shall use boldface letters to denote functions / distributions ( e.g. , @xmath72 ) or sets of functions / distributions ( e.g. , @xmath73 ) in the frequency domain .
let @xmath74 and @xmath75 be two sets of distributions on @xmath20 .
for an integer @xmath54 and @xmath76 , we define _ a frequency - based nonhomogeneous wavelet system @xmath77 _ to be @xmath78 similarly , _ a frequency - based homogeneous wavelet system _ is defined as follows : @xmath79 by , it is straightforward to see that under the fourier transform , the images of @xmath57 and @xmath17 with @xmath80 are simply @xmath77 and @xmath81 , respectively , where @xmath82 and @xmath83 . for @xmath84 , by @xmath85
we denote the linear space of all measurable functions @xmath69 such that @xmath86 for every compact subset @xmath87 of @xmath20 , with the usual modification for @xmath88 saying that @xmath69 is essentially bounded over @xmath87 .
note that @xmath89 is just the set of all measurable functions that can be globally identified as distributions .
so , @xmath89 is the most natural space for us to study wavelets and framelets in the distribution space . for @xmath90 ,
it is evident that @xmath91 .
however , a distribution @xmath92 may not be a tempered distribution and therefore , fourier transform may not be applied so that @xmath93 holds for some tempered distribution @xmath94 . under the setting of tempered distributions on which the fourier transform can apply ,
although all the definitions and results of this paper in the frequency domain could be equivalently translated into the time / space domain by the inverse fourier transform on tempered distributions , to avoid notational complexity and to avoid the a priori underlying assumption that @xmath69 is a tempered distribution if the notion @xmath95 is used , it seems very natural and convenient for us to work in the frequency domain in this paper . for @xmath96 and @xmath97 , we shall use the following paring @xmath98 when @xmath96 and @xmath99 , the duality pairings @xmath100 and @xmath101 are understood similarly as @xmath102 .
now we are ready to introduce the key notion in this paper .
let @xmath103 be subsets of @xmath68 .
let @xmath104 and @xmath3 , we say that the pair @xmath105 , where @xmath77 is defined in , forms _ a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 _ if the following identity holds @xmath106 where the infinite series in converge in the following sense : 1 . for every @xmath107 , the following series @xmath108 converge absolutely for all integers @xmath109 , @xmath110 , and @xmath111 .
2 . for every @xmath107 , the following limit exists and @xmath112
as we shall discuss in section 3 , the above introduced notion enables us to completely separate the perfect reconstruction property in from its stability property in function spaces .
since the test function space @xmath66 is dense in many function spaces , one could extend the perfect reconstruction property ( or `` wavelet expansion '' ) in to other function spaces , provided that the involved wavelet systems have stability in these function spaces .
let us give a simple example here to illustrate this connection for the particular function space @xmath6 .
let @xmath74 and @xmath75 in be two subsets of distributions in @xmath68 .
we shall see in section 3 that @xmath113 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space , if and only if , @xmath73 are subsets of @xmath6 and latexmath:[\[\label{f : twf } \sum_{\ell=1}^{\mphi } \sum_{k\in \z } |\la \ff , \fphi^{\ell}_{\df^{-j};0,k}\ra|^2 + \sum_{j = j}^\infty \sum_{\ell=1}^{\mpsi } \sum_{k\in \z } moreover , when @xmath45 , as a direct consequence of , one automatically has @xmath115 ( that is , multiply every element in @xmath77 by the factor @xmath116 ) is an orthonormal basis of @xmath6 .
denote @xmath117 . without assuming in advance that @xmath118 is a refinable function vector and all @xmath119
are derived from @xmath118 , we can deduce that @xmath118 must be a refinable function vector and all @xmath119 must be derived from @xmath118 via similar relations as in and .
see section 3 for more detail . to have some rough ideas about our results on nonhomogeneous wavelet systems , here we present two typical results .
the following result is a special case of theorem [ thm : main:1:general ] .
[ thm : main:1 ] let @xmath4 be an integer such that @xmath45 .
let @xmath120 in be subsets of @xmath121 .
then @xmath105 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 for some integer @xmath54 ( or for all integers @xmath54 ) , if and only if , the following three statements hold 1 . for all integers @xmath122 , @xmath123 2 . for all integers
@xmath124 $ ] , @xmath125 3 .
the following identity holds in the sense of distributions : @xmath126 more precisely , @xmath127 for all @xmath96 .
in the following , we make some remarks about theorem [ thm : main:1 ] .
we assumed in theorem [ thm : main:1 ] that all the generating functions in @xmath120 are from the space @xmath121 .
note that @xmath121 includes the fourier transforms of all compactly supported distributions and of all elements in all sobolev spaces .
this assumption on membership in @xmath121 can be weakened and is only used to guarantee the absolute convergence of the infinite series in . see the remark after lemma [ lem : converg ] in section 2 for more detail on this natural assumption . if we assume additionally that @xmath128 for all @xmath110 and if holds for almost every @xmath32 , by lebesgue dominated convergence theorem , then holds in the sense of distributions . if all elements in @xmath120 are essentially nonnegative measurable functions , then it is not difficult to verify that the conditions in items ( i ) and ( ii ) of theorem [ thm : main:1 ] are equivalent to the following simple conditions : @xmath129 and @xmath130 as we shall see in section 2 , items ( i ) and ( ii ) of theorem [ thm : main:1 ] correspond to a natural multiresolution - like structure , which is closely linked to a fast wavelet transform .
the condition in item ( iii ) of theorem [ thm : main:1 ] is a natural normalization condition which is related to . comparing with the characterization of a pair of homogeneous dual wavelet frames in the space @xmath6 or a homogeneous orthonormal wavelet basis in @xmath6 ( e.g. , see @xcite ) , theorem [ thm : main:1 ] has several interesting features .
firstly , the characterization in items ( i)(iii ) of theorem [ thm : main:1 ] does not involve any infinite series or infinite sums ; this is in sharp contrast to the homogeneous setting in @xmath6 .
secondly , as we shall see in section 2 , all the involved infinite sums in the proof of theorem [ thm : main:1 ] are in fact finite sums .
this allows us to easily generalize theorem [ thm : main:1 ] to any real dilation factors and to nonstationary wavelets , see sections 2 and 4 for detail .
thirdly , we do not require any stability ( bessel ) property of the wavelet systems , while the homogeneous setting in @xmath6 needs the stability property to guarantee the convergence of the involved infinite series .
fourthly , we do not require in theorem [ thm : main:1 ] that the generating wavelet functions possess any order of vanishing moments or smoothness , while all the generating wavelet functions in the homogeneous setting require at least one vanishing moment .
lastly , from a pair of nonhomogeneous dual wavelet frames in @xmath6 , we shall see in section 3 that one can always derive an associated pair of homogeneous dual wavelet frames in @xmath6 .
in fact , most homogeneous wavelet systems in the literature are derived in such a way .
we mention that weak convergence of wavelet expansions has been characterized in @xcite for homogeneous wavelet systems .
similar weak convergence of wavelet expansions that are related to also appeared in the study of homogeneous dual wavelet frames in @xmath6 and their frame approximation properties , for example , see @xcite .
we also point out that the approach in this paper can be extended to frequency - based homogeneous wavelet systems in the distribution space @xmath131 .
the following result generalizes the oblique extension principle ( oep ) and naturally connects a wavelet filter bank with a pair of frequency - based nonhomogeneous dual wavelet frames in @xmath68 .
[ thm : main:2 ] let @xmath4 be an integer such that @xmath45 .
let @xmath132 and @xmath133 , @xmath134 be @xmath35-periodic measurable functions on @xmath20 .
suppose that there are measurable functions @xmath135 satisfying @xmath136 define @xmath120 as in with @xmath137 and @xmath138 assume that all the elements in @xmath120 belong to @xmath121 .
then @xmath139 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 for some integer @xmath54 ( or for all integers @xmath54 ) , if and only if , @xmath140 with @xmath141 and the following fundamental identities are satisfied : @xmath142 and @xmath143 for all @xmath144 , where @xmath145 and @xmath146 in particular , if all @xmath147 are @xmath35-periodic measurable functions in @xmath121 and if there exist positive real numbers @xmath38 and @xmath39 such that latexmath:[\[\label{lip : ata } then the frequency - based standard refinable measurable functions @xmath135 with masks @xmath149 and the dilation factor @xmath4 , which are defined by @xmath150 are well - defined for almost every @xmath32 and in fact @xmath151 .
then all elements in @xmath120 belong to @xmath121 .
moreover , @xmath105 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 for some integer @xmath54 ( or for all integers @xmath54 ) , if and only if , the identities and are satisfied for all @xmath144 , and @xmath152 in the sense of distributions .
note that is automatically satisfied with @xmath43 if @xmath153 and @xmath154 are @xmath35-periodic trigonometric polynomials with @xmath155 . a similar result to theorem [ thm : main:2 ] also holds when @xmath156 and @xmath157 are refinable measurable function vectors .
the identities in and with @xmath5 and @xmath158 are called the oblique extension principle in @xcite , provided that all elements in @xmath120 belong to @xmath6 and satisfy some technical conditions to guarantee the bessel ( stability ) property of the homogeneous wavelet systems @xmath81 and @xmath159 in the space @xmath6 ( see @xcite ) .
in contrast , our results here generally do not require any a priori condition on the generating wavelet functions and provide a natural explanation for the connection between the perfect reconstruction property induced by oep in and in the discrete filter bank setting to wavelets and framelets in the function setting .
the structure of the paper is as follows . in order to prove theorems [ thm : main:1 ] and [ thm : main:2 ]
, we shall introduce some auxiliary results in section 2 .
in particular , we shall provide sufficient conditions in section 2 for the absolute convergence of the infinite series in
. then we shall prove theorems [ thm : main:1 ] and [ thm : main:2 ] in section 2 .
to explain in more detail about our motivation and importance for studying frequency - based nonhomogeneous wavelet systems , we shall discuss in section 3 nonhomogeneous wavelet systems in various function spaces such as @xmath6 and sobolev spaces , as initiated in @xcite .
we shall see in section 3 that under the stability property , a pair of frequency - based nonhomogeneous dual wavelet frames can be naturally extended from the distribution space to a pair of dual function spaces . in section 3
, we shall also explore the connections between nonhomogeneous and homogeneous wavelet systems in the space @xmath6 . to illustrate the flexibility and generality of the approach in this paper , we further study nonstationary wavelets which are useful in many applications , since nonstationary wavelet filter banks can be implemented in almost the same way and efficiency as a traditional fast wavelet transform .
however , except a few special cases as discussed in @xcite , only few theoretical results on nonstationary wavelets are available in the literature , probably partially due to the difficulty in guaranteeing the membership of the associated refinable functions in @xmath6 and in establishing the stability property of the nonstationary wavelet systems in @xmath6 . in section 4
, we present a complete characterization of a pair of frequency - based nonstationary dual wavelet frames in the distribution space . though the statements and notation in section 4 on nonstationary wavelets seem a little bit more complicated comparing with the stationary case in sections 13 , it is worth our effort to provide a better picture to understand nonstationary wavelets , since there are few theoretical results on this topic in the literature . to understand and study wavelet systems in various function spaces , it is our opinion that there are two key fundamental ingredients to be considered .
one ingredient is the notion investigated in this paper of a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space which enables us to completely separate its perfect reconstruction property from its stability property in function spaces .
the other ingredient is the stability issue of nonhomogeneous wavelet systems in function spaces which we did nt discuss in this paper but shall be addressed elsewhere .
in this section , we study pairs of frequency - based nonhomogeneous dual wavelet frames in the distribution space . to prove theorems [ thm : main:1 ] and [ thm : main:2 ] , we first present some sufficient conditions for the absolute convergence of the infinite series in . for @xmath160 , by @xmath161
we denote the set of all @xmath35-periodic measurable functions @xmath69 such that @xmath162 ( with the usual modification for @xmath88 ) . by the following result
, we always have the absolute convergence of the infinite series in provided that all the frequency - based wavelet functions are from the space @xmath121 .
[ lem : converg ] let @xmath163 be a nonzero real number and let @xmath164 .
then for all @xmath107 , @xmath165 with the series on the left - hand side converging absolutely . note that the infinite sum on the right - hand side of is in fact finite .
by @xmath166 we denote the linear space of all compactly supported measurable functions in @xmath167 .
note that @xmath168 .
more generally , we prove for @xmath169 .
denote @xmath170 now we show that @xmath171 are well - defined functions in @xmath172 . in fact , since @xmath173 , @xmath174 has compact support and therefore , @xmath174 is essentially supported inside @xmath175 $ ] for some @xmath176 with @xmath177 depending on @xmath174 .
now it is easy to see that @xmath178.\ ] ] since @xmath179 and @xmath173 , we see that @xmath180 for every @xmath122 . therefore , @xmath181 is a well - defined @xmath35-periodic function in @xmath172 .
similarly , we have @xmath182 .
note that @xmath183 and @xmath184 . since @xmath185 ,
by the parseval identity , we have @xmath186 and @xmath187 with the series on the left - hand side converging absolutely . by the finite sum in
, we have @xmath188 which completes the proof . the condition @xmath164 in lemma [ lem : converg ] is only used in this paper to guarantee the identity .
as long as holds for @xmath107 and the frequency - based wavelet functions belong to @xmath89 , all the claims in this paper still hold .
note that @xmath89 is just the set of all measurable functions that can be globally identified as distributions .
so , @xmath89 is the most natural and weakest space for us to study wavelets and framelets in the distribution space .
the condition @xmath189 in lemma [ lem : converg ] could be replaced by other conditions .
for example , for every positive integer @xmath190 , if there exist positive numbers @xmath191 and @xmath192 such that @xmath193 and @xmath194 for all @xmath195 $ ] , then it is not difficult to check by that @xmath181 with @xmath96 is a @xmath35-periodic lipschitz function with some lipschitz exponent @xmath196 . by bernstein theorem , @xmath181 has an absolutely convergent fourier series .
now for any @xmath197 , it is easy to prove that indeed holds for all @xmath107 .
other assumptions could be used to guarantee .
but @xmath121 is a large space containing the fourier transforms of all compactly supported distributions and of all elements in all sobolev spaces . for simplicity of presentation , we shall stick to the space @xmath121 for our discussion of frequency - based wavelets and framelets .
[ lem : to1 ] let @xmath198 be a sequence of nonzero real numbers such that @xmath199 .
let @xmath200 and @xmath201 be elements in @xmath121 with @xmath202 and @xmath109 . then @xmath203 if and only if , @xmath204 by lemma [ lem : converg ] , we have @xmath205 by @xmath107 , @xmath174 and @xmath206 are compactly supported . since @xmath199
, there exists an integer @xmath207 such that @xmath208 for all @xmath32 , @xmath209 , and @xmath210 .
that is , for @xmath211 , becomes @xmath212 if holds in the sense of distributions , then it follows directly from that holds .
conversely , we can take @xmath213 such that @xmath206 takes value one on the support of @xmath174 , now it follows from and that @xmath214 hence , holds in the sense of distributions . in the next auxiliary result
, we shall study a multiresolution - like structure .
more precisely , we have the following result . [
lem : twolevel ] let @xmath163 be a nonzero real number .
let @xmath215 , @xmath119 , @xmath216 and @xmath217 , @xmath218 be elements in @xmath121
. then @xmath219 if and only if @xmath220 , \label{i : eq1}\\ & i_\fphi^{k}(\xi)+i^{k}_{\fpsi}(\xi)=0 , \qquad a.e.\ , \xi\in \r,\ ; \forall\ ; k\in \z \bs [ \gl^{-1 } \z ] , \label{i : eq2}\\ & i^{\gl k}_\feta(\xi)=0 , \qquad a.e.\ , \xi\in \r,\ ; \forall \ ; k\in [ \gl^{-1 } \z ] \bs \z , \label{i : eq3}\end{aligned}\ ] ] where @xmath221 and @xmath222 , @xmath223 , and @xmath224 by lemma [ lem : converg ] , all the infinite series in converge absolutely and is equivalent to @xmath225 } \ff(\xi ) \ol{\fg(\xi+2\pi k ) } i^{\gl k}_\feta(\gl \xi ) d\xi,\ ] ] which can be easily rewritten as @xmath226 } \ff(\xi ) \ol{\fg(\xi+2\pi k ) } \big(i^{k}_\fphi(\xi ) + i^{k}_\fpsi(\xi)-i^{\gl k}_\feta(\gl \xi ) \big ) d\xi\\ & + \int_\r \sum_{k\in \z \bs [ \gl^{-1 } \z ] } \ff(\xi ) \ol{\fg(\xi+2\pi k ) } \big(i^{k}_\fphi(\xi ) + i^{k}_\fpsi(\xi)\big ) d\xi = \int_\r \sum_{k\in [ \gl^{-1 } \z]\bs \z } \ff(\xi ) \ol{\fg(\xi+2\pi k ) } i^{\gl k}_\feta(\gl \xi ) d\xi . \end{split}\ ] ] sufficiency . if , , and are satisfied , then it is obvious that is true and therefore , holds . necessity
. denote @xmath227 $ ] .
for a point @xmath228 , we define @xmath229 . by , holds .
let @xmath230 $ ] and @xmath231 be temporarily fixed .
then it is easy to check that @xmath232 .
consider all @xmath233 such that the support of @xmath174 is contained inside @xmath234 and the support of @xmath206 is contained inside @xmath235 .
then it is not difficult to verify that @xmath236 from which we see that becomes @xmath237 for all @xmath107 such that @xmath238 and @xmath239 . from
, we must have @xmath240 for almost every @xmath241 .
thus , must be true . and
can be proved by the same argument .
now we have the generalized version of theorem [ thm : main:1 ] with a general real dilation factor @xmath4 .
[ thm : main:1:general ] let @xmath4 be a real number such that @xmath45 .
let @xmath120 in be subsets of @xmath121 .
then @xmath242 , where @xmath243 is defined in , forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 for some integer @xmath54 ( or for all integers @xmath54 ) , if and only if , @xmath244 and @xmath245 , \end{split}\ ] ] @xmath246 , \label{i : eq2:sp}\\ & \sum_{\ell=1}^\mphi \ol{\fphi^\ell(\xi ) } \tilde \fphi^\ell(\xi+2\pi \df^{-1 } k)=0 , \qquad a.e.\ , \xi\in \r,\ ; \forall \ ; k\in [ \df \z ] \bs \z .
\label{i : eq3:sp}\end{aligned}\ ] ] by the following simple observation , we have @xmath247 now it is straightforward to see that for all @xmath107 , @xmath248 if and only if , for all @xmath107 , @xmath249 sufficiency . by lemma [ lem : twolevel ] with @xmath250 , we see that holds and therefore , holds for all @xmath107 and @xmath251 . for @xmath252 , we define @xmath253 now by , we can easily deduce that @xmath254 by lemma [ lem : to1 ] , it follows from that @xmath255 for all @xmath107 .
hence , @xmath256 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 .
necessity . by
, we can easily deduce that @xmath105 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 for some integer @xmath54 if and only if it is true for all integers @xmath54 . considering the difference between two consecutive integers @xmath54 and @xmath257
, we see that must hold and therefore , holds . now by lemma [ lem : twolevel ] , , , and hold with @xmath258 and @xmath259 , or equivalently , , , and hold . since holds , we deduce that holds . by lemma [ lem : to1 ]
, it follows from our assumption @xmath260 that must hold .
we point out that theorem [ thm : main:1:general ] and the approach in this paper can be extended to frequency - based homogeneous wavelet systems in the distribution space @xmath131 , which is the dual space of the test function space @xmath261 consisting of all compactly supported @xmath67 functions whose supports are contained inside @xmath262 .
we shall address this issue elsewhere .
now we are ready to prove theorems [ thm : main:1 ] and [ thm : main:2 ] . by theorem [ thm : main:1:general ] with @xmath4 being an integer , it suffices to show that and are equivalent to the three conditions , , and .
note that @xmath263 \bs \z$ ] is the empty set .
so , with @xmath4 being an integer is automatically true .
it is evident that is equivalent to . since @xmath4 is an integer , we have @xmath264=\df \z$ ] . now it is also easy to see that is equivalent to .
this completes the proof .
we use theorem [ thm : main:1 ] to prove theorem [ thm : main:2 ] as follows : we prove the first part of theorem [ thm : main:2 ] first . by and
, we have @xmath265 and similarly by , @xmath266 for all integers @xmath122 .
now is equivalent to @xmath267 for all @xmath122 .
now it is not difficult to deduce that is equivalent to .
note that any @xmath124 $ ] can be uniquely written as @xmath268 for @xmath269 and @xmath122 .
replacing @xmath270 in by @xmath271 , by the same argument , we see that is equivalent to @xmath272 for all @xmath122 and @xmath144 .
now it is not difficult to deduce that is equivalent to with @xmath144 .
we now prove the second part of theorem [ thm : main:2 ] .
note that implies that @xmath273 and @xmath274 for all @xmath275 $ ] .
it is a standard argument to show that both @xmath46 and @xmath276 in are well - defined measurable functions in @xmath277 and @xmath278 in fact , by the same argument as in ( * ? ? ?
* page 93 ) or ( * ? ? ?
* page 932 ) , also implies that for any @xmath279 , there exists @xmath280 such that @xmath281.\ ] ] see the proof of theorem [ thm : mra : ndwf ] in section 4 for more detail on proving and .
take @xmath282 in . by the definition of @xmath283 in
, we conclude that @xmath284 for almost every @xmath285 $ ] and for all @xmath286 . consequently , by the generalized lebesgue dominated convergence theorem and using and , we can conclude that holds in the sense of distributions if and only if @xmath287 holds in the sense of distributions . since @xmath288 and @xmath289 , by the definition in and ,
it is evident that all the measurable functions in @xmath120 belong to the desired space @xmath121 . by what has been proved for the first part of theorem [ thm : main:2 ] , the claim in the second part of theorem [ thm : main:2 ] holds
in this section , we shall discuss nonhomogeneous wavelet systems in the space @xmath6 and other function spaces .
we shall see that a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space plays a basic role in the study of wavelets and framelets in various function spaces .
we shall also discuss the connections between nonhomogeneous and homogeneous wavelet systems in @xmath6 .
let us recall some necessary definitions first . for @xmath290 ,
we denote by @xmath291 the sobolev space consisting of all tempered distributions @xmath69 such that @xmath292 note that @xmath293 is the dual space of @xmath294 , since @xmath295 can be regarded as a continuous linear functional on @xmath294 in the sense of @xmath296 for @xmath297 .
let @xmath4 be a nonzero real number and @xmath298 be a nonhomogeneous wavelet system defined in .
we define a normed sequence space @xmath299 , indexed by the elements of @xmath300 , with weighted norm as follows : @xmath301 for @xmath302 , we say that @xmath298 has _ stability in @xmath291 _ with respect to the normed sequence space @xmath299 if there exist positive constants @xmath303 and @xmath304 such that @xmath305 or more explicitly , @xmath306 note that @xmath291 is a hilbert space under the inner product @xmath307 it was shown in ( * ? ? ? * proposition 2.1 ) that holds for all @xmath308 , if and only if , @xmath309 in other words , that @xmath300 has stability in @xmath291 with respect to @xmath299 is equivalent to saying that after a proper renormalization of @xmath300 , @xmath310 is a frame in the hilbert space @xmath291 . for @xmath311 and @xmath45 , in the following we show that @xmath312 for some positive constants @xmath313 and @xmath314 depending on @xmath315 and @xmath94 . for @xmath316 , to prove , we further assume that @xmath317 which is also a necessary condition for to hold .
we now prove .
in fact , @xmath318 since @xmath45 , it is easy to deduce that @xmath319 for @xmath320 , noting that there exists @xmath279 such that @xmath321 we deduce from and that holds with @xmath322 and @xmath323 . for @xmath316 , by our assumption in , there exists @xmath279 such that @xmath324 now we can easily deduce from and that holds with @xmath325 and @xmath326 . using and applying fatou s lemma to with @xmath327 ,
it is easy to see that is a necessary condition for to hold .
hence , the weights @xmath328 in the definition of @xmath329 is chosen to be the normalization factors of the elements @xmath330 in the space @xmath291 .
the stability property in , which is equivalent to the frame property in , characterizes the sobolev space @xmath331 .
we point out that the characterization of homogeneous besov spaces by homogeneous framelets in @xcite uses nonlinear approximation and is quite different in nature to and .
now we are ready to recall the definition in @xcite on a pair of nonhomogeneous dual wavelet frames in a pair of dual sobolev spaces @xmath332 .
let @xmath4 be the dilation factor .
let @xmath333 be subsets of tempered distributions .
we say that the pair @xmath334 , or more precisely @xmath335 , forms _ a pair of nonhomogeneous dual wavelet frames in a pair of dual sobolev spaces @xmath332 _ if 1 .
@xmath336 and @xmath337 has stability in @xmath291 with respect to @xmath299 .
that is , @xmath338 is a frame in @xmath291 ; 2 .
@xmath339 and @xmath340 has stability in @xmath331 with respect to @xmath341 .
that is , @xmath342 is a frame in @xmath331 ; 3 . for all @xmath343 and @xmath344 , the following identity holds @xmath345 with the series on the right - hand side converging absolutely .
the above definition is introduced in @xcite . when @xmath346 , since @xmath347 , it is easy to see that the above definition of a pair of nonhomogeneous dual wavelet frames in @xmath348 becomes the definition of a pair of nonhomogeneous dual wavelet frames in @xmath6 .
suppose that the pair @xmath334 forms a pair of nonhomogeneous dual wavelet frames in @xmath332 .
using and the upper stability ( that is , the right - hand inequality in holds ) of the two nonhomogeneous wavelet systems , it is not difficult to see that we have the following representations : @xmath349 with the series converging unconditionally in the space @xmath331 , and similarly , @xmath350 with the series converging unconditionally in the space @xmath291 . see @xcite for more detail .
let @xmath351 be a normed function space and @xmath352 be its dual .
for example , for besov spaces @xmath353 , @xmath354 , where @xmath355 and @xmath356 with @xmath357 .
similarly , for triebel - lizorkin spaces @xmath358 , @xmath359 . replacing @xmath291 and @xmath331 by @xmath351 and @xmath352 , respectively , the notion of a pair of nonhomogeneous dual wavelet frames can be generalized from a pair of dual sobolev spaces @xmath332 to a pair of dual function spaces @xmath360 by a proper choice of some normed sequence spaces @xmath361 and @xmath362 .
we shall not further address this issue in this paper . in the rest of this section
, we shall use the following notation @xmath363 by the following result , we see that the notion of a frequency - based nonhomogeneous dual wavelet frames in the distribution space plays a basic role in the study of pairs of nonhomogeneous dual wavelet frames in a pair of dual function spaces .
[ thm : sobolev ] let @xmath290 and @xmath4 be a nonzero real number .
let @xmath364 in be subsets of tempered distributions .
define @xmath365 then @xmath366 forms a pair of nonhomogeneous dual wavelet frames in the pair of dual sobolev spaces @xmath332 , if and only if , 1 .
there exists a positive constant @xmath39 such that @xmath367 and @xmath368 2 .
the pair @xmath105 , which is the image of the pair @xmath366 under the fourier transform , forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 . for @xmath369 and @xmath370 with @xmath371 ,
the relation @xmath372 holds .
since @xmath373 is contained in both @xmath291 and @xmath331 ( or equivalently , @xmath374 and @xmath375 ) , the necessity part is evident .
hence , it suffices to prove the sufficiency part . by , for all @xmath96 and @xmath376 , @xmath377 using , we now prove that @xmath378 . by , for @xmath379
, we have @xmath380 for all @xmath96 .
therefore , @xmath381 can be extended into a continuous linear functional on @xmath382 . since @xmath383 is the dual space of @xmath382 , there exists @xmath384 such that @xmath385 .
in particular , @xmath386 for all @xmath96 .
since @xmath387 is a distribution , we must have @xmath388 in the sense of distributions . by @xmath389
, we deduce that @xmath390 and therefor , @xmath391 .
consequently , by , we proved that @xmath378 .
in other words , we proved that @xmath336 .
since @xmath66 is dense in @xmath382 and since all @xmath392 and @xmath393 are continuous linear functionals on @xmath382 , we see that holds for all @xmath394 and all @xmath376 . letting @xmath395 in and noting @xmath372 for all @xmath308 and @xmath344
, we conclude that @xmath336 and @xmath396 similarly , we can show that implies @xmath339 and @xmath397 define two operators @xmath398 and @xmath399 by @xmath400 then and are equivalent to saying that the operators @xmath401 and @xmath402 are well - defined bounded linear operators , more precisely , @xmath403 by cauchy - schwarz inequality , using and , we see that for @xmath308 and @xmath344 , the series on the right - hand side of converges absolutely .
now by assumption in item ( ii ) , we see that holds for all @xmath404 . in other words , for @xmath404 , @xmath405 note that @xmath373 is dense in both @xmath331 and @xmath291 .
now we use a standard argument to show that holds for all @xmath308 and @xmath344 . for @xmath308 and @xmath344 , there exist two sequences @xmath406 and @xmath407 in @xmath373 such that @xmath408 observe that @xmath409 by @xmath410 and , we have @xmath411 .
therefore , we have @xmath412 now by and the triangle inequality , @xmath413 where @xmath414 . by , the right - hand side of the above inequality goes to @xmath415 as @xmath416 .
consequently , holds for all @xmath308 and @xmath344 .
that is , and item ( 3 ) have been verified . to prove item ( 1 ) , by and , we have @xmath417 that is , for @xmath308 ,
the following inequality holds : @xmath418 where in the last step we used the fact that @xmath291 is the dual space of @xmath331 .
hence , item ( 1 ) holds .
item ( 2 ) can be proved similarly .
this completes the proof of the sufficiency part .
next , we discuss connections between nonhomogeneous and homogeneous wavelet systems in the particular function space @xmath6 . to do so ,
we need an auxiliary result , which is essentially known in the literature , e.g. , see ( * ? ? ?
* page 28 ) . for completeness
, we present a proof here .
[ lem : to0 ] let @xmath4 be a real number such that @xmath45 . let @xmath28 such that there exists a positive constant @xmath39 such that @xmath419 for all @xmath90 ( or equivalently , @xmath420 for almost every @xmath32 ) .
then @xmath421 it suffices to prove the case @xmath422 , since the negative case @xmath423 can be proved similarly .
so , we assume @xmath422 .
we first show that holds for all @xmath424 with @xmath425 , where @xmath426 denotes the characteristic function of the open interval @xmath427 . for @xmath424 , by cauchy - schwarz inequality , we have @xmath428 since @xmath429 , noting that @xmath430 by @xmath422 and all @xmath431 are disjoint as @xmath432 , we deduce that @xmath433 as @xmath432 , since @xmath28 .
consequently , holds for all @xmath424 .
so , holds for all @xmath69 that are finite linear combinations of characteristic functions of bounded open intervals . on the other hand ,
define operators @xmath434 by @xmath435 .
by @xmath436 , we have @xmath437 when @xmath420 for almost every @xmath32 , by lemma [ lem : converg ] , we see that @xmath438 let @xmath90 . for an arbitrary @xmath279 , there is @xmath439 , which is a finite linear combination of characteristic functions of bounded open intervals , such that @xmath440 . since @xmath441
, there exists @xmath54 such that @xmath442 for all @xmath443 .
hence , for all @xmath443 , we have @xmath444 hence , @xmath445 . that is , holds .
[ thm : connection ] let @xmath4 be a real number .
let @xmath364 in be subsets of @xmath6 .
suppose that @xmath298 , which is defined in , is a frame in @xmath6 for some integer @xmath54 , that is , there are positive constants @xmath303 and @xmath304 such that @xmath446 then holds for all integers @xmath54 . if in addition @xmath45 , then @xmath17 must be a frame in @xmath6 with the same frame bounds satisfying @xmath447 if @xmath45 and the pair @xmath448 forms a pair of nonhomogeneous dual wavelet frames in @xmath348 for some integer @xmath54 , then @xmath449 forms a pair of homogeneous dual wavelet frames in @xmath6 , that is , both @xmath17 and @xmath450 are frames in @xmath6 , and the following identity holds @xmath451 with the series converging absolutely . by the simple observation in ,
it is easy to see that for all integers @xmath54 , holds with the same constants @xmath303 and @xmath304 . in particular , for all @xmath104 , we have @xmath452 since @xmath45 , by lemma [ lem : to0 ] , we have @xmath453 for all @xmath90 .
now it is easy to deduce that holds . to prove the second claim , by what has been proved , both @xmath17 and @xmath450 are frames in @xmath6 .
note that the pair @xmath448 forms a pair of nonhomogeneous dual wavelet frames in @xmath348 for all integers @xmath54 .
thus , for all integers @xmath54 , we have @xmath454 with the series on the right - hand side converging absolutely . by lemma [ lem : to0 ]
again , for @xmath455 , we have @xmath456 .
now we see that holds . as a direct consequence of all the above results ,
we have [ cor : twf ] let @xmath4 be a real number such that @xmath45 .
let @xmath457 and @xmath458 be subsets of distributions in @xmath68
. then the following statements are equivalent 1 .
@xmath77 , which is defined in , is a frequency - based nonhomogeneous tight wavelet frame in @xmath6 for some integer @xmath54 , that is , @xmath459 and @xmath460 2
. item ( 1 ) is true for all integers @xmath54 ; 3 .
@xmath113 forms a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space @xmath68 ; 4 .
@xmath461 and , , , hold with @xmath462 and @xmath463 ; 5
. there exist @xmath464 such that @xmath465 , and @xmath466 is a nonhomogeneous tight wavelet frame in @xmath6 : @xmath467 moreover , any of the above statements implies that @xmath81 is a frequency - based homogeneous tight wavelet frame in @xmath6 , that is , @xmath468 and @xmath469 we first show that if item ( 3 ) holds , then must be true with @xmath470 and @xmath346 .
since item ( 3 ) holds , by definition , for all @xmath96 , we have latexmath:[\[\lim_{j'\to + \infty } \big ( \sum_{\ell=1}^\mphi \sum_{k\in \z } |\la \ff , \fphi^\ell_{\df^{-j } ; 0 , k}\ra|^2+\sum_{j = j}^{j'-1}\sum_{\ell=1}^\mpsi \sum_{k\in \z } now it is straightforward to see that must be true with @xmath470 and @xmath346 , since @xmath472 .
consequently , by theorem [ thm : sobolev ] , item ( 3 ) implies @xmath459 .
now by theorem [ thm : sobolev ] , ( 1 ) , ( 2 ) , and ( 3 ) are equivalent to each other . the equivalence of ( 3 ) and ( 4 ) is guaranteed by theorem [ thm : main:1:general ] .
the equivalence between ( 1 ) and ( 5 ) is trivial .
is a direct consequence of theorem [ thm : connection ] . from the following result
, we see that there is a natural connection between refinable function vectors and frequency - based nonhomogeneous orthonormal wavelet bases in @xmath6 .
[ prop : orth ] let @xmath4 be a nonzero real number .
let @xmath74 and @xmath75 in be subsets of @xmath6 .
suppose that @xmath473 is an orthonormal basis of @xmath6 for some integer @xmath54 ( this is equivalent to saying that holds with @xmath474 ) .
denote @xmath475 and @xmath476 .
then there must exist @xmath477 matrix @xmath153 and @xmath478 matrix @xmath479 of @xmath35-periodic measurable functions in @xmath172 such that @xmath480 moreover , if @xmath45 , then @xmath481 is also an orthonormal basis of @xmath6 .
note that @xmath473 is an orthonormal basis of @xmath6 if and only if it is an orthonormal basis of @xmath6 for all integers @xmath54 .
consider the expansion of the elements @xmath482 and @xmath483 under the orthonormal basis @xmath484 . by orthogonality
, we have @xmath485 and @xmath486 . noting that @xmath487 and @xmath488 , we see that holds with @xmath489_{\ell',\ell}=\frac{1}{2\pi\sqrt{|\df|}}\sum_{k\in \z } \la \fphi^{\ell'}_{\df;0,0 } , \fphi^{\ell}_{1;0,k}\ra e^{-i k\xi } , \qquad \ell , \ell'=1 , \ldots , \mphi\ ] ] and @xmath490_{\ell',\ell}=\frac{1}{2\pi\sqrt{|\df|}}\sum_{k\in \z } \la \fpsi^{\ell'}_{\df;0,0 } , \fphi^{\ell}_{1;0,k}\ra e^{-i k\xi } , \qquad \ell=1 , \ldots , \mphi,\ , \ell'=1 , \ldots , \mpsi,\ ] ] where @xmath491_{\ell ' , \ell}$ ] denotes the @xmath492-entry of the matrix @xmath493 .
since @xmath494 , all @xmath495_{\ell',\ell}$ ] are well defined elements in @xmath172 .
similarly , all @xmath496_{\ell',\ell}$ ] are well defined elements in @xmath172 .
all the results in the previous sections have been mainly built on the multiresolution - like structure in for stationary nonhomogeneous wavelet systems .
here stationary means that at scale level @xmath497 the dilation is @xmath498 and the generating wavelet functions are independent of the scale level @xmath497 .
the result in lemma [ lem : twolevel ] characterizing the multiresolution - like structure in makes most proofs in the previous sections relatively simple .
nonhomogeneous wavelet systems are closely related to nonstationary wavelets , which are useful in many applications since the nonstationary wavelet filter banks can be implemented in almost the same way and efficiency as a traditional fast wavelet transform .
however , except a few special cases as discussed in @xcite and some references therein , only few theoretical results on nonstationary wavelets are available in the literature . in this section
, we shall see that the notion of a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space is very flexible and similar results hold in the most general setting of fully nonstationary wavelets . since there are few theoretical results on nonstationary wavelets in the literature , it is worth our effort to provide a better picture to understand them in this section
. let us first introduce the notion of a pair of frequency - based nonstationary dual wavelet frames in the distribution space .
let @xmath104 and @xmath198 be a sequence of nonzero real numbers .
let @xmath499 and @xmath500 be subsets of distributions in @xmath68 with @xmath109 and @xmath501 .
we say that the pair @xmath502 forms _ a pair of frequency - based nonstationary dual wavelet frames in the distribution space @xmath68 _ if the following identity holds : @xmath503 where the infinite series in converge in the following sense : 1 . for every @xmath233 , all the series @xmath504 and @xmath505 converge absolutely for every integer @xmath109 , @xmath110 , and @xmath506 ; 2 . for every @xmath107 , the following limit exists and @xmath507 the stationary nonhomogeneous wavelet systems considered in previous sections correspond to the case that @xmath508 , @xmath509 , and @xmath510 for all @xmath109 ; that is , the generating wavelet functions remain stationary ( unchanged ) at all the scale levels @xmath497 .
[ thm : ndwf ] let @xmath54 be an integer and @xmath198 be a sequence of nonzero real numbers such that @xmath199 .
let @xmath511 in and @xmath512 in be subsets of @xmath121 for all integers @xmath109 .
then the pair in forms a pair of frequency - based nonstationary dual wavelet frames in the distribution space @xmath68 ,
if and only if , @xmath513 ( all the above infinite sums are in fact finite , since @xmath514 for all @xmath515 . ) and @xmath516 where @xmath517 $ ] and @xmath518 and @xmath519 let @xmath107 . by lemma [ lem : converg ] , we have @xmath520 } \ff(\xi ) \ol { \fg(\xi+2\pi k ) } i_{\fpsi^j}^{\gl_j k}(\gl_j\xi)\ , d\xi.\ ] ] since @xmath174 and @xmath206 have compact support and @xmath199 , we observe that there exists an integer @xmath207 such that @xmath521 for all @xmath522\bs \{0\}$ ] and @xmath210 . therefore , @xmath523 sufficiency . for @xmath252 , define @xmath524 therefore , by , for @xmath252 , we have @xmath525\ , d\xi.\ ] ] now by , for all @xmath526 , we deduce that @xmath527\ , d\xi.\ ] ] now by and , we conclude that @xmath528
. necessity .
the proof of the necessity part is essentially the same as that of lemma [ lem : twolevel ] .
since @xmath199 , the set @xmath529 is discrete and closed . for any temporarily fixed @xmath32 and @xmath530 , it is important to notice that @xmath531 .
now the same argument as in the proof of lemma [ lem : twolevel ] leads to .
similarly , for any temporarily fixed @xmath32 , since @xmath532 , by , the same argument as in the proof of lemma [ lem : twolevel ] leads to .
[ cor : ndwf ] let @xmath4 be an integer with @xmath45 and @xmath54 be an integer .
let @xmath511 in and @xmath533 in be subsets of @xmath121 for all integers @xmath109 .
then the pair in , with @xmath508 for all @xmath109 , forms a pair of frequency - based nonstationary dual wavelet frames in the distribution space @xmath68 , if and only if , for all @xmath534 and @xmath535 $ ] , @xmath536 for almost every @xmath32 , and the following limit holds in the sense of distributions : @xmath537 since @xmath508 , it is evident that is just . note that @xmath538=\df^j \z$ ] by @xmath539 . for any @xmath540 \bs \{0\}$ ] , we can write @xmath541 for a unique integer @xmath542 and a unique @xmath124 $ ] . now it is easy to check that is equivalent to . as another application of theorem [ thm : ndwf ] , using a similar argument as in the proof of corollary [ cor : twf ] , we have the following result on frequency - based nonstationary tight wavelet frames in @xmath6 .
[ cor : ntwf ] let @xmath54 be an integer and @xmath198 be a sequence of nonzero real numbers such that @xmath199 .
let @xmath74 in and @xmath543 in be subsets of distributions in @xmath68 for all integers @xmath109
. then the following statements are equivalent 1 .
@xmath544 is a frequency - based nonstationary tight wavelet frame in @xmath6 , that is , @xmath545 for all @xmath109 and @xmath546 2 .
the pair @xmath547 forms a pair of frequency - based nonstationary dual wavelet frames in the distribution space @xmath68 ; 3 .
@xmath548 and , hold with @xmath462 and @xmath549 for all @xmath109 .
4 . there exist @xmath550 for all @xmath109 such that @xmath551 for all @xmath109 , and @xmath552 [ prop : ndwf ] let @xmath56 be an integer and @xmath553 be a sequence of nonzero real numbers such that @xmath199 . for integers @xmath554 , let @xmath555 in and @xmath556 be subsets of @xmath121 .
then @xmath557 forms a pair of frequency - based nonstationary dual wavelet frames in the distribution space @xmath68 for every integer @xmath558 , if and only if , @xmath559\cup [ \gl_{j+1}^{-1}\z],\ ; j \ge j_0\ ] ] and @xmath560 where @xmath561 , @xmath515 , are defined in and @xmath562 by the same argument as in theorem [ thm : main:1:general ] , we see that the pair in forms a pair of frequency - based nonstationary dual wavelet frames in @xmath68 for all integers @xmath558 , if and only if , @xmath563 and @xmath564 by lemma [ lem : twolevel ] , is equivalent to . by lemma [ lem : to1 ] ,
is equivalent to . for @xmath565 with @xmath566 , the degree of @xmath153
is defined to be @xmath567 .
we finish this paper by the following result which connects a nonstationary wavelet filter bank obtained via a generalized nonstationary oblique extension principle with a pair of frequency - based nonstationary dual wavelet frames in the distribution space .
[ thm : mra : ndwf ] let @xmath568 be a sequence of nonzero integers such that @xmath569 . define @xmath570 and @xmath571 for all @xmath572 .
let @xmath573 and @xmath574 , @xmath575 , be @xmath35-periodic trigonometric polynomials such that @xmath576 for all @xmath575 and @xmath577 define @xmath578 then all @xmath579 , @xmath580 are elements in @xmath277 satisfying @xmath581 let @xmath582 , @xmath583 and @xmath584 , @xmath585 with @xmath586 and @xmath575 be @xmath35-periodic measurable functions in @xmath121 .
define @xmath587 then @xmath588 in and @xmath555 in are subsets of @xmath121 for all @xmath286 .
moreover , the pair in forms a pair of frequency - based nonstationary dual wavelet frames in @xmath68 for every integer @xmath589 , if and only if , for all @xmath575 , @xmath590,\label{mra : filter : eq2}\end{aligned}\ ] ] and @xmath591 where @xmath592 we first establish an inequality on the decay of @xmath593 , which plays a critical role in this proof to show . by @xmath576 and bernstein inequality @xmath594'\|_{\tlp{\infty } } \le \deg(\fa^j ) \| [ \fa^j]\|_{\tlp{\infty}}$ ] , @xmath595'$ ] denotes the derivative of @xmath573 . observe a simple inequality @xmath596 for @xmath597 and @xmath598 .
it is easy to prove that @xmath599 for @xmath597 .
consequently , @xmath600 , which leads to .
now for any @xmath601 with @xmath602 , by and , we deduce that @xmath603 hence , by , we have @xmath604 on the other hand , by a similar idea as in ( * ? ? ?
* page 93 ) and ( * ? ? ?
* page 932 ) , we have @xmath605 where @xmath606 . by and , we deduce from the above identity that @xmath607 that is , for all @xmath601 with @xmath602 , we have @xmath608 the above inequality implies the uniform convergence of @xmath609 for @xmath270 on any bounded set . since @xmath610 , we conclude from that @xmath611 . since @xmath612 , it is evident that all @xmath613 are elements in @xmath121 .
similarly , we can prove @xmath614 and all @xmath615 are elements in @xmath121 .
note that @xmath616 , that is , @xmath617 .
so , holds .
also note that @xmath618 and @xmath619 for at most countably many @xmath32 .
note that with @xmath497 being replaced by @xmath620 is equivalent to @xmath621 for all @xmath122 and @xmath622 , where @xmath623 and @xmath624 for all @xmath625 .
now it is easy to directly verify that and are equivalent to .
we now show that is equivalent to . by @xmath610 and
, we have @xmath626 now by and , using lebesgue dominated convergence theorem , by the same argument as in theorem [ thm : main:2 ] , we conclude that holds if and only if holds . by proposition [ prop : ndwf ]
, the proof is completed . | in this paper , we study nonhomogeneous wavelet systems which have close relations to the fast wavelet transform and homogeneous wavelet systems . we introduce and characterize a pair of frequency - based nonhomogeneous dual wavelet frames in the distribution space ; the proposed notion enables us to completely separate the perfect reconstruction property of a wavelet system from its stability property in function spaces . the results in this paper lead to a natural explanation for the oblique extension principle , which has been widely used to construct dual wavelet frames from refinable functions , without any a priori condition on the generating wavelet functions and refinable functions . a nonhomogeneous wavelet system , which is not necessarily derived from refinable functions via a multiresolution analysis , not only has a natural multiresolution - like structure that is closely linked to the fast wavelet transform , but also plays a basic role in understanding many aspects of wavelet theory . to illustrate the flexibility and generality of the approach in this paper
, we further extend our results to nonstationary wavelets with real dilation factors and to nonstationary wavelet filter banks having the perfect reconstruction property . |
hydrodynamics has been rather successful at describing data obtained in relativistic nuclear collisions at rhic .
usually , smooth initial conditions are assumed ( see e.g. fig.1 in @xcite and fig.3 in @xcite . on the other side , microscopic codes such as nexus predict initial conditions event - by - event , which are quite irregular as shown in fig.1 .
the question we address here is whether such structures ( hot spots or more precisely hot tubes ) can have a sizable effect on variables . to solve the hydro equations with very irregular initial conditions
, we use the spherio code .
this code is based on the method of smoothed particle hydrodynamics , originally developed in astrophysics and adapted to relativistic heavy ion collisions in @xcite .
the version of nexspherio used here has initial conditions provided by nexus @xcite and normalized by an @xmath0-dependent factor to reproduce @xmath1 in each centrality window @xcite .
the equation of state has a critical point @xcite .
@xmath2 is fixed ( mostly ) by @xmath3 and depends on the centrality window ( i.e. number of participants ) .
centrality windows are defined using participant number and not impact parameter @xcite .
an ideal fluid is assumed , a code with smoothed particle hydrodynamics and dissipation is under development @xcite .
in the following , we present a summary of results obtained using smooth initial conditions and running once the spherio hydro code ( standard approach ) or using a set of nexus initial conditions , running for each initial conditions the spherio hydro code and computing averages over the set for various observables ( event - by - event hydrodynamics ) . as can be seen in figure 2 ( left ) , the high @xmath4 part is lifted .
this is expected since hot tubes must expand more violently , producing more high @xmath4 particles @xcite .
@xmath5 is flatter as seen in figure 2 ( centre ) .
this is also expected as the isotropic expansion of hot tubes produces more high @xmath4 particles and lowers @xmath5 @xcite .
in addition , @xmath6 has no shoulder @xcite as seen in figure 2 ( right ) .
the effects ( isentropic expansion ) of the hot tubes are more visible in regions of lower matter density present at larger @xmath0 s @xcite . in @xcite , we argued that the hot tubes should manifest themselves giving smaller hbt radii .
however , the situation might be more complicated .
another observable where hot tubes might manifest themselves is the ridge , a structure observed in the 2 particle correlations , plotted as function of pseudorapitity difference @xmath7 and azimutal angle difference @xmath8 between a high @xmath4 trigger hadron and its associated hadrons ( see e.g. @xcite ) .
the structure is @xmath7 independent . in nexspherio
, the hot tubes can lead to such a ridge for the e - by - e initial conditions and not the smooth ones @xcite .
finally , the fluctuations in the e - b - e initial conditions also manifest themselves in fluctuations of @xmath9 ( as well as @xmath10 ) . the predicted values for @xmath9 at 130 a gev @xcite and estimates at 200 a gev @xcite are in agreement with data @xcite .
improvements to remove the non - flow effects have been reported by star and phobos , see e.g. @xcite .
in this section , we present some new _ preliminary _ results obtained with nexspherio on directed flow . if a nucleus - nucleus collision is a number of independent nucleon - nucleon collisions , the momentum distribution is isotropic .
if instead , it leads to thermalized matter in the overlap region , the momentum distribution is stretched along the impact parameter direction , @xmath9 is a measure of this stretching ( so teaches about ic , thermalization , etc ) .
there is also the possibility that the momentum distribution be shifted / deformed towards one of the sides in the x - y plane , @xmath10 is a measure of this shift . at some energy , a `` wiggle '' in @xmath11 is predicted . in some microscopical models such as rqmd and urqmd
, this could be the case for nucleons at rhic energy @xcite . in hydro models
, this could be the case for the fluid , if a qgp phase occurs @xcite . at sps energy
( 40 a gev and 158 a gev ) , it was shown @xcite that pions and protons behave oppositely .
pion directed flow as function of rapidity has no wiggle and crosses y=0 with a negative slope while nucleon directed flow has no wiggle and crosses y=0 with a positive slope ( except perhaps at the higher energy , in the more peripheral bin , where there is a hint of wiggle ) . at rhic ,
directed flow for charged particles is rather similar to what was obtained at sps for pions : it crosses @xmath12 with a negative slope @xcite .
this is understandable since charged particles are mostly pions , the fluid directed flow must be dominated by pions .
the turnover in @xmath11 occurs for different values of @xmath0 in phobos and in star ( see below ) .
results for identified particles are becoming available @xcite .
in addition , comparison of results for directed flow in cu+cu and au+au collisions show no system - size dependence @xcite .
nexspherio results are in qualitative agreement with phobos for all @xmath0 s and quantitative agreement for @xmath13 ( figure 3 left ) .
they are in qualitative agreement with star for @xmath14 but turnover occurs for smaller @xmath0 than for star ( figure 3 right ) .
@xmath11 from nexspherio for various centrality windows for au+au and cu+cu at 200 a gev is shown in figure 4 .
little dependence on a is seen in the windows 6 - 15% to 45 - 55% .
statistics must be improved .
+ figure 5 ( left ) illustrates particle dependence . in nexspherio
, protons have a big wiggle , pions have a plateau ( left ) .
a similar result was obtained using urqmd @xcite . in figure 5 (
right ) , it is seen that @xmath11 has a plateau for fluctuating initial conditions and a somewhat stronger negative inclination for smooth initial conditions .
a short review of possible effects of fluctuating initial conditions , rather than smooth ones , was presented .
in addition to providing a reasonable description of various observables , as is possible with smooth initial conditions , some new effects were listed , most notably the ridge effect and the @xmath9 fluctuations , which do not appear when using the smooth initial conditions . 99 t.hirano phys.rev.c65 ( 2001 ) 011901r c.nonaka and s.bass phys.rev.c75 ( 2007 ) 014902 c.e.aguiar , t.kodama , t.osada@xmath15&@xmath15y.hama , j.phys.g27(2001)75 ;
y.hama , t.kodama & o.socolowski jr .
35(2005)24 h.j .
drescher , f.m .
liu , s. ostrapchenko , t. pierog and k. werner , phys .
c65 ( 2002 ) 054902 y.hama et al .
phys . at .
71 ( 2008 ) 1558 y. hama et al .
a774 ( 2006 ) 169 r.andrade et al . braz.j.phys .
34 ( 2004 ) 319 g.s.denicol , t.kodama , ph.mota in progress .
andrade et al .
( 2008 ) 112301 y.hama , these proceedings r.p.g .
andrade et al .
97 ( 2006 ) 20230 b.b.back et al .
( phobos ) phys .
b578 ( 2004)297 b.b .
back et al . (
phobos ) phys . rev .
c72 ( 2005 ) 051901 o.socolowski jr .
93 ( 2004 ) 182301 j.putschke et al .
( star ) j.phys . g34 ( 2007 ) s679 j.takahashi et al . in progress .
c.e.aguiar et al .
a698 ( 2002 ) 639c c.loizides et al .
( phobos ) j. phys .
g34 ( 2007 ) s907 p.sorensen et al .
( star ) j. phys .
g34 ( 2007 ) s897 b.wosiek , these proceedings .
snellings et al . phys.rev.lett .
84 ( 2000 ) 2804 bleicher and stcker phys.lett.b526 ( 2002 ) 309 csernai and rhrich , phys.lett.b458 ( 1999 ) 454 brachmann et al .
phys.rev.c61 ( 2000 ) 024909 rischke et al . heavy ion phys . 1 ( 1995 ) 309 ivanov et al .
acta phys . hung .
15 ( 2002 ) 117 c.alt et al .
( na49 ) phys .
c68 ( 2003 ) 034903 ( phobos ) phys .
97 ( 2006 ) 012301 j.adams et al .
( star ) phys .
92 ( 2004 ) 062301 b.i .
abelev et al .
( star ) arxiv:0807.1518 j. chen et al .
( star ) j. phys .
g35 ( 2008 ) 044072 | a comparison is made between results obtained using smooth initial conditions and event - by - event initial conditions in the hydrodynamical description of relativistic nuclear collisions .
some new results on directed flow are also included . |
it has been established that every massive , local spheroid harbours a supermassive black hole ( smbh ) in its centre whose mass is proportional to that of its host ( e.g. @xcite ; @xcite ) .
this suggests that the black holes ( bhs ) and their surrounding galaxies were formed synchronously .
this hypothesis has found support from hydrodynamical simulations of galaxy formation , which use feedback from winds and outflows from active galactic nuclei ( agn ) to link the growth of the smbh to that of its host ( e.g. @xcite ; @xcite ; @xcite ) .
thus these models support a picture , first presented by @xcite , where a starburst - dominated ultra - luminous infrared galaxy ( ulirg ) , arising from a merger , evolves first into an obscured qso and then into an unobscured qso , before finally becoming a passive spheroid .
the high - redshift population of ulirgs in this proposed evolutionary cycle are the submillimetre ( submm ) galaxies ( smgs ; @xcite ; @xcite ; @xcite ) .
these systems have ulirg - like bolometric luminosities , @xmath14l@xmath5 ( @xcite ; @xcite ) , and they have many of the properties expected for gas - rich mergers ( @xcite ; @xcite ) .
this population evolves rapidly out to a peak at @xmath15 , crudely matching the evolution of qsos @xcite and providing additional circumstantial evidence for a link between smbh growth and spheroids .
two further results have shed light on the evolutionary link between smgs and qsos .
firstly , a modest fraction of optically luminous qsos at @xmath16 are detected in the submm / mm ( @xmath17% ; @xcite ) showing that the qso- and smg - phases do not overlap significantly , given the lifetime estimates of the two populations ( qsos make up @xmath18% of flux - limited samples of smgs ; @xcite ) .
but when a qso _ is _ detected in the submm / mm then it could to be in the transition phase from an smg to an unobscured qso , making its properties a powerful probe of the evolutionary cycle ( e.g. @xcite ; @xcite ) .
secondly , the evolutionary state of the smbhs within smgs can also be judged using the 2-ms _ chandra _ deep field north observations @xcite to derive accurate agn luminosities and hence lower limits on the bh masses ( @xmath19 ) in those smgs with precise redshifts in this region ( @xcite ; @xcite ) .
these studies suggest that the agn in typical smgs are growing almost continuously but that the smbhs in these galaxies appear to be several times less massive than seen in comparably massive galaxies at @xmath20 @xcite .
together these results argue for a fast transition from a star - formation - dominated smg - phase to the agn - dominated qso - phase @xcite .
the latter phase will result in the rapid bh growth necessary to account for the present - day relation between spheroid and smbh masses ( e.g. @xcite ; @xcite ) .
_ can we confirm this and more generally test the proposed evolutionary link between smgs and qsos at the peak of their activity at @xmath16 ? _
this evolutionary cycle has been tested in the local universe by comparing the properties of qsos and ulirgs ( e.g. @xcite ) .
however , both of these populations are @xmath21 less abundant in the local universe than they were at the era of their peak activity at @xmath22 and so we have to be cautious about extrapolating from local examples to the high - redshift progenitors of the bulk of today s massive spheroids ( @xcite ; @xcite ) . thus , to properly test the validity of this cycle for typical spheroids we have to compare qsos and ulirgs at the era where their populations peaked : @xmath22 .
the critical pieces of information needed to test the link between smgs and qsos are the relative dynamical , gas and smbh masses of these two populations . in principle
the dynamical masses can be derived from optical or near - infrared observations of emission line gas in the smgs or qsos ( see @xcite ) . however , there is the problem of removing the qso emission in these observations , as well as the effects of dust obscuration and outflows .
in contrast , molecular co emission line profiles are relatively immune to the effects of obscuration and outflows , while at the same time yielding additional constraints on the relationship between qsos and smgs from their gas masses .
there is currently a lack of sensitive co observations of qsos at @xmath233 , with data published on only eight sources ( e.g. @xcite ; @xcite ; @xcite ; @xcite ) .
instead the focus has been on co studies of qsos at @xmath24 ( e.g. @xcite ; @xcite ; @xcite ) , although these qsos have little overlap with the redshift range where smgs are typically detected .
the paucity of co constraints for @xmath25 qsos also reflects the difficulty in determining their systemic redshifts with sufficient precision to guarantee that the co emission falls within the bandwidth of typical millimetre ( mm ) correlators .
however , sensitive near - infrared spectroscopy of the civ , mgii , and [ oiii]5007 emission lines in qsos can provide redshifts with required precision as well as h@xmath26 or h@xmath27 fluxes and line widths to yield @xmath19 estimates ( @xcite ; @xcite ) .
we have carried out a quantitative test of the proposed link between smgs and qsos at @xmath16 where both populations are most common .
we have obtained precise systemic redshifts from near - infrared spectroscopy of potential transition qsos ( i.e. submm / mm - detected qsos ) and then used the iram plateau de bure interferometer ( pdbi ) to search for co emission .
we relate their dynamical , gas and smbh masses to smgs from the pdbi co survey @xcite .
we test : a ) whether the cold gas masses in these qsos are similar to those in smgs ; b ) whether the line widths and dynamical masses of these two populations are comparable ; and c ) how the ratio of smbh to dynamical masses for these submm - detected qsos relate to the estimates for smgs and those for optically luminous qsos ( which lie on the present - day @xmath19@xmath28 relation ; @xcite ) .
together these observations can constrain the proposed evolutionary sequence which links qsos to the formation of massive young galaxies and smbhs at high redshift .
we describe the sample selection , observations and data reduction in [ obsdr ] .
the results of the near - infrared and mm co spectra are given in [ results ] .
the co properties of the submm - detected qsos are compared and contrasted with smgs in [ discuss ] , and we discuss the evolutionary status of the submm - detected qsos in [ evolution ] .
our conclusions are given in
[ concl ] .
we adopt cosmological parameters from the _ wmap _ fits in @xcite : @xmath29 , @xmath30 , and @xmath31kms@xmath32mpc@xmath32 .
all quoted magnitudes are on the vega system .
the objective of this study is to test the link between smgs and qsos at the era where these populations peaked .
we have therefore selected radio - quiet qsos with a redshift distribution matched to smgs ( @xmath33 , @xcite ) and with @xmath34 fluxes @xmath35mjy .
the @xmath36mjy flux limit selects comparatively bright qsos and smgs and is imposed so that the sources would be sufficiently bright to be detected in co. we choose submm - detected qsos from three main samples : we select five qsos from a submm photometry survey of optically - bright qsos by @xcite hs1002 + 4400 , j140955.5 + 562827 , j154359.3 + 535903 , hs1611 + 4719 , and j164914.9 + 530316 , we select two qsos from the submm photometry survey of x - ray absorbed qsos by @xcite rxj121803.82 + 470854.6 and rxj124913.86055906.2 , and finally we include two qsos found in blank - field submm surveys by @xcite and @xcite , smmj131222.35 + 423814.1 and mmj163655 + 4059 , respectively , and smmj123716.01 + 620323.3 , an optically - selected qso targetted in submm photometry mode by @xcite .
the final sample comprises 10 submm - detected qsos with similar fluxes to the smg sample ( @xcite ; @xmath35mjy at 850@xmath0 m , assuming 850@xmath0m/1.2 mm colours typical of smgs ; @xcite ) , with a median of 10mjy .
the seds of these qsos demonstrate that the bulk of the mm / submm luminosity arises from dust emission and is very likely associated with star formation , rather than an agn ( @xcite ; @xcite ) .
our sample spans six orders of magnitude in qso @xmath37-band brightness , enabling us to potentially trace how co properties depend on qso luminosity or @xmath19 ( see fig .
[ fig : selection ] ) . [ cols="<,^,^,^,^,^,^,^,^ " , ] \(a ) this quantity was obtained by fitting a gaussian distribution to the co spectrum ( see text ) + ( b ) derived assuming @xmath38 and @xmath39k ( see text ) + ( c ) derived following @xcite ( see text ) + ( d ) derived assuming @xmath40 + ( e ) derived assuming a disk radius of 2kpc + ( f ) 1.2 mm flux is corrected for synchrotron emission before computing @xmath41 +
we now compare the co properties of the submm - detected qsos with smgs . in order to fairly compare the smg and qso samples , we have discarded three strongly lensed galaxies from the co - observed smg sample from @xcite that would not have been detected with @xmath35mjy at 850@xmath0 m .
these would lie below our sample flux limit except for the boost from the lensing magnification , and they thus probe an intrinsically fainter population .
this leaves a sample of 17 smgs ( of which 11 are co - detected ) from @xcite with a median observed 850@xmath42 m flux of 8.2mjy .
we compare median values of the gas masses , co linewidths , sfes , m@xmath43 , and m@xmath44/m@xmath45 .
the error bars on these median values represent the spread of values , and have been obtained by taking the standard deviation of the median values from 50 trials of randomly drawing out the appropriate number of values out of each population with replacement .
we note that the dominant uncertainties in @xmath46 , @xmath41 , and @xmath47 are likely the assumed co - to - gas conversion factor , the assumed @xmath27 and @xmath48 in the sed fitting , and the assumed co radius and inclination angle , respectively , and have not been included in each error budget , although the relative @xmath46 , @xmath41 , and @xmath47 in our samples will be correct if the same assumptions hold for all our smgs and submm - detected qsos . the median gas mass of our co - detected qsos is @xmath49m@xmath50 ( see fig . [
fig : gas ] ) .
this is similar to the median gas mass of the comparison sample of 11 co - detected smgs ( m@xmath51m@xmath50 ) . including the upper limits from the co - undetected objects in each sample yields median gas masses of @xmath52 and @xmath53m@xmath50 for submm - detected qsos and smgs , respectively .
not surprisingly , we find that the gas mass distributions of both samples are statistically indistinguishable at the present sample sizes , as we show below .
given the moderate fraction of non - detections in both samples , we employ survival analysis ( see e.g. @xcite ) to estimate the intrinsic distributions of the gas masses of submm - bright smgs and qsos .
this includes the detections _ and _ upper limits , assuming that the data are censored at random ( i.e. that the chance of only a gas upper limit being available for an object is independent of the true value of the gas mass ) .
given the current detection limit and small dynamic range in our sample , there is no strong apparent trend with the selection criteria ( i.e. with submm flux ) versus whether a particular source is detected or not , and thus the assumption of the data being randomly censored seems reasonable .
we employ the widely - used product - limit avni estimator ( @xcite ; see also @xcite ) to construct a maximum - likelihood - type reconstruction of the true distribution of gas masses in the qso and smg samples ( see fig .
[ fig : gas ] ) .
we now test whether or not our censored samples of smg and qso gas masses are likely to have been drawn from the same distribution using the non - parametric gehan test , following @xcite . the gehan statistic yields @xmath54 ( equivalent to @xmath55 ) , revealing that the difference between the gas mass distributions for smgs and qsos is about 35% likely due to chance .
we consider 35% to be a high value , and therefore conclude that the gas masses of our qso sample are indistinguishable from that of the smg sample at the same epoch of @xmath223 for the high mass end ( where we are the most sensitive ) . within our qso sample
we find a broad similarity in the star formation efficiency of the co - detected and undetected qsos ( see [ lfircalc ] ) , assuming conservatively that the co - undetected subset lies just below the detection threshold . a kolmogorov - smirnov ( ks ) test reveals a @xmath56 chance that the two subsamples are drawn from the same parent population , indicating that we have not detected a statistically significant difference in the sfe between the co - detected and undetected qsos in our small sample given the current detection limits .
what is the cause then of a given submm - detected qso in our sample , which all have similar @xmath41 , being either co - detected versus co - undetected ?
a possible explanation could be that the co - detected submm - detected qsos are physically larger and contain proportionally more massive co reservoirs than the co - undetected qsos , although time - dependent star formation , differences in the seds or filling factors could also be causes for the scatter in the @xmath41@xmath57 correlation ( see fig .
[ fig : sfe ] ) . locally , l@xmath58 increases with l@xmath59 for ( u)lirgs , with the @xcite sample of smgs extending this trend out to the highest far - infrared luminosities ( @xmath60l@xmath50 ) . for comparison , in fig .
[ fig : sfe ] we have plotted our qsos on the l@xmath58l@xmath59 diagram along with lirgs , ulirgs and smgs . the co - observed smgs and qsos seem to lie along the relation within the considerable uncertainties in their far - infrared luminosities , suggesting that the smgs and qsos are undergoing similar star formation modes , despite the qsos hosting luminous agn activity . to quantify this we compare the sfe of the qsos to the smgs , where @xmath41 for the smgs has been calculated using the method outlined in [ lfircalc ] and
their @xmath57 are taken from @xcite . as fig .
[ fig : sfe ] shows , the smgs appear to have similar sfe with respect to qsos on average , although this probably reflects the similar submm flux selection criteria .
given that the median @xmath41 and @xmath46 of the qso and smg samples are similar , @xmath61l@xmath50 and @xmath62l@xmath50 and @xmath52m@xmath50 and @xmath53m@xmath50 ( including all co - detected and -undetected qsos and smgs ) , respectively , the sfes are also similar : medians of @xmath63 and @xmath64l@xmath65 , for qsos and smgs respectively .
we perform a gehan test on these censored data sets and find @xmath66 ( @xmath67 ) , revealing that the difference between the distribution of sfe in smgs and qsos is about 15% likely due to chance , indicating that we have not detected a significant difference in the sfe between our small samples of smgs and qsos . a comparison between the observed co line profiles of smgs and qsos requires care as we know that there is an inherent difference in the types of lines seen in smgs versus qsos : five of the 11 smgs detected in co in our comparison sample show hints of double - peaked profiles ( @xcite ; @xcite ) , while none of our qso co emission lines appear double - peaked ( which is perhaps not surprising given our small sample size and low snr data ) . these double - peaked profiles could indicate merger activity ( as likely for smgs from high - resolution observations ) , or a rotating gas disk / ring ( e.g. @xcite ) . in both cases , measuring the separation between the peaks is a more appropriate measure of the dynamics of the merger or rotating disk system than the fwhm of the individual peaks
. in cases
where double - peaked profiles are present we need to fit two gaussian profiles simultaneously in order to derive the velocity offsets of the peaks .
we have therefore refit all of the smgs and qsos with single and double gaussian profiles using the @xmath68 statistic as a measure of the goodness - of - fit .
we find that 4/11 of the co - detected smgs are better fit at the @xmath69 level with a double gaussian profile over a single gaussian profile @xcite .
none of the submm - detected qsos are significantly better fit by double gaussian profiles , although we note that the random chance of drawing six single peaked profiles ( with replacement ) out of a sample containing seven single - peaked and four double - peaked profiles ( i.e. the smg sample above ) is 7 percent .
the fact that the qso profiles appear to be single - peaked could indicate that our data are too low snr to detect double - peaked profiles ( especially for a merger , where both the width and peak strength of the lines can be different ) , or that we are viewing the qsos almost face - on@xmath70 ( similar to those found for double - peaked smgs ) and equal line fluxes and shift them to a velocity separation such that the fwhm of a single - gaussian profile fit is @xmath71 ( our qso sample median fwhm ) , yielding typical velocity separations of 300@xmath72 .
this translates into a conservative limit on the inclination of the system of @xmath73 by using the relation @xmath74 and taking @xmath75 and @xmath76 ( the latter value is what is observed typically for smgs by fitting double gaussian profiles ) . ] if the co is contained in a disk , or that they are at a late stage of a merger , or that they are not in the course of a merger . for our six submm - detected qsos
we compare the co fwhms and dynamical masses with the co - detected sub - sample of 11 smgs taken from the literature ( @xcite ; @xcite ) .
the median fwhm of the six co - detected qsos and seven single - gaussian profile fit smgs are @xmath77 and @xmath78 , respectively .
when we include the double gaussian profile fits in the median estimate , taking the separation between the peaks as a proxy for the dynamics of the system , the median for the smgs slightly decreases to @xmath79 .
we have plotted the histogram of the submm - detected qso and smg line widths for comparison in fig .
[ fig : fwhm ] ( c.f . fig . 2 in @xcite ) .
the co fwhm of qsos and smgs appear to be similar . to quantify this we use a standard two - sided ks test which reveals a 95% probability that the two distributions are drawn from the same parent population ( this drops to 77% if only the single - line gaussian profile smgs are used ) .
this is in stark contrast with @xcite , who find only a 0.9% probability for an inhomogeneous sample of 12 smgs and 15 qsos from a literature compilation ( @xcite ) .
however , we note that they have used the smg single - gaussian profile fit line widths even in cases where clearly double - peaked co profiles are apparent , leading to artificially broad fwhm in these cases . their analysis is also based on a mixed sample over a wide redshift range and includes strongly lensed sources . also , the qsos studied here have larger linewidths than have been reported for other qsos and they are also those with the lowest snrs and thus the most difficult to fit .
therefore the differences in our results for @xmath22 qsos and those in @xcite for higher redshift qsos can probably mainly be attributed to the small sample sizes involved and the somewhat different approaches ( rather than a real statistically significant effect ) when all possible sources of error are taken into account .
since the co line widths of smgs and qsos are similar , it may follow that the dynamical masses of smgs and qsos are also similar .
the uncertainties in this are if the gas is distributed similarly in both populations ( i.e. either a disk or merger scenario ) , if their sizes ( @xmath37 ) are the same , and if their average inclination angles ( @xmath80 ) are similar .
@xmath81kpc is believed to be appropriate for the gas distribution in smgs ( e.g. @xcite ; @xcite ) .
one of our qsos , j140955.5 , has been observed at pdbi by @xcite in co(76 ) with a @xmath82 resolution and was unresolved , giving an upper limit to the gas reservoir of 5kpc ( though this could be too high a j transition to trace the bulk of the gas reservoir ) .
no sizes have been measured for our sample of qsos , although the co emission seems to be confined to a compact region in other samples of high - z qso @xcite as inferred by comparing a range of co line ratios ( see also @xcite who observe gas distributed over a 2.5kpc radius and @xcite ) .
this agrees with the recent results from @xcite who find that co emission in a @xmath83 qso is unresolved with a beam size of @xmath84arcsec , implying that the molecular gas is contained within a compact region with a radius of @xmath85kpc .
the qso from @xcite was observed in a high j transition , co(54 ) , and one might worry that such high transitions might not trace the bulk of the reservoir .
co(10 ) and ( 21 ) observations of other @xmath24 qsos ( e.g. @xcite ) have confirmed that the high j transitions do trace the bulk of the reservoir , and thus provide good size constraints ( this is in contrast to what is observed in most nearby galaxies ) .
spatially resolved observations exist for local qsos , albeit the data are sparse : @xcite measure a @xmath86kpc disk in mrk 231 ; @xcite find a @xmath86kpc ring - like molecular gas distribution ; and @xcite estimate @xmath87kpc for the co emission in a low luminosity qso . therefore , assuming 23kpc radius regions will probably still be reasonable once these galaxies are resolved .
while the average orientation of the smgs is likely to be close to the value expected for random orientations , @xmath88 , since they are not prone to orientation selection effects ( see @xcite ) , the submm - detected qsos could be orientated preferentially with respect to the sky plane due to their selection as optically luminous qsos in the first place .
we find evidence for this from a comparison of the [ oiii ] line width ( see table [ tab : nir ] ) , a proxy for the measure of the stellar velocity dispersion ( supporting the idea that the narrow - line region gas is in orbital motion in the gravitational potential well of the bulge ; @xcite ; @xcite ; @xcite ) , with the co line width .
we find that the [ oiii ] fwhm are a median factor of @xmath89 broader than the co line profiles ( with a large scatter ; ratios range from 0.64.3 ) , suggesting indeed that inclination effects could be important in the co line widths of qsos ( see also @xcite ) .
the lack of the double - peaked co profiles in the submm - detected qsos could be interpreted as a later phase of a merger in the typical proposed evolutionary sequence , with smgs representing an earlier phase of the merger .
we also know that at least two of our qsos must have @xmath90 ( alternatively the assumed standard co - to - h@xmath91 conversion factor could be too high or else the assumed 2kpc radius is too small ) or else the cold gas mass estimates exceed the dynamical masses ( even given the large error bars ; see table [ tab : derived ] ) .
conversely , some of our broadest co line detections suggest that their inclination angles might be larger than @xmath92 .
other constraints come from the obscured : unobscured agn ratio ( @xmath9310 ) for a sample of smgs , where @xcite suggest that qsos could be seen close to the line of sight ( @xmath94@xmath95 ) .
we thus adopt the best available estimates of the average inclination angles for smgs and qsos of @xmath96 and @xmath92 , respectively , in calculating the dynamical masses below .
high - resolution co observations could help to place tighter constraints on the inclination angles of our systems . assuming a 2kpc radius and an average inclination angle of @xmath92 for the submm - detected qsos and @xmath96 for the smgs
, we find median dynamical masses , m@xmath97 , of m@xmath98kpc@xmath99 and @xmath100m@xmath5 , respectively .
this results in gas - to - dynamical mass fractions of @xmath101% and 30% for qso and smg populations , respectively .
the gas mass accounts for a significant fraction of the dynamical mass of both classes , indicating that the host galaxies are at an early evolutionary stage .
we place a constraint on the spheroid mass by taking the dynamical mass estimate from
[ fwhm_discuss ] , assuming @xmath102 and a spheroid radius of 2kpc or less ( as we assumed for the co emission ) , yielding a median spheroid mass of @xmath103m@xmath5 . in principle one should subtract the gas and bh masses , although we note that doing so makes a negligible difference in general , given the large error bars on @xmath47 , and we also have an additional complication that two of our co - detected qsos have @xmath104 .
other uncertainties to our @xmath105 estimates include the possibility that the cold gas in submm - detected qsos is more widely distributed ( e.g. due to entrainment in outflows ) than we have assumed here ( @xmath106kpc ) , which would mean that all of our dynamical ( and hence spheroidal ) masses have been underestimated .
similarly , if the spheroid is significantly more extended than the co emission then the spheroid masses will be larger .
for example , assuming a characteristic radius of @xmath107kpc for the spheroid would result in an increase in the mass of a factor of @xmath108 times . based on these assumptions
we estimate @xmath109 for our qso sample .
this is approximately an order of magnitude larger than the local ratio of @xmath110 for the bhs in local spheroids ( fig .
[ fig : ratio ] ; @xcite ; @xcite ) . thus our co survey of @xmath16 qsos appears to support the claim that @xmath8 for @xmath25 agn is higher than observed locally ( e.g. @xcite ; @xcite ) .
our @xmath8 ratio could be made to match the local relation by either decreasing the inclination angles to @xmath111 or by adopting a spheroid extent ten times larger than we have assumed .
our results and those of @xcite suggest that bh growth may occur more rapidly than spheroid formation in qsos at @xmath16 ( although both studies could be prone to strong selection effects ; see @xcite ) .
similar co - based studies of rarer individual qsos at higher redshifts also support this conclusion ( e.g. @xcite ; @xcite ) .
this result contrasts strongly with the @xmath8 claimed for smgs , which lie just below the local ratio ( fig .
[ fig : ratio ] ; @xcite ) . however , as mentioned in [ smbhmasses ] , six out of nine of our submm - detected qsos have very massive bhs ( @xmath112m@xmath5 ) which are far more massive than those inferred for typical smgs ( @xcite ; @xmath113m@xmath5 ) and would place them among the largest known masses for local bhs .
they are therefore likely to be considerably rarer than typical smgs . to make a fairer comparison between smgs and submm - detected qsos we need to focus on those co - detected submm - detected qsos with @xmath19 more similar to that found for smgs .
we have three submm - detected qsos in our sample with bhs more similar to those of smgs , unfortunately two of these are undetected in co , so we can not determine their dynamical masses .
nevertheless , the one example with a marginal co detection , smmj131222.35 , has an implied spheroid mass of @xmath114m@xmath5 ( assuming @xmath102 ) and hence @xmath115 .
this is an order of magnitude below the ratio for the more optically luminous qsos and places smmj131222.35 approximately on the local relation ( fig .
[ fig : ratio ] ) .
while based on a single source marginally detected in co , this result hints that submm - detected qsos with @xmath116m@xmath5 might be the smg
qso ` transition ' objects we are interested in .
we discuss this possibility further in [ evolution ] .
in this section we discuss the various observations which constrain the evolutionary relationship between smgs and submm - detected qsos . for the purposes of this discussion ,
we adapt the proposed evolutionary pathway of @xcite and so test the scenario where the smg population evolves through a submm - detected qso phase , into a submm - undetected qso and finally a passive elliptical .
now we ask are the observed properties of the smgs and submm - detected qsos consistent with their observed gas depletion , amount of bh growth , and space densities required to link them in this evolutionary sequence ? to provide a characteristic timescale for our discussion we first determine the lifetime of the submm - detected qso phase .
we assume that the far - infrared luminosities of the submm - detected qsos arise from dust - obscured star formation and that the molecular gas reservoirs we detect via their co emission are the fuel for this star formation . in this way
we determine that the qsos have enough cold gas to sustain their current episode of star formation ( assuming a star formation efficiency of 100% ) for @xmath117m@xmath118m@xmath119myr .
@xmath120 is on the order of a couple of dynamical times within our assumed co radius of 2kpc ( @xmath121myr ) , which could be defined as a ` maximal starburst ' ( e.g. @xcite ) as is commonly found with other populations of starburst galaxies such as ulirgs at low - redshift and smgs and other qsos at high-@xmath122 .
since we are most likely catching the qsos halfway through the current star formation episode , we infer a submm - detected qso phase lifetime of @xmath123myr .
the reader should of course note that @xmath120 and our estimated lifetime are both lower limits if @xmath41 includes a significant contribution from agn activity , or if the starburst is discontinuous , or if the imf is more top heavy than salpeter s imf , or if the gas reservoirs contain significant masses of sub - thermal gas @xcite , or if the co reservoir is widely distributed on scales @xmath124kpc ( not that there is any evidence that such galaxies exist ) . next we note that in our submm - detected qso sample co is typically detected in only the optically luminous qsos ( see fig . [
fig : selection ] ) , i.e. those qsos with the most massive bhs ( @xmath125m@xmath5 ; see fig .
[ fig : dave ] ) . as we discussed in [ bulge ] these bhs are much more massive than found in typical smgs is it feasible for the bhs to grow sufficiently between these two phases ? in fig .
[ fig : dave ] we plot @xmath46 as a function of @xmath19 for our sample of submm - detected qsos and for an ` average ' smg ( @xmath19 from @xcite and @xmath46 from @xcite ) .
we indicate gas consumption and bh growth evolutionary tracks for the average smg on timescales of 10 and 20myr , assuming a sfr of 1000m@xmath5yr@xmath32 and eddington - limited bh growth .
[ fig : dave ] demonstrates that an smg could evolve into a submm - detected qso with @xmath116m@xmath5 and @xmath126m@xmath5 ( the median @xmath46 of the co - undetected qsos ) in a reasonable timescale , whereas an average smg would need substantially ( @xmath13 times ) more gas and more time ( @xmath127 ) to evolve into the optically luminous submm - detected qsos in our sample with @xmath128m@xmath5 and @xmath129m@xmath5 .
the gas consumption and bh growth timescales suggest that the optically luminous submm - detected qsos in our sample can not be the direct descendents of typical smgs .
indeed the co - detected optically luminous qsos in our sample are extremely luminous agn ( @xmath130 , @xcite ) with @xmath125m@xmath5 and will most likely evolve into the most massive elliptical galaxies found in local rich clusters .
these qsos are so rare ( @xmath131 over all redshifts ; @xcite ) that they are unlikely to be associated with _ any _ known smgs , since @xmath132mjy smgs at @xmath1333 have much higher surface densities ( @xmath134 ; @xcite ) or space densities ( @xmath135mpc@xmath136 ; @xcite ) .
even correcting for the relative timescales of the two phases still leaves an order of magnitude difference in the space densities .
based on the relative number or space densities , these optically luminous submm - detected qsos are unlikely to be involved in the smg / qso evolutionary sequence we are interested in testing . in contrast , the less luminous submm - detected qsos appear to be consistent with the @xcite proposed evolutionary sequence , linking them to typical smgs ( see fig .
[ fig : dave ] ) .
thus we concentrate on the three submm - detected qsos ( including one marginal co detection and two upper limits : smmj131222.35 , smmj123716.01 and mmj163655 ) with bhs more similar to that of smgs of @xmath116m@xmath5 .
the median gas mass and sfr of this subsample are @xmath137m@xmath50 and 850m@xmath138 , respectively .
hence , these systems can sustain their current star formation episode for @xmath139myr or a total lifetime of the submm - detected qso phase of @xmath140myrs .
this estimate of the lifetime of submm - detected qsos can then be used to relate them to smgs as follows .
the incidence of blank - field submm sources identified with qsos is @xmath18% @xcite . in our test scenario
this is the fraction of an smg lifetime in which it appears as an submm - detected qso .
conversely , about 1525% of optically luminous qsos are detected in the submm @xcite .
assuming that all bright smgs go through a qso phase , and that all qsos are formed via smgs and there is no strong mass or luminosity dependence of their lifetimes , then these relative fractions imply that the total qso lifetime ( i.e. submm - detected plus submm - undetected ) is about @xmath141times shorter than the average lifetime of an smg . taking our estimated lifetime of less luminous submm - detected qsos , @xmath140myrs ,
these detection fractions imply total lifetimes for qsos of @xmath142myrs and for smgs @xmath143myrs .
these estimates are consistent with independent estimates of the lifetimes for qsos of @xmath14440myr ( @xcite ; @xcite ) and @xmath145300myr for smgs from modelling @xcite .
there are two other observational links between these qsos with @xmath146 ( @xmath147m@xmath5 ) and smgs .
firstly , the qsos have space densities at @xmath16 some 510 times lower than smgs ( @xcite ; @xcite ) .
this ratio of volume densities is roughly consistent with the ratio of the relative smg and qso lifetime , @xmath145300myrs and @xmath14840myrs , of @xmath108 times .
moreover , in [ bulge ] we found a hint that this subset of our submm - detected qsos lies on the present - day @xmath8 relation and similar to smgs , which is consistent with them being ` transition objects ' between smgs and qsos in the proposed evolutionary sequence in which they would eventually evolve into passively evolving spheroids . although this conclusion is based on a small sample of qsos ,
if the dynamical and gas properties of these three qsos are representative of this subset of submm - detected qsos , then our results are consistent with @xmath149 , @xmath116m@xmath5 submm - detected qsos being ` transition objects ' in the proposed evolutionary sequence , representing a very brief prodigious star formation phase where the bh is undergoing rapid growth synchronously with the stellar mass .
submm - detected qsos do not possess sufficiently large gas reservoirs ( containing less gas than smgs on average ) to sustain the implied sfrs for very long , providing a possible explanation for why some of our submm - detected qsos are undetected in co ( a single weak co detection and two non - detections of @xmath116m@xmath5 submm - detected qsos ) .
where do the optically luminous submm - detected qsos fit into this evolutionary scheme ?
the average smg would need @xmath150gyr ( more than its lifetime ) in order to grow their bhs by the required factor of @xmath151 to host a @xmath152m@xmath5 bh seen in the optically luminous subset of submm - detected qsos .
in addition , these smgs would need to begin with gas reservoirs containing @xmath153m@xmath5 in order to leave luminous qsos with detectable reservoirs of residual gas .
alternatively , these rare optically - luminous qsos could be periodically ` repeating ' smgs , with smg progenitors with @xmath154 and a surface density of @xmath155 , which are consuming their gas and growing their bhs synchronously , but which accumulate additional bh mass and gas through merging with other smgs in dense environments ( i.e. damp ` dry merging ' ; e.g. @xcite ) .
however , we currently have not observed any smgs with such large gas masses , and given their rarity , we will require large scale submm blank - field surveys such as those planned with scuba-2 to find these potential progenitor smgs .
hence our current results split the submm - detected qsos in our survey into two categories ( although a bigger sample would probably give a continuous trend ) : \(1 ) our observations of the optically luminous submm - detected qsos indicate they have @xmath156 .
it is very difficult to grow the bhs in typical smgs to such large masses in the typical lifetimes of smgs or submm - detected qsos .
we therefore suggest that these qsos can not be related to the evolutionary cycle of typical smgs . instead , these intrinsically rare qsos most likely evolve from an equally rare subset of the smg population
whose surface density is low enough that it has yet to be detected in current small - area submm surveys .
simply based upon the masses of their bhs , it is clear that these systems can only evolve into the most massive elliptical galaxies found in local rich clusters .
\(2 ) we have found a subset of our submm - detected qsos with lower optical luminosities which have gas and bh masses and space densities which are consistent with them being potential ` transition objects ' in the proposed evolutionary scenario where smgs evolve into qsos .
this subset of submm - detected qsos have @xmath157 and gas consumption timescales consistent with a phase of prodigious star formation where the bh has grown substantially during the preceding @xmath158 smg phase .
the fast build up of the stellar population is consistent with an evolutionary path ending in a population of present - day massive spheroids , where homogeneous old stellar populations are seen @xcite .
we have carried out a millimetre interferometry survey of nine submm - detected qsos at @xmath12.6 in order to test the link between smgs and qsos at the era where these two important populations were most numerous .
we include in our analysis comparable observations of a similarly selected qso from the literature to provide a final sample of ten submm - detected qsos . to support this survey we obtained near - infrared spectroscopy of these qsos to derive accurate systemic redshifts needed to tune the millimetre receivers to the correct frequencies .
the near - infrared spectra also provide h@xmath26 fluxes and line widths needed to derive reliable bh mass estimates for the qsos .
our main findings are : \(1 ) we detect co emission from six of the ten submm - detected qsos in our sample , confirming that they contain a significant amount of molecular gas and that a large fraction of the mm emission is from starbursts .
the median gas mass of our sample ( including non - detections ) is @xmath159m@xmath5 , similar to that found for @xmath16 smgs and to @xmath24 qsos .
the star formation efficiencies of our qsos are also comparable to those measured for smgs , @xmath63l@xmath65 , suggesting that the gross properties of the star formation in the qsos are like those seen in smgs . adopting a 2kpc scale size for the gas distribution in the qsos and a typical inclination of 20@xmath160
we derive a median dynamical mass of m@xmath98kpc@xmath161m@xmath5 , similar to smgs ( assuming an inclination angle appropriate for random inclinations ) .
we find a lower incidence of double - peaked co line profiles in the qsos , compared to smgs , which we believe results from a selection bias towards lower average inclination angles for the qsos .
\(2 ) our near - infrared spectroscopy indicates a median black hole mass in our qso sample of @xmath162m@xmath5 .
combined with our dynamical estimates of the spheroid mass , these yield @xmath163 .
this @xmath8 ratio for this sample of submm - detected qsos at @xmath7 is an order of magnitude larger than the local ratio , although @xmath105 suffers from large uncertainties due to the unknown co radii and inclination angles .
this ratio is also significantly above that seen for smgs at @xmath16 .
however , this comparison masks a broad range in bh masses within our qso sample and so we split the sample into two subsets based on their bh masses .
( 3 ) looking at the optically luminous submm - detected qsos in our sample we find that we detect co emission in 5/6 of these qsos .
however , the estimated bh masses for these qsos , @xmath16410@xmath4m@xmath50 , are too large ( and their number densities too small ) for them to be related to typical smgs in a simple evolutionary cycle .
we propose that the progenitors of these most massive qsos are a rare subset of smgs with @xmath165m@xmath5 with a number density of @xmath155 which will be possible to detect with future scuba-2 surveys .
\(4 ) for the optically less luminous ( @xmath149 ) submm - detected qsos , we marginally detect one source in co and obtain sensitive limits for three further qsos .
the bh masses for these systems are @xmath116m@xmath5 , similar to the estimates for bhs in smgs .
these submm - detected qsos are consistent with being ` transition ' objects between smgs and submm - undetected qsos , as we show it is feasible to link their bh masses to those of smgs by eddington limited growth for a period comparable to the gas depletion timescale of the qsos , @xmath107myrs .
the space density of these qsos is also in rough agreement with that expected for the descendents of smgs given current estimates of the relative lifetimes of qsos and smgs .
we conclude that these @xmath149 , @xmath166m@xmath5 submm - detected qsos are consistent with being in a very brief prodigious star formation phase , and that they simply do not possess sufficiently large gas reservoirs to sustain the sfr ( which is why these might be less often detected in co ) , although a larger sample of co observations of submm - detected qsos with these bh masses is required for confirmation .
to make further progress on understanding the evolutionary links between smgs and qsos requires a larger survey of the submm and co emission from typical qsos ( @xmath146 ) .
in addition , measurements of other co transitions for the submm - detected qsos ( e.g. from iram 30-m , alma , evla , and ska ) are required to place better constraints on the temperature and density of the molecular gas and thus provide a more accurate determination of the line luminosity ratios and hence total gas masses of these systems .
similarly , higher resolution co observations are essential to put strong constraints on the reservoir sizes and inclination angles , and hence @xmath47 , needed to compare the two populations .
finally , better measurements of the far - infrared seds ( with saboca , scuba-2 or _ herschel _ ) will yield more accurate measures of @xmath41 and @xmath167 for submm - detected qsos to constrain the contribution from an agn component .
this work is based on observations carried out with the iram pdbi .
iram is supported by insu / cnrs ( france ) , mpg ( germany ) and ign ( spain ) .
this work also makes use of data obtained at ukirt as part of programme i d u/06a/46 .
ukirt is operated by the joint astronomy centre on behalf on the uk science and technology facilities council ( stfc ) .
ned wright s javascript cosmology calculator was used in preparing this paper @xcite .
kekc and ams acknowledge support from the stfc .
dma and irs acknowledge support from the royal society .
we thank thomas greve for sending us the smg co spectra so that we could obtain the parameters of the double - gaussian profile fits .
we thank scott chapman and tadafumi takata for allowing us to use their lris / uv and ohs / near - ir spectra of smmj123716.01 , smmj131222.35 and mmj163655 .
we thank laura hainline , david frayer , and an anonymous referee for useful feedback which helped improve the clarity and presentation of the paper .
alexander d.m .
et al . , 2003 ,
aj , 125 , 383 alexander d.m . , smail i. , bauer f.e . ,
chapman s.c . ,
blain a.w .
, brandt w.n . , ivison r.j . , 2005a , nat . , 434 , 738 alexander d.m . ,
bauer f.e .
, chapman s.c . , smail i. , blain a.w .
, brandt w.n . , ivison r.j . , 2005b , apj , 632 , 736 alexander d.m .
et al . , 2008 ,
aj , 135 , 1968 avni y. , soltan a. , tananbaum h. , zamorani g. , 1980 , apj , 238 , 800 beelen a. et al . , 2004 , a&a , 423 , 441 beelen a. , cox p. , benford d. , dowell c.d .
, kovcs a. , bertoldi f. , omont a. , carilli c.l . , 2006 ,
apj , 642 , 694 bell e.f .
et al . , 2004 ,
apj , 608 , 752 biggs a.d . ,
ivison r.j . , 2006 , mnras , 371 , 963 bonning e.w .
, shields g.a .
, salviander s. , mclure r.j . , 2005 , apj , 626 , 89 borys c. , smail i. , chapman s.c . ,
blain a.w . ,
alexander d.m . ,
ivison r.j . , 2005 ,
apj , 635 , 853 bower r.g . ,
benson a.j . , malbon r. , helly j.c . , frenk c.s . ,
baugh c.m . , cole s. , lacey c.g . , 2006 , mnras , 370 , 645 boyle b.j .
, shanks t. , croom s.m . , smith r.j . ,
miller l. , loaring n. , heymans c. , 2000 , mnras , 317 , 1014 calzetti d. , armus l. , bohlin r.c .
, kinney a.l . , koornneef j. , storchi - bergmann t. , 2000 , apj , 533 , 682 carilli c.l . , wang r. , 2006 , aj , 131 , 2763 cavanagh b. , hirst p. , jenness t. , economou f. , currie m.j . , todd s. , ryder s.d . , 2003 , in payne h.e .
, jedrzejewski r.i . ,
, eds , asp conf . ser .
295 , astronomical data analysis software and systems xii , p. 237
chapman s.c . ,
blain a.w . ,
ivison r.j .
, smail i.r . , 2003 , nat . , 422 , 695 chapman s.c . ,
blain a.w . , smail i. , ivison r.j . , 2005 ,
apj , 622 , 772 coppin k. , et al . , 2006 , mnras , 372 , 1621 coppin k. , et al . , 2008 , mnras , 384 , 1597 di matteo t. , springel v. , hernquist l. , 2005 , nat . , 433 , 604 downes d. , solomon p.m. , 1998 , apj , 507 , 615 fan x. et al .
, 1999 , aj , 118 , 1 fan x. et al . , 2001 ,
aj , 122 , 2833 feigelson e.d .
, nelson p.i .
, 1985 , apj , 293 , 192 frayer d. t. , ivison r.j . , scoville n.z . , yun m. , evans a.s .
, smail i. , blain a. w. , kneib j .-
, 1998 , apj , 506 , l7 gebhardt k. et al . , 2000 , apj , 539 , 13 genzel r. , baker a.j .
, tacconi l.j . , lutz d. , cox p. , guilloteau s. , omont a. , 2003 , apj , 584 , 633 goncalves t.s . , steidel c.c .
, pettini m. , 2008 , apj , 676 , 816 greene j.e .
, ho l.c . , 2005 , apj , 630 , 122 greve t.r . , ivison r.j
. , bertoldi f. , stevens j.a . , dunlop j.s . , lutz d. , carilli c.l . , 2004 , mnras , 354 , 779 greve t.r
et al . , 2005 , mnras , 359 , 1165 greve t.r . , hainline l.j . ,
blain a.w . , smail i. , ivison r.j .
, papadopoulos p.p . , 2006 , aj , 132 , 1938 guilloteau s. , omont a. , cox p. , mcmahon r.g . , petitjean p. , 1999
, a&a , 349 , 363 guilloteau s. , lucas r. , 2000 , in mangum j.g .
, radford s.j.e .
, eds , asp conf . ser .
217 , imaging at radio through submillimeter wavelengths , p. 299
hainline l.j .
, scoville n.z .
, yun m.s . , hawkins d.w . ,
frayer d.t .
, isaak k.g . , 2004 ,
apj , 609 , 61 hring n. , rix h .- w . , 2004 ,
apj , 604 , l89 hopkins p.f . , hernquist l. , cox t.j . , di matteo t. , martini p. , robertson b. , springel v. , 2005 , apj , 630 , 705 humphrey p. j. , buote d. a. , brighenti f. , gebhardt k. , mathews w. g. , 2008 , preprint(astro - ph/0801.3461 ) ivison r.j . , et al . , 2007 , mnras , 380 , 199 kennicutt r.c . , 1998 , ara&a , 36 , 189 kovcs a. , chapman s.c . , dowell c.d . ,
blain a.w .
, ivison r.j . , smail i. , phillips t.g . , 2006 ,
apj , 650 , 592 krips m. , eckart a. , neri r. , bertram t. , straubmeier c , fischer s. , staguhn j.g . ,
vogel s.n . , a&a , 464 , 187 lutz d. et al . , 2007 , apj , 661 , l25 macchetto f. , marconi a.
, axon d. j. , capetti a. , sparks w. , crane p. , 1997 ,
apj , 489 , 579 magorrian j. et al .
, 1998 , aj , 115 , 2285 maiolino r. et al . , 2007 , a&a , 472 , l33 marconi a. , hunt l.k . , 2003 , apj , 589 , 21 martini p. , weinberg d.h . , 2001 ,
apj , 547 , 12 mclure r.j . , dunlop j.s . , 2004 , mnras , 352 , 1390 mclure r.j
, jarvis m.j . ,
targett t.a . , dunlop j.s .
, best p.n . , 2006 ,
mnras , 368 , 1395 nelson c.h . , 2000 , apj , 544 , l91 nelson c.h .
, whittle m. , 1996 , apj , 465 , 96 omont , a. , mcmahon r.g . ,
cox p. , kreysa e. , bergeron j. , pajot f. , storrie - lombardi l.j .
, 1996 , a&a , 315 , 1 omont , a. , cox p. , bertoldi f. , mcmahon r.g .
, carilli c. , isaak k.g . , 2001 , a&a , 374 , 371 omont , a. , beelen , a. , bertoldi f. , cox p. , carilli c.l . , priddey r.s . ,
mcmahon r.g .
, isaak k.g . , 2003 ,
a&a , 398 , 857 page m.j .
, stevens j.a .
, mittaz j.p.d . ,
carrera f.j . , 2001 ,
sci , 294 , 2516 page m.j .
, stevens j.a .
, ivison r.j . ,
carrera f.j . , 2004 ,
apj , 611 , l85 peng c.y . ,
impey c.d .
, ho l.c . , barton e.j .
, rix h .- w .
, 2006 , apj , 640 , 114 petric a.o .
, carilli c.l .
, bertoldi f. , beelen a. , cox p. , omont a. , 2006 , aj , 132 , 1307 ramsay howat s.k .
et al . , 2004 , in moorwood a.f.m .
, masanori i. , eds , spie conf .
proc . vol .
5492 , ground - based instrumentation for astronomy ,
riechers d.a .
et al . , 2006 ,
apj , 650 , 604 riechers d.a . , walter f. , brewer b.j . , carilli c.l .
, lewis g.f . , bertoldi f. , cox p. , 2008
, preprint(astro - ph/0806.4616 ) salpeter e.e . , 1955 , 121 , 161 sanders d.b .
, soifer b.t . , elias j.h . , neugebauer g. , matthews k. , 1988 , apj , 328 , l35 sanders d.b . , scoville n.z . ,
soifer b.t . , 1991 , apj , 370 , 158 seaquist e.r . , ivison r.j . ,
hall p.j . , 1995 , mnras , 276 , 867 shields g.a .
, menezes k.l . ,
massart c.a . , vanden bout p. , 2006
, apj , 641 , 683 smail i. , ivison r.j . , blain a.w . , 1997 ,
apjl , 490 , l5 smail i. , chapman s.c .
, blain a.w . ,
ivison r.j . , 2004 ,
apj , 616 , 71 stevens j.a .
, page m.j . ,
ivison r.j . , smail i. , carrera f.j . , 2004 , apj , 604
, l17 stevens j.a .
, page m.j . ,
ivison r.j . ,
carrera f.j . ,
mittaz j.p.d . , smail i. , mchardy i.m . , 2005 , mnras , 360 , 610 solomon p.m. , downes d. , radford s.j.e . ,
barrett j.w . , 1997 , apj , 478 , 144 solomon p.m. , vanden bout p.a .
, 2005 , ara&a , 43 , 677 spergel d.n .
et al . , 2003 , apjs , 148 , 175 staguhn j.g . , schinnerer e. , eckart a. , scharwchter j. , 2004 , apj , 609 , 85 swinbank , a. m. , smail , i. , chapman , s. c. , blain , a. w. , ivison , r. j. , keel , w. c. , 2004 , apj , 617 , 64 swinbank a. m. et al . , 2005 ,
mnras , 359 , 401 swinbank , a. m. , chapman , s. c. , smail , i. , lindner , c. , borys , c. , blain , a. w. , ivison r. j. , lewis , g. f. , 2006 , mnras , 371 , 465 tacconi l.j . , genzel r. , lutz d. , rigopoulou d. , baker a.j . , iserlohe c. , tecza m. , 2002 , apj , 580 , 73 tacconi l.j .
et al . , 2006 ,
apj , 640 , 228 tacconi l.j .
et al . , 2008 ,
apj , 680 , 246 takata t. , sekiguchi k. , smail i. , chapman s.c . , geach j.e . ,
swinbank a.m. , blain a. , ivison r.j .
, 2006 , apj , 651 , 713 valtonen m. j. , et al .
2007 , american astronomical society meeting abstracts , 211 , # 112.07 wall j.v .
, jenkins c.r . , 2003 , practical statistics for astronomers .
cambridge univ . press , cambridge walter f. , carilli c. , bertoldi f. , menten k. , cox p. , lo k.y . ,
fan x. , strauss m.a .
, 2004 , apj , 615 , l17 wang r. et al . , 2007 ,
aj , 134 , 617 wang r. et al .
, 2008 , aj , 135 , 1201 weiss a. , downes d. , neri r. , walter f. , henkel c. , wilder d.j . , wagg j. , wiklind t. , 2007 , a&a , 467 , 955 wright e.l .
, 2006 , pasp , 118 , 1711 yao l. , seaquist e.r . , kuno n. , dunne l. , 2003 , apj , 588 , 771 | we have used the iram plateau de bure millimeter interferometer and the ukirt 15@xmath0 m imager spectrometer ( uist ) to test the connection between the major phases of spheroid growth and nuclear accretion by mapping co emission in nine submillimetre - detected qsos at @xmath12.6 with black hole ( bh ) masses derived from near - infrared spectroscopy . when combined with one qso obtained from the literature
, we present sensitive co(32 ) or co(21 ) observations of 10 submillimetre - detected qsos selected at the epoch of peak activity in both qsos and submillimetre ( submm ) galaxies ( smgs ) .
co is detected in 5/6 very optically luminous ( @xmath2 ) submm - detected qsos with bh masses @xmath310@xmath4m@xmath5 , confirming the presence of large gas reservoirs of @xmath6m@xmath5 .
our bh masses and dynamical mass constraints on the host spheroids suggest , at face value , that these optically luminous qsos at @xmath7 lie about an order of magnitude above the local bh - spheroid relation , @xmath8 , although this result is dependent on the size and inclination of the co - emitting region . however , we find that their bh masses are @xmath9 times too large and their surface density is @xmath10 times too small to be related to typical smgs in an evolutionary sequence .
conversely , we measure weaker co emission in four fainter ( @xmath11 ) submm - detected qsos with properties , bh masses ( @xmath12m@xmath5 ) , and surface densities similar to smgs .
these qsos appear to lie near the local @xmath8 relation , making them plausible ` transition objects ' in the proposed evolutionary sequence linking qsos to the formation of massive young galaxies and bhs at high - redshift .
we show that smgs have a higher incidence of bimodal co line profiles than seen in our qso sample , which we interpret as an effect of their relative inclinations , with the qsos seen more face - on . finally , we find that the gas masses of the four fainter submm - detected qsos imply that their star formation episodes could be sustained for @xmath13myr , and are consistent with representing a phase in the formation of massive galaxies which overlaps a preceding smg starburst phase , before subsequently evolving into a population of present - day massive ellipticals .
galaxies : high - redshift galaxies : evolution galaxies : formation galaxies : kinematics and dynamics quasars : emission lines submillimetre |
given a set of skill requirements ( called task @xmath0 ) , a set of experts who have expertise in one or more skill , along with a social or professional network of the experts , the team formation problem is to identify a competent and highly collaborative team .
this problem in the context of a social network was first introduced by @xcite and has attracted recent interest in the data mining community @xcite .
a closely related and well - studied problem in operations research is the assignment problem . here ,
given a set of agents and a set of tasks , the goal is to find an agent - task assignment minimizing the cost of the assignment such that exactly one agent is assigned to a task and every task is assigned to some agent .
this problem can be modeled as a maximum weight matching problem in a weighted bipartite graph .
in contrast to the assignment problem , the team formation problem considers the underlying social network , which for example models the previous collaborations among the experts , while forming teams .
the advantage of using such a social network is that the teams that have worked together previously are expected to have less communication overhead and work more effectively as a team .
the criteria explored in the literature so far for measuring the effectiveness of teams are based on the shortest path distances , density , and the cost of the minimum spanning tree of the subgraph induced by the team . here
the density of a subgraph is defined as the ratio of the total weight of the edges within the subgraph over the size of the subgraph .
teams that are well connected have high density values .
methods based on minimizing diameter ( largest shortest path between any two vertices ) or cost of the spanning tree have the main advantage that the teams they yield are always connected ( provided the underlying social network is connected )
. however , diameter or spanning tree based objectives are not robust to the changes ( addition / deletion of edges ) in the social network .
as demonstrated in @xcite using various performance measures , the density based objective performs better in identifying well connected teams . on the other hand
, maximizing density may give a team whose subgraph is disconnected .
this happens especially when there are small groups of people who are highly connected with each other but are sparsely connected to the rest of the graph .
existing methods make either strong assumptions on the problem that do not hold in practice or are not capable of incorporating more intuitive constraints such as bounding the total size of the team .
the goal of this paper is to consider the team formation problem in a more realistic setting and present a novel formulation based on a generalization of the densest subgraph problem .
our formulation allows modeling of many realistic requirements such as ( i ) inclusion of a designated team leader and/or a group of given experts , ( ii ) restriction on the size or more generally cost of the team ( iii ) enforcing _ locality _ of the team , e.g. , in a geographical sense or social sense , etc .
in fact most of the future directions pointed out by @xcite are covered in our formulation .
the first work @xcite in the team formation problem in the presence of a social network presents greedy algorithms for minimizing the diameter and the cost of the minimum spanning tree ( mst ) induced by the team .
while the greedy algorithm for minimizing the diameter has an approximation guarantee of two , no guarantee is proven for the mst algorithm . however ,
@xcite impose the strong assumption that a skill requirement of a task can be fulfilled by a single person ; thus a more natural requirement such as `` at least @xmath4 experts of skill @xmath5 are needed for the task '' can not be handled by their method .
this shortcoming has been addressed in @xcite , which presents a 2-approximation algorithm for a slightly more general problem that can accommodate the above requirement .
however , both algorithms can not handle an upper bound constraint on the team size . on the other hand ,
the solutions obtained by all these algorithms ( including the mst algorithm ) can be shown to be connected subgraphs if the underlying social graph is connected .
two new formulations are proposed in @xcite based on the shortest path distances between the nodes of the graph .
the first formulation assumes that experts from each skill have to communicate with every expert from the other skill and thus minimizes the sum of the pairwise shortest path distances between experts belonging to different skills .
they prove that this problem is np - hard and provide a greedy algorithm with an approximation guarantee of two .
the second formulation , solvable optimally in polynomial time , assumes that there is a designated team leader who has to communicate with every expert in the team and minimizes the sum of the distances only to the leader .
the main shortcoming of this work is its restrictive assumption that _ exactly _ one expert is sufficient for each skill , which implies that the size of the found teams is always upper bounded by the number of skills in the given task , noting that an expert is allowed to have multiple skills .
they exploit this assumption and ( are the first to ) produce top-@xmath4 teams that can perform the given task .
however , although based on the shortest path distances , neither of the two formulations does guarantee that the solution obtained is connected .
in contrast to the distance or diameter based cost functions , @xcite explore the usefulness of the density based objective in finding strongly connected teams . using various performance measures ,
the superiority of the density based objective function over the diameter objective is demonstrated .
the setting considered in @xcite is the most general one until now but the resulting problem is shown to be np hard . the greedy algorithms that they propose have approximation guarantees ( of factor 3 ) for two special cases .
the teams found by their algorithms are often quite large and it is not straightforward to modify their algorithms to integrate an additional upper bound constraint on the team size .
another disadvantage is that subgraphs that maximize the density under the given constraints need not necessarily be connected .
recently @xcite considered an _
team formation problem where tasks arrive in a sequential manner and teams have to be formed minimizing the ( maximum ) load on any expert across the tasks while bounding the coordination cost ( a free parameter ) within a team for any given task .
approximation algorithms are provided for two variants of coordinate costs : diameter cost and steiner cost ( cost of the minimum steiner tree where the team members are the terminal nodes ) .
while this work focusses more on the load balancing aspect , it also makes the strong assumption that a skill is covered by the team if there exists at least one expert having that skill .
all of the above methods allow only binary skill level , i.e. , an expert has a skill level of either one or zero .
we point out that many methods have been developed in the operations research community for the team formation problem , @xcite , but none of them explicitly considers the underlying social or professional connections among the experts .
there is also literature discussing the social aspects of the team formation @xcite and their influence on the evolution of communities , e.g. , @xcite .
now we formally define the _ team formation _ problem that we address in this paper .
let @xmath1 be the set of @xmath6 experts and @xmath7 be the weighted , undirected graph reflecting the relationship or previous collaboration of the experts @xmath1 .
then non - negative , symmetric weight @xmath8 connecting two experts @xmath9 and @xmath10 reflects the level of compatibility between them .
the set of skills is given by @xmath11 .
each expert is assumed to possess one or more skills .
the non - negative matrix @xmath12 specifies the skill levels of all experts in each skill .
note that we define the skill level on a continuous scale . if an expert @xmath9 does not have skill @xmath10 , then @xmath13 .
moreover , we use the notation @xmath14 for the @xmath15th column of @xmath16 , i.e. the vector of skill levels corresponding to skill @xmath10 . a task @xmath0
is given by the set of triples @xmath17 , where @xmath18 , specifying that at least @xmath19 and at most @xmath20 of skill @xmath21 is required to finish the given task .
+ * generalized team formation problem . * given a task @xmath0 , the generalized team formation problem is defined as finding a team @xmath3 of experts maximizing the _ collaborative compatibility _ and satisfying the following constraints : * * inclusion of a specified group : * a predetermined group of experts @xmath22 should be in @xmath23 . *
* skill requirement : * at least @xmath19 and at most @xmath20 of skill @xmath21 is required to finish the task @xmath0 . * * bound on the team size : * the size of the team should be smaller than or equal to @xmath24 , i.e. , @xmath25 . *
* budget constraint : * total budget for finishing the task is bounded by @xmath26 , i.e. , @xmath27 , where @xmath28 is the cost incurred on expert @xmath9 . *
* distance based constraint : * the distance ( measured according to some non - negative , symmetric function , @xmath29 ) between any pair of experts in @xmath23 should not be larger than @xmath30 , i.e. , @xmath31 .
* discussion of our generalized constraints . * in contrast to existing methods , we also allow an upper bound on each skill and on the total team size . if the skill matrix is only allowed to be binary as in previous work , this translates into upper and lower bounds on the number of experts required for each skill . using vertex weights
, we can in fact encode more generic constraints , e.g. , having a limit on the total budget of the team .
it is not straightforward to extend existing methods to include any upper bound constraints .
up to our knowledge we are the first to integrate upper bound constraints , in particular on the size of the team , into the team formation problem .
we think that the latter constraint is essential for realistic team formation .
our general setting also allows a group of experts around whom the team has to be formed .
this constraint often applies as the team leader is usually fixed before forming the team .
another important generalization is the inclusion of _ distance _
constraints for any general distance function .
such a constraint can be used to enforce locality of the team e.g. in a geographical sense ( the distance could be travel time ) or social sense ( distance in the network ) . another potential application
are mutual incompatibilities of team members e.g.on a personal level , which can be addressed by assigning a high distance to experts who are mutually incompatible and thus should not be put together in the same team .
we emphasize that all constraints considered in the literature are special instances of the above constraint set . + * measure of collaborative compatiblity . * in this paper we use as a measure of collaborative compatibility a generalized form of the density of subgraphs , defined as @xmath32 where @xmath33 is the non - negative weight of the edge between @xmath9 and @xmath10 and @xmath34 is defined as @xmath35 , with @xmath36 being the positive weight of the vertex @xmath9 .
we recover the original density formulation , via @xmath37 .
we use the relation , @xmath38 , where @xmath39 is the degree of vertex @xmath9 and @xmath40 . + * discussion of density based objective . * as pointed out in @xcite , the density based objective possesses useful properties like strict monotonicity and robustness . in case of the density based objective , if an edge gets added ( because of a new collaboration ) or deleted ( because of newly found incompatibility ) the density of the subgraphs involving this edge necessarily increases resp.decreases , which is not true for the diameter based objective .
in contrast to density based objective , the impact of small changes in graph structure is more severe in the case of diameter objective @xcite .
the generalized density that we use here leads to further modeling freedom as it enables to give weights to the experts according to their expertise . by giving smaller weight to those with high expertise
, one can obtain solutions that not only satisfy the given skill requirements but also give preference to the more competent team members ( i.e.the ones having smaller weights )
. + * problem formulation . * using the notation introduced above , an instance of the team formation problem based on the generalized density can be formulated as @xmath41 note that the upper bound constraints on the team size and the budget can be rewritten as skill constraints and can be incorporated into the skill matrix @xmath16 accordingly .
thus , without loss of generality , we omit the budget and size constraints from now on , for the sake of brevity .
moreover , since @xmath42 is required to be part of the solution , we can assume that @xmath43 , otherwise the above problem is infeasible .
the distance constraint also implies that any @xmath44 for which @xmath45 , for some @xmath46 , can not be a part of the solution .
thus , we again assume wlog that there is no such @xmath44 ; otherwise such vertices can be eliminated without changing the solution of problem .
+ our formulation is a generalized version of the classical densest subgraph problem ( dsp ) , which has many applications in graph analysis , e.g. , see @xcite .
the simplest version of dsp is the problem of finding a densest subgraph ( without any constraints on the solution ) , which can be solved optimally in polynomial time @xcite .
the densest-@xmath4-subgraph problem , which requires the solution to contain exactly @xmath4 vertices , is a notoriously hard problem in this class and has been shown not to admit a polynomial time approximation scheme @xcite .
recently , it has been shown that the densest subgraph problem with an upper bound on the size is as hard as the densest-@xmath4-subgraph problem @xcite .
however , the densest subgraph problem with a lower bound constraint has a 2-approximation algorithm @xcite .
it is based on solving a sequence of unconstrained densest subgraph problems .
they also show that there exists a linear programming relaxation for this problem achieving the same approximation guarantee .
recently @xcite considered the following generalized version of the densest subgraph problem with lower bound constraints in the context of team formation problem : @xmath47 where @xmath16 is the _ binary _ skill matrix .
they extend the greedy method of @xcite and show that it achieves a 3-approximation guarantee for some special cases of this problem .
@xcite recently improved the approximation guarantee of the greedy algorithm of @xcite for problem ( [ eq : teamproblb ] ) to a factor 2 .
the time complexity of this greedy algorithm is @xmath48 , where @xmath6 is the number of experts and @xmath49 is the minimum number of experts required . + * direct integration of subset constraint .
* the subset constraint can be integrated into the objective by directly working on the subgraph @xmath50 induced by the vertex set @xmath51 .
note that any @xmath52 that contains @xmath42 can be written as @xmath53 , for @xmath54 .
we now reformulate the team formation problem on the subgraph @xmath50 .
we introduce the notation @xmath55 , and we assume wlog that the first @xmath56 entries of @xmath1 are the ones in @xmath57 . the terms in problem (
[ eq : teamprob ] ) can be rewritten as @xmath58 moreover , note that we can write : @xmath59 , where @xmath60 denotes the degree of vertex @xmath9 restricted to the subset @xmath42 in the original graph . using the abbreviations , @xmath61 , @xmath62 , @xmath63 , we rewrite the team formation problem as @xmath64 where for all @xmath65
, the bounds were updated as @xmath66 .
note that here we already used the assumption : @xmath67 .
the constraint , @xmath68 , has been introduced for technical reasons required for the formulation of the continuous problem in section [ sec : equiv_cont_prob ] .
the equivalence of problem to follows by considering either @xmath42 ( if feasible ) or the set @xmath69 , where @xmath70 is an optimal solution of , depending on whichever has higher density . to the best of our knowledge there is no greedy algorithm with an approximation guarantee to solve problem ( [ eq : teamprobsim ] ) . instead of designing a greedy approximation algorithm for this discrete optimization problem
, we derive an _
equivalent _ continuous optimization problem in section [ sec : forte ] .
that is , we reformulate the discrete problem in continuous space while preserving the optimality of the solutions of the discrete problem .
the rationale behind this approach is that the continuous formulation is more flexible and allows us to choose from a larger set of methods for its solution than for the discrete one .
although the resulting continuous problem is as hard as the original discrete problem , recent progress in continuous optimization @xcite allow us to find a locally optimal solution very efficiently .
in this section we present our method , _ formation of realistic teams _ ( , for short ) to solve the team formation problem , which is rewritten as , using the continuous relaxation .
we derive in three steps : 1 .
derive an equivalent unconstrained discrete problem ( [ eq : teamprobuncstr ] ) of the team formation problem ( [ eq : teamprobsim ] ) via an _ exact penalty _ approach .
2 . derive an equivalent continuous relaxation ( [ eq : teamprobcont ] ) of the unconstrained problem by using the concept of _ lovasz extensions_. 3 .
compute the solution of the continuous problem ( [ eq : teamprobcont ] ) using the recent method ratiodca from _
fractional programming_. a general technique in constrained optimization is to transform the constrained problem into an equivalent unconstrained problem by adding to the objective a penalty term , which is controlled by a parameter @xmath71 .
the penalty term is zero if the constraints are satisfied at the given input and strictly positive otherwise .
the choice of the regularization parameter @xmath72 influences the tradeoff between satisfying the constraints and having a low objective value .
large values of @xmath72 tend to enforce the satisfaction of constraints . in the following
we show that for the team formation problem ( [ eq : teamprobsim ] ) there exists a value of @xmath72 that guarantees the satisfaction of all constraints .
+ let us define the penalty term for constraints of the team formation problem ( [ eq : teamprobsim ] ) as @xmath73 note that the above penalty function is zero only when @xmath74 satisfies the constraints ; otherwise it is strictly positive and increases with increasing infeasibility .
the special treatment of the empty set is again a technicality required later for the lovasz extensions , see section [ sec : equiv_cont_prob ] .
for the same reason , we also replace the constant terms @xmath75 and @xmath76 in ( [ eq : teamprobsim ] ) by @xmath77 and @xmath78 respectively , where @xmath79 and @xmath80 .
+ the following theorem shows that there exists an unconstrained problem equivalent to the constrained optimization problem ( [ eq : teamprobsim ] ) .
[ thm : teamprobuncstr ] the constrained problem ( [ eq : teamprobsim ] ) is equivalent to the unconstrained problem @xmath81 for @xmath82 , where @xmath83 is any feasible set of problem ( [ eq : teamprobsim ] ) such that @xmath84 and @xmath85 is the minimum value of infeasibility , i.e. , @xmath86 , if @xmath74 is infeasible .
we define @xmath87 .
note that maximizing ( [ eq : teamprobsim ] ) is the same as minimizing @xmath88 subject to the constraints of ( [ eq : teamprobsim ] ) . for any feasible subset @xmath74 ,
the objective of ( [ eq : teamprobuncstr ] ) is equal to @xmath88 , since the penalty term is zero .
thus , if we show that _ all _ minimizers of ( [ eq : teamprobuncstr ] ) satisfy the constraints then the equivalence follows .
suppose , for the sake of contradiction , that @xmath89 , if @xmath90 ) is a minimizer of ( [ eq : teamprobuncstr ] ) and that @xmath91 is infeasible for problem ( [ eq : teamprobsim ] ) .
since @xmath92 and @xmath93 , we have under the given condition on @xmath72 , @xmath94 which leads to a contradiction because the last term is the objective value of ( [ eq : teamprobuncstr ] ) at @xmath83
. we will now derive a tight continuous relaxation of problem ( [ eq : teamprobuncstr ] ) .
this will lead us to a minimization problem over @xmath95 , which then can be handled more easily than the original discrete problem .
the connection between the discrete and the continuous space is achieved via thresholding . given a vector @xmath96
, one can define the sets @xmath97 by thresholding @xmath98 at the value @xmath99 . in order to go from functions on sets to functions on continuous space , we make use of the concept of lovasz extensions .
( lovasz extension)[def : lovasz ] let @xmath100 be a set function with @xmath101 , and let @xmath102 be ordered in ascending order @xmath103 .
the lovasz extension @xmath104 of @xmath105 is defined by @xmath106 note that @xmath107 for all @xmath108 , i.e.@xmath109 is indeed an extension of @xmath105 from @xmath110 to @xmath111 ( @xmath112 ) . in the following , given a set function @xmath113 , we will denote its lovasz extension by @xmath109 .
the explicit forms of the lovasz extensions used in the derivation will be dealt with in section [ sec : algo ] . in the following theorem
we show the equivalence for [ eq : teamprobsim ] .
a more general result showing equivalence for fractional set programs can be found in @xcite .
[ thm : teamprobcont ] the unconstrained discrete problem ( [ eq : teamprobuncstr ] ) is equivalent to the continuous problem @xmath114 for any @xmath115 .
moreover , optimal thresholding of a minimizer @xmath116 , @xmath117 yields a set @xmath70 that is optimal for problem ( [ eq : teamprobuncstr ] ) .
let @xmath118 .
then we have @xmath119 where in the first step we used the fact that @xmath120 and @xmath121 are extensions of @xmath122 and @xmath123 , respectively .
below we first show that the above inequality also holds in the other direction , which then establishes that the optimum values of both problems are the same .
the proof of the reverse direction will also imply that a set minimizer of the problem ( [ eq : teamprobuncstr ] ) can be obtained from any minimizer @xmath124 of ( [ eq : teamprobcont ] ) via optimal thresholding .
+ we first show that the optimal thresholding of any @xmath125 yields a set @xmath74 such that @xmath126 has an objective value at least as good as the one of @xmath98 .
this holds because @xmath127 the third step follows from the fact that @xmath98 is non - negative ( @xmath128 ) and ordered in ascending order , i.e. , @xmath129 .
since @xmath130 is non - negative , the final step implies that @xmath131 thus we have @xmath132 from inequality ( [ eq : optthres ] ) , it follows that optimal thresholding of @xmath124 yields a set that is a minimizer of problem ( [ eq : teamprobuncstr ] ) .
the team formation problem ( [ eq : teamprobsim ] ) is equivalent to the problem ( [ eq : teamprobcont ] ) if @xmath72 is chosen according to the condition given in theorem [ thm : teamprobuncstr ] .
this directly follows from theorems [ thm : teamprobuncstr ] and [ thm : teamprobcont ] .
while the continuous problem is as hard as the original discrete problem , recent ideas from continuous optimization @xcite allow us to derive in the next section an algorithm for obtaining locally optimal solutions very efficiently .
we now describe an algorithm for ( approximately ) solving the continuous optimization problem ( [ eq : teamprobcont ] ) .
the idea is to make use of the fact that the fractional optimization problem ( [ eq : teamprobcont ] ) has a special structure : as we will show in this section , it can be written as a special ratio of difference of convex ( d.c . ) functions , i.e. it has the form @xmath133 where the functions @xmath134 and @xmath135 are positively one - homogeneous convex functions is said to be positively one - homogeneous if @xmath136 . ] and numerator and denominator are nonnegative .
this reformulation then allows us to use a recent first order method called ratiodca @xcite . in order to find the explicit form of the convex functions , we first need to rewrite the penalty term as @xmath137 , where @xmath138 using this decomposition of @xmath139
, we can now write down the functions @xmath134 and @xmath135 as @xmath140 where @xmath141 , @xmath142 , @xmath143 denotes the lovasz extension of @xmath144 , and @xmath145 using the functions @xmath134 and @xmath135 defined above , the problem ( [ eq : teamprobcont ] ) can be rewritten in the form .
the functions @xmath134 and @xmath135 are convex and positively one - homogeneous , and @xmath146 and @xmath147 are nonnegative .
the denominator of is given as @xmath148 and the numerator is given as @xmath149 using prop.2.1 in @xcite and the decomposition of @xmath139 introduced earlier in this section , we can decompose @xmath150 .
the lovasz extension of @xmath151 is given as @xmath152 and let @xmath143 denote the lovasz extension of @xmath144 ( an explicit form is not necessary , as shown later in this section ) .
the equality between and then follows by simple rearranging of the terms .
the nonnegativity of the functions @xmath153 and @xmath154 follows from the nonnegativity of denominator and numerator of and the definition of the lovasz extension .
moreover , the lovasz extensions of any set function is positively one - homogeneous @xcite .
finally , the convexity of @xmath155 and @xmath156 follows as they are a non - negative combination of the convex functions @xmath157 and @xmath158 for some @xmath159 .
the function @xmath160 is well - known to be convex @xcite . to show the convexity of @xmath161
, we will show that the function @xmath144 is submodular is submodular if for all @xmath162 , @xmath163 .
the convexity then follows from the fact that a set function is submodular if and only if its lovasz extension is convex @xcite . for the proof of the submodularity of the first two sums one uses the fact that the pointwise minimum of a constant and a increasing submodular function is again submodular .
writing @xmath164 , the last sum can be written as @xmath165 . using @xmath166 ,
we can write its lovasz extension as @xmath167 which is a sum of a linear term and a convex term . the reformulation of the problem in the form enables us to apply a modification of the recently proposed ratiodca @xcite , a method for the _ local _ minimization of objectives of the form on the whole @xmath95 .
@xmath168 , @xmath169 @xmath170 + @xmath171 , @xmath172 @xmath173 @xmath174 given an initialization @xmath175 , the above algorithm solves a sequence of convex optimization problems ( line 3 ) .
note that we do not need an explicit description of the terms @xmath176 and @xmath177 , but only elements of their sudifferential @xmath178 resp . @xmath179 .
the explicit forms of the subgradients are given in the appendix .
the convex problem ( line 3 ) then has the form @xmath180 where @xmath181 .
note that is a _ non - smooth _ problem .
however , there exists an equivalent smooth dual problem , which we give below .
[ lem : inner_problem ] the problem is equivalent to @xmath182 where @xmath183 with @xmath184 , @xmath185 denotes the projection on the positive orthant and @xmath186 is the simplex @xmath187 .
first we use the homogenity of the objective in the inner problem to eliminate the norm constraint .
this yields the equivalent problem @xmath188 we derive the dual problem as follows : @xmath189 where @xmath184 .
the optimization over @xmath190 has the solution @xmath191 plugging @xmath190 into the objective and using that @xmath192 , we obtain the result
. the smooth dual problem can be solved very efficiently using recent scalable first order methods like fista @xcite , which has a guaranteed convergence rate of @xmath193 , where @xmath4 is the number of steps done in fista .
the main part in the calculation of fista consists of a matrix - vector multiplication . as the social network is typically sparse , this operation costs @xmath194 , where @xmath56 is the number of non - zeros of @xmath195 .
ratiodca @xcite , produces a strictly decreasing sequence @xmath196 , i.e. , @xmath197 , or terminates .
this is a typical property of fast local methods in non - convex optimization .
moreover , the convex problem need not be solved to full accuracy ; we can terminate the convex problem early , if the current @xmath196 produces already sufficent descent in @xmath198 .
as the number of required steps in the ratiodca typically ranges between 5 - 20 , the full method scales to large networks .
note that convergence to the global optimum of ( [ eq : fracset ] ) can not be guaranteed due to the non - convex nature of the problem . however , we have the following quality guarantee for the team formation problem .
[ th : quality - guarantee ] let @xmath83 be a feasible set for the problem and @xmath72 is chosen as in theorem [ thm : teamprobuncstr ] .
let @xmath199 denote the result of ratiodca after initializing with the vector @xmath200 , and let @xmath201 denote the set found by optimal thresholding of @xmath199 .
either ratiodca terminates after one iteration , or produces @xmath201 which satisfies all the constraints of the team formation problem ( [ eq : teamprobsim ] ) and @xmath202 ratiodca generates a decreasing sequence @xmath203 such that @xmath197 until it terminates @xcite .
we have @xmath204 , if the algorithm does not stop in one step . as shown in theorem ( [ thm : teamprobcont ] ) optimal thresholding of @xmath205 yields a set @xmath206 that achieves smaller objective on the corresponding set function .
since the chosen value of @xmath72 guarantees the satisfaction of the constraints , @xmath207 has to be feasible .
recall that our team formation problem based on the density objective is rewritten as the following gdsp after integrating the subset constraint : @xmath208 note that here we do not require the additional constraint , @xmath68 , that we added to . in this section
we show that there exists a linear programming ( lp ) relaxation for this problem .
the lp relaxation can be solved optimally in polynomial time and provides an upper bound on the optimum value of gdsp .
in practice such an upper bound is useful to check the quality of the solutions found by approximation algorithms . + the following lp is a relaxation of the generalized densest subgraph problem ( [ eq : teamprobsim2 ] ) .
@xmath209 where @xmath210 , @xmath211 is the set of edges induced by @xmath212 .
the following problem is equivalent to ( [ eq : teamprobsim2 ] ) , because ( i ) for every feasible set @xmath74 of , there exist corresponding feasible @xmath213 given by @xmath214 , @xmath215 , with the same objective value and ( ii ) an optimal solution of the following problem always satisfies @xmath216 .
@xmath217 relaxing the integrality constraints and using the substitution , @xmath218 and @xmath219 , we obtain the relaxation : @xmath220 since this problem is invariant under scaling , we can fix the scale by setting the denominator to 1 , which yields the equivalent lp stated in the theorem .
note that the solution @xmath124 of the lp ( [ eq : lp ] ) is , in general , not integral , i.e. , @xmath221 .
one can use standard techniques of randomized rounding or optimal thresholding to derive an integral solution from @xmath124 .
however , the resulting integral solution may not necessarily give a subset that satisfies the constraints of .
in the special case when there are only lower bound constraints , i.e. , problem ( [ eq : teamproblb ] ) , one can obtain a feasible set @xmath74 for problem ( [ eq : teamproblb ] ) by thresholding @xmath124 ( see ) according to the objective of .
this is possible in this special case because there is always a threshold @xmath222 which yields a non - empty subset @xmath223 ( in the worst case the full set @xmath212 ) satisfying all the lower bound constraints . in our experiments on problem , we derived a feasible set from the solution of lp in this fashion by choosing the threshold that yields a subset that satisfies the constraints and has the highest objective value .
note that the lp relaxation ( [ eq : lp ] ) is vacuous with respect to upper bound constraints in the sense that given @xmath224 that does not satisfy the upper bound constraints of the lp ( [ eq : lp ] ) one can construct @xmath225 , feasible for the lp by rescaling @xmath98 without changing the objective of the lp .
this implies that one can always transform the solution of the unconstrained problem into a feasible solution when there are @xmath226 upper bound constraints .
however , in the presence of lower bound or subset constraints , such a rescaling does not yield a feasible solution and hence the lp relaxation is useful on the instances of with at least one lower bound or a subset constraint ( i.e. , @xmath227 ) .
we now empirically show that consistently produces high quality compact teams .
we also show that the quality guarantee given by theorem [ th : quality - guarantee ] is useful in practice as our method often improves a given sub - optimal solution .
since we are not aware of any publicly available real world datasets for the team formation problem , we use , as in @xcite , a scientific collaboration network extracted from the dblp database .
similar to @xcite , we restrict ourselves to four fields of computer science : databases ( ) , theory ( ) , data mining ( ) , artificial intelligence ( ) . conferences that we consider for each field are given as follows : db = \{sigmod , vldb , icde , icdt , pods } , t = \{soda , focs , stoc , stacs , icalp , esa } , dm = \{www , kdd , sdm , pkdd , icdm , wsdm } , ai = \{ijcai , nips , icml , colt , uai , cvpr}. for our team formation problem , the skill set is given by @xmath228\ { , , , } .
any author who has at least three publications in any of the above 23 conferences is considered to be an expert . in our dblp co -
author graph , a vertex corresponds to an expert and an edge between two experts indicates prior collaboration between them .
the weight of the edge is the number of shared publications .
since the resulting co - author graph is disconnected , we take its largest connected component ( of size 9264 ) for our experiments . directly solving
the non - convex problem ( [ eq : teamprobcont ] ) for the value of @xmath72 given in theorem [ thm : teamprobuncstr ] often yields poor results .
hence in our implementation of we adopt the following strategy .
we first solve the unconstrained version of problem ( [ eq : teamprobcont ] ) ( i.e. , @xmath229 ) and then iteratively solve ( [ eq : teamprobcont ] ) for increasing values of @xmath72 until all constraints are satisfied . in each iteration ,
we increase @xmath72 only for those constraints which were infeasible in the previous iteration ; in this way , each penalty term is regulated by different value of @xmath72 . moreover , the solution obtained in the previous iteration of @xmath72 is used as the starting point for the current iteration . [
fig : times ] -0.2 cm in this section we perform a quantitative evaluation of our method in the special case of the team formation problem with lower bound constraints and @xmath230 ( problem ) .
we evaluate the performance of our method against the greedy method proposed in @xcite , refered to as .
similar to the experiments of @xcite , an expert is defined to have a skill level of 1 in skill @xmath10 , if he / she has a publication in any of the conferences corresponding to the skill @xmath10 .
as done in @xcite , we create random tasks for different values of skill size , @xmath231 . for each value of @xmath4
we sample @xmath4 skills with replacement from the skill set @xmath232 = \ { , , , } .
for example if @xmath233 , a sample might contain \ { , , } , which means that the random task requires at least two experts from the skill and one expert from the skill . in figure 1 , we show for each method the densities , sizes and runtimes for the different skill sizes @xmath4 , averaged over 10 random runs . in the first plot , we also show the optimal values of the lp relaxation in .
note that this provides an upper bound on the optimal value of .
we can obtain feasible solutions from the lp relaxation of via thresholding ( see section [ sec : lprel ] ) , which are shown in the plot as .
furthermore , the plots contain the results obtained when the solutions of and are used as the initializations for ( in each of the @xmath72 iteration ) .
the plots show that always produces teams of higher densities and smaller sizes compared to and .
furthermore , produces better results than the greedy method in several cases in terms of densities and sizes of the obtained teams .
the results of + and + further show that our method is able improve the sub - optimal solutions of and significantly and achieves almost similar results as that of which was started with the unconstrained solution of . under the worst - case assumption that the upper bound on computed using the lp is the optimal value , the solution of is @xmath234 optimal ( depending on @xmath4 ) .
[ tb : qualteams ] [ cols= " < , < ,
< , < " , ] [ table : teams ] -0.2 cm in this experiment , we assess the quality of the teams obtained for several tasks with different skill requirements .
here we consider the team formation problem ( [ eq : teamprobsim ] ) in its more general setting .
we use the generalized density objective of where each vertex is given a rank @xmath235 , which we define based on the number of publications of the corresponding expert . for each skill
, we rank the experts according to the number of his / her publications in the conferences corresponding to the skill . in this way
each expert gets four different rankings ; the total rank of an expert is then the minimum of these four ranks .
the main advantage of such a ranking is that the experts that have higher skill are given preference , thus producing more competent teams .
note that we choose a relative measure like rank as the vertex weights instead of an absolute quantity like number of publications , since the distribution of the number of publications varies between different fields .
in practice such a ranking is always available and hence , in our opinion , should be incorporated .
furthermore , in order to identify the main area of expertise of each expert , we consider his / her relative number of publications . each expert is defined to have a skill level of 1 in skill @xmath10 if he has more than 25% of his / her publications in the conferences corresponding to skill @xmath10 . as a distance function between authors ,
we use the shortest path on the _ unweighted version _ of the dblp graph , i.e. two experts are at a distance of two , if the shortest path between the corresponding vertices in the unweighted dblp graph contains two edges . note that in general the distance function can come from other general sources beyond the input graph , but here we had to rely on the graph distance because of lack of other information . in order to assess the _ competence _ of the found teams , we use the list of the 10000 most cited authors of citeseer @xcite . note that in contrast to the skill - based ranking discussed above , this list is only used in the evaluation and _ not _ in the construction of the graph .
we compute the average inverse rank as in @xcite as @xmath236 , where @xmath4 is the size of the team and @xmath237 is the rank of expert @xmath9 on the citeseer list of 10000 most cited authors
. for authors not contained on the list we set @xmath238 .
we also report the densities of the teams found in order to assess their _
compatibility_. we create several tasks with various constraints and compare the teams produced by , and ( feasible solution derived from the lp relaxation ) .
note that in our implementation we extended the algorithm of @xmath239 to incorporate general vertex weights , using dinkelbach s method from fractional programming @xcite .
the results for these tasks are shown in table 1 .
we report the upper bound given by the lp relaxation , density value , @xmath240 as well as number and sizes of the connected components .
furthermore , we give the names and the citeseer ranks of the team members who have rank at most 1000 .
note that could only be applied to some of the tasks and failed to find a feasible team in several cases . as a first task
we show the unconstrained solution where we maximize density without any constraints .
note that this problem is optimally solvable in polynomial time and all methods find the optimal solution .
the second task asks for at least three experts with skill . here
again all methods return the same team , which is indeed optimal since the lp bound agrees with the density of the obtained team .
next we illustrate the usefulness of the additional modeling freedom of our formulation by giving an example task where obtaining meaningful , connected teams is not possible with the lower bound constraints alone .
consider a task where we need at least four experts having skill ( task 3 ) .
for this , all methods return the same disconnected team of size seven where only four members have the skill .
the other three experts possess skills and and are densely connected among themselves .
one can see from the lp bound that this team is again optimal .
this example illustrates the major drawback of the density based objective which while preferring higher density subgraphs compromises on the connectivity of the solution .
our further experiments revealed that the subgraph corresponding to the skill is less densely connected ( relative to the other skills ) and forming coherent teams in this case is difficult without specifying additional requirements . with the help of subset and distance based constraints supported by
, we can now impose the team requirements more precisely and obtain meaningful teams . in task 4
, we require that andrew y. ng is the team leader and that all experts of the team should be within a distance of two from each other in terms of the underlying co - author graph .
the result of our method is a densely connected and highly ranked team of size four with a density of 3.89 .
note that this is very close to the lp bound of 3.91 .
the feasible solution obtained by is worse than our result both in terms of density and @xmath240 .
the greedy method can not be applied to this task because of the distance constraint . in task 5
we choose bernhard schoelkopf as the team leader while keeping the constraints from the previous task .
out of the three methods , only can solve this problem .
it produces a large disconnected team , many members of which are highly skilled experts from the skill and have strong connections among themselves . to filter these densely connected members of high expertise , we introduce a budget constraint in task 6 , where we define the cost of the team as the total number of publications of its members .
again this task can be solved only by which produces a compact team of four well - known experts .
a slightly better solution is obtained when is initialized with the infeasible solution of the lp relaxation as shown ( only in this task ) .
this is an indication that on more difficult instances of , it pays off to run with more than one starting point to get the best results .
the solution of the lp , possibly infeasible , is a good starting point apart from the unconstrained solution of .
tasks 7 , 8 and 9 provide some additional teams found by for other tasks involving upper and lower bound constraints on different skills . as noted in section [ sec : lprel ] the lp bound is loose in the presence of upper bound constraints and this is also the reason why it was not possible to derive a feasible solution from the lp relaxation in these cases . in fact the lp bounds for these tasks remain the same even if the upper bound constraints are dropped from these tasks .
by incorporating various realistic constraints we have made a step forward towards a realistic formulation of the team formation problem .
our method finds qualitatively better teams that are more compact and have higher densities than those found by the greedy method @xcite .
our linear programming relaxation not only allows us to check the solution quality but also provides a good starting point for our non - convex method .
however , arguably , a potential downside of a density - based approach is that it does not guarantee connected components .
a further extension of our approach could aim at incorporating `` connectedness '' or a relaxed version of it as an additional constraint .
we gratefully acknowledge support from the excellence cluster mmci at saarland university funded by the german research foundation ( dfg ) and the project nolepro funded by the european research council ( erc ) .
the subgradient of @xmath241 is given by @xmath242 , where @xmath243 is the indicator function of the largest entry of @xmath98 .
for the subgradient of @xmath244 , using prop . 2.2 . in @xcite
, we obtain for the subgradient @xmath245 of the terms of the form @xmath246 , @xmath247 defining @xmath248 , an element of the subgradient of the second term of @xmath244 is given as @xmath249 , where @xmath166 and @xmath250 where @xmath251 , if @xmath252 ; -1 if @xmath253 ; @xmath254 $ ] , if @xmath255 . in total
, we obtain for the subgradient @xmath256 of @xmath177 , @xmath257 | given a task @xmath0 , a set of experts @xmath1 with multiple skills and a social network @xmath2 reflecting the compatibility among the experts , _ team formation _ is the problem of identifying a team @xmath3 that is both competent in performing the task @xmath0 and compatible in working together .
existing methods for this problem make too restrictive assumptions and thus can not model practical scenarios .
the goal of this paper is to consider the team formation problem in a realistic setting and present a novel formulation based on densest subgraphs .
our formulation allows modeling of many natural requirements such as ( i ) inclusion of a designated team leader and/or a group of given experts , ( ii ) restriction of the size or more generally cost of the team ( iii ) enforcing _ locality _ of the team , e.g. , in a geographical sense or social sense , etc .
the proposed formulation leads to a generalized version of the classical densest subgraph problem with cardinality constraints ( dsp ) , which is an np hard problem and has many applications in social network analysis . in this paper
, we present a new method for ( approximately ) solving the generalized dsp ( gdsp ) .
our method , , is based on solving an _ equivalent _ continuous relaxation of gdsp .
the solution found by our method has a quality guarantee and always satisfies the constraints of gdsp .
experiments show that the proposed formulation ( gdsp ) is useful in modeling a broader range of team formation problems and that our method produces more coherent and compact teams of high quality .
we also show , with the help of an lp relaxation of gdsp , that our method gives close to optimal solutions to gdsp .
[ data mining ] [ graph algorithms ] |
in this paper , we prove stability results for the loomis - whitney inequality and some of its generalisations .
let us start by describing these results .
[ [ the - loomis - whitney - inequality . ] ] the loomis - whitney inequality .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the loomis - whitney inequality @xcite bounds the volume of a @xmath0-dimensional body in terms of the volumes of its @xmath2-dimensional projections .
it states that every body @xmath3 in @xmath4 satisfies @xmath5 } \mu(\pi_{[d ] \setminus \{i\ } } ( s)),\ ] ] where @xmath6 denotes lebesgue measure , @xmath7:=\{1,2,\ldots , d\}$ ] , and for @xmath8 $ ] , we denote by @xmath9 the projection of @xmath3 onto the coordinates of @xmath10 ( i.e. , the projection onto the subspace @xmath11 ) . here
, we say that @xmath12 is a _ body _ if it is an open set with compact closure .
note that if @xmath3 is a cartesian product of subsets of @xmath13 , then equality holds in ; we call such a body a _ box_. [ [ the - box - theorem - and - the - uniform - cover - inequality . ] ] the box theorem and the uniform cover inequality .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + the box theorem of bollobs and thomason @xcite is a simultaneous generalisation and strengthening of the loomis - whitney inequality .
it states that for any body @xmath14 , there exists a box @xmath15 with the same measure as @xmath3 , such that @xmath16 for all @xmath8 $ ] .
bollobs and thomason show that this is equivalent to the so - called ` uniform cover inequality ' of chung , frankl , graham and shearer @xcite .
we say that a family @xmath17)$ ] is a _
uniform @xmath18-cover _ if every @xmath19 $ ] belongs to exactly @xmath18 of the sets in @xmath20 , and that @xmath21 is a _ uniform cover _ if @xmath21 is a uniform @xmath18-cover for some @xmath22 .
the uniform cover inequality states that for any body @xmath12 , any @xmath22 , and any uniform @xmath18-cover @xmath17)$ ] , we have @xmath23 the uniform cover inequality is sharp when @xmath3 is a box .
applying the uniform cover inequality to sets which are unions of axis - parallel unit cubes , implies that for any finite @xmath24 , any @xmath22 , and any uniform @xmath18-cover @xmath25)$ ] , we have @xmath26 denotes the cardinality of the set @xmath3 . ) since increasing the size of a set @xmath27 can only increase the right - hand side of ( [ eq : uc - finite ] ) , it follows that for any set family @xmath17)$ ] , and any finite @xmath24 , we have @xmath28 denotes the minimum integer @xmath18 such that every @xmath19 $ ] belongs to at least @xmath18 of the sets in @xmath21 .
in fact , implies , by a standard approximation argument , approximating a body @xmath12 by a union of cubes in a sufficiently fine grid , as outlined in @xcite .
note , however , that the analogue of for bodies in @xmath4 ( with lebesgue measure ) does not necessarily hold if @xmath21 is not a uniform cover , as can be seen by taking @xmath3 to be a @xmath0-dimensional axis - parallel cube of side - length less than 1 . ( roughly speaking , the approximation argument requires both sides of ( [ eq : uc ] ) to scale by the same factor , when @xmath3 is dilated ; see the proof of corollary [ corr : uniform - cover - stability ] in section [ lemmas ] below . ) [ [ shearers - lemma . ] ] shearer s lemma .
+ + + + + + + + + + + + + + + + loomis and whitney , and bollobs and thomason , proved their results using induction on the dimension , and hlder s inequality .
however , the discrete versions of the loomis - whitney and uniform cover inequalities ( which are equivalent to the continuous ones ) are special cases of shearer s entropy lemma .
the term ` shearer s lemma ' is often used to refer to two essentially equivalent results .
the first is stated in terms of the entropy of a random variable , first proved by shearer ( implicitly ) in 1978 , and first published by chung , frankl , graham and shearer in 1986 in @xcite .
the second , from the same paper , is equation , the result we referred to above as the uniform cover inequality .
the entropy approach for proving these inequalities is the one we adopt in this paper . as well as being very natural statements in their own right ,
the above results have many applications : for example in convex geometry ( see @xcite ) , in the study of isoperimetric problems ( see @xcite ) , in extremal combinatorics ( see @xcite ) and in additive combinatorics ( see @xcite ) .
there are also several useful generalisations of shearer s lemma , such as the weighted version in @xcite , which is itself a special case of the brascamp - lieb inequality @xcite . in @xcite , ball stated and applied the geometric version of the brascamp - lieb inequality to study sections of the euclidean cube .
the results in this paper apply directly to several of these generalisations . when a geometric , combinatorial or functional inequality is sharp , it is natural to ask whether it is also ` stable ' i.e. , when the inequality is _ almost _ sharp for a particular object , must that object be close in structure to an _ extremal _ one ( meaning , an object for which equality holds ) ?
stability phenomena for geometric and functional inequalities have been widely studied in recent years . to obtain a stability result for an inequality ,
it is natural to look closely at known proofs of the inequality and see what information these proofs yield about objects where equality ` almost ' holds .
several methods for proving geometric inequalities have recently been shown to yield best - possible ( or close to best - possible ) stability results .
a partial list includes symmetrization techniques ( see e.g. @xcite ) , optimal transport ( see e.g. @xcite ) , spectral techniques ( see e.g. @xcite ) , and non - linear evolution equations ( see e.g. @xcite ) .
stability phenomena for combinatorial inequalities have also been widely studied , and best - possible ( or close to best - possible ) stability results have been obtained via elementary combinatorial arguments ( see e.g. @xcite ) , and using spectral techniques ( see e.g. @xcite ) , fourier analysis ( see e.g. @xcite ) and ` non - abelian ' fourier analysis ( see e.g. @xcite ) .
our main result in this paper is a stability result for the uniform cover inequality . to state it ,
we need some more notation .
we define a _ box _ in @xmath1 to be a cartesian product of finite subsets of @xmath29 . for a collection of subsets of coordinates @xmath17)$ ] , we denote by @xmath30 the maximum integer @xmath31 such that for every @xmath32 $ ] with @xmath33 , there are at least @xmath31 sets in @xmath20 containing @xmath34 but not @xmath35 .
if @xmath36 , then we define @xmath37 we prove the following stability result for the inequality ( [ eq : uc - finite - extended ] ) .
[ thm : uniform - cover - stability ] for every integer @xmath38 there exists @xmath39 such that the following holds .
let @xmath17)$ ] with @xmath40 .
let @xmath24 with @xmath41 . if @xmath42 then there exists a box @xmath43 such that @xmath44 our proof yields @xmath45 .
this theorem is best - possible in terms of its dependence upon @xmath46 , as can be seen by taking @xmath47^d \setminus [ a']^d$ ] , where @xmath48 , and taking @xmath49^{(d-1)}$ ] .
theorem [ thm : uniform - cover - stability ] easily implies the following analogue for bodies ( with uniform covers ) , via the standard approximation argument outlined in @xcite , and referred to above .
[ corr : uniform - cover - stability ] for every integer @xmath38 there exists @xmath50 such that the following holds .
let @xmath22 , and let @xmath25)$ ] be a uniform @xmath18-cover with @xmath51 .
let @xmath12 be a body such that @xmath52 then there exists a box @xmath15 such that @xmath53 for completeness , we present in full the deduction of corollary [ corr : uniform - cover - stability ] from theorem [ thm : uniform - cover - stability ] , in section [ lemmas ] . if @xmath54^{(d-1)}$ ] , then @xmath55 , @xmath56 and @xmath57 , so the following stability result for the loomis - whitney inequality , is a special case of theorem [ thm : uniform - cover - stability ] .
[ cor : lw - stability ] for every integer @xmath38 there exists @xmath58 such that the following holds .
let @xmath24 with @xmath41 . if @xmath59\setminus\{i\ } } ( s)| \right)^{1/(d-1)},\ ] ] then there exists a box @xmath43 such that @xmath60 of course , we can take @xmath61 in this corollary .
[ [ stability - for - a - weighted - version - of - the - uniform - cover - inequality . ] ] stability for a weighted version of the uniform cover inequality .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + in @xcite , a weighted version of the uniform cover inequality is proved , a version that is , in fact , a special case of the brascamp - lieb inequality of @xcite .
it is not hard to verify that our proof in this paper goes through almost word for word to yield the following stability result for the weighted version .
[ thm : weighted ] for every integer @xmath38 there exists @xmath50 such that the following holds .
let @xmath25)$ ] , and let @xmath62 be a non - negative weight function on @xmath20 , such that every @xmath19 $ ] is covered by sets with total weight at least 1 , i.e. @xmath63 let @xmath64 assume @xmath36 , and let @xmath65 .
let @xmath24 with @xmath41 . if @xmath66 then there exists a box @xmath43 such that @xmath44 again , we can take @xmath45 .
we omit the details of the proof of theorem [ thm : weighted ] . [
[ about - the - proof . ] ] about the proof .
+ + + + + + + + + + + + + + + + a few words regarding our proof of theorem [ thm : uniform - cover - stability ] . to prove a stability result for some inequality , it is natural to consider a certain proof of that inequality , and to ` work backwards ' through this proof to deduce closeness to the desired structure when equality almost holds . this may be called a ` stable version ' of the relevant proof . perhaps the first natural approach to proving theorem
[ thm : uniform - cover - stability ] is to produce a stable version of the classical proof of , which uses hlder s inequality and induction on the dimension .
however , even for the loomis - whitney inequality , this only yields @xmath67-dependence . in order to obtain the sharp @xmath68-dependence we seek
, we consider the beautiful llewellyn - radhakrishnan proof of shearer s lemma ( see @xcite ) .
this proof is information - theoretic , using simple properties of entropy .
given a set @xmath69 and a family @xmath25)$ ] , such that the inequality ( [ eq : uc - finite - extended ] ) is almost tight for @xmath3 and @xmath20 , we first deduce that , in a sense , @xmath3 is ` close ' to a ` product structure ' ; that is , the uniform distribution on @xmath3 is ` close ' to the product of its marginals ( the relevant notion of distance is discussed below ) .
that part of the argument is a fairly straightforward deduction from the entropy proof of llewellyn - radhakrishnan , or alternatively from the entropy inequalities proved by balister and bollobs in @xcite .
( it is also inspired by the proof of the parallel repetition theorem in @xcite , and related works such as @xcite . )
although information theory allows us to easily identify a product structure , it moves us from the language of sets to the language of distributions .
the next ingredient of the proof is more combinatorial .
its purpose is to move back from the language of distributions to the language of sets ; we identify the actual box that we claim exists .
it turns out that for this part of the proof , a ` two - dimensional lemma ' suffices .
given a set @xmath70 , such that the uniform distribution on @xmath3 is ` close ' to the product of its marginals , we find a two - dimensional box ( a ` rectangle ' ) @xmath71 , which is a good approximation of the set @xmath3 . to prove this lemma , we need to identify the two sets @xmath72 and @xmath73 .
this is done via an iterative ` trimming ' procedure , which gradually removes parts of @xmath74 and @xmath75 , until only @xmath76 and @xmath77 remain .
the crux of the proof is in showing that we did not throw too much , i.e. , that @xmath78 is small .
we apply this two - dimensional lemma @xmath0 times ; for each @xmath19 $ ] we consider our set @xmath69 as a two - dimensional set @xmath79 \setminus \{i\}}$ ] , and we find a set @xmath80 which is a good candidate to be the ` edge in direction @xmath34 ' of the box approximating @xmath3 .
we then check that , indeed , @xmath3 is close to the box @xmath81 .
the last issue to discuss is the meaning of the word ` close ' above : how to measure the ` distance ' between the uniform distribution on @xmath3 and the product of its marginals .
the information theoretic part of the argument naturally leads to measuring this distance using the kullback - leibler divergence . a natural thing to do at this point
would be to use pinsker s inequality to move from kullback - leibler divergence to @xmath82-distance .
this , however , leads to suboptimal @xmath83-dependence . to overcome this difficulty
, we introduce a new ( but natural ) measure of distance , which we call the ` hole - weight ' .
the hole - weight suffices to control the trimming procedure , yielding optimal @xmath84-dependence .
it may find applications in other , similar scenarios .
for more details on this part of the proof , see section [ lemmas ] below . in section [ sec : edgeiso ] , we apply theorem [ thm : uniform - cover - stability ] to prove another result , demonstrating stability for the edge - isoperimetric inequality in the infinite @xmath0-dimensional lattice . before stating this formally
, we give some background on isoperimetric inequalities .
isoperimetric problems are classical objects of study in mathematics . in general
, they ask for the minimum possible ` boundary size ' of a set of a given ` size ' , where the exact meaning of these words varies according to the problem .
the classical isoperimetric problem in the plane asks for the minimum possible perimeter of a shape in the plane with area 1 .
the answer , that it is best to take a circle , was already known to the ancient greeks , but it was not until the 19th century that this was proved rigorously . the isoperimetric inequality for euclidean space states that among all subsets of @xmath4 of given volume , euclidean balls have the smallest boundary . to state a version of this precisely , if @xmath85 is a borel set of finite lebesgue measure , we denote by @xmath86 the _ distributional perimeter _ of @xmath87 ( see e.g. chapter 12 in @xcite for the definition of distributional perimeter )
when @xmath87 has piecewise smooth topological boundary @xmath88 , then @xmath89 , where @xmath90 denotes the @xmath2-dimensional hausdorff measure of @xmath88 , a measure of the boundary which may be more familiar to some readers .
[ thm : iso - rd ] if @xmath12 is a borel set with lebesgue measure @xmath91 , then @xmath92 where @xmath93 is an euclidean ball in @xmath4 with @xmath94 . in this paper
, we consider a discrete analogue of theorem [ thm : iso - rd ] . to state it , we need some more notation .
let @xmath95 denote the graph of the @xmath0-dimensional integer lattice , i.e. the graph with vertex - set @xmath1 and edge - set @xmath96\},\ ] ] where @xmath97 denotes the @xmath34th unit vector in @xmath4 .
if @xmath24 , we let @xmath98 denote the _ edge - boundary _ of @xmath3 in the graph @xmath95 , meaning the set of edges of @xmath95 which join a vertex in @xmath3 to a vertex not in @xmath3 .
the following edge - isoperimetric inequality in @xmath95 is an easy consequence of the loomis - whitney inequality , and the inequality of arithmetic and geometric means ( the am - gm inequality , for short ) .
it is also an immediate consequence of theorem 8 in @xcite .
[ thm : edge - iso ] let @xmath24 with @xmath41 .
then @xmath99 equality holds in theorem [ thm : edge - iso ] if @xmath47^d$ ] for some @xmath100 .
very slightly more generally , equality holds if @xmath101 , where @xmath102 are equal - sized intervals in @xmath29 ; we will call such a set a _
cube_. in their seminal work @xcite , fusco , maggi and pratelli proved that if @xmath12 is a borel set of finite measure and with distributional perimeter close to the minimum possible size ( viz , the size given by theorem [ thm : iso - rd ] ) , then @xmath3 must be close in symmetric difference to an euclidean ball of the same measure , confirming a conjecture of hall .
[ thm : fmp ] suppose @xmath12 is a borel set with lebesgue measure @xmath91 , and with @xmath103 where @xmath93 is a euclidean ball with @xmath94
. then there exists @xmath104 such that @xmath105 where @xmath106 is a constant depending upon @xmath0 alone . as observed in @xcite , theorem [ thm : fmp ]
is sharp up to the value of the constant @xmath107 , as can be seen by taking @xmath3 to be an ellipsoid with @xmath108 semi - axes of length 1 and one semi - axis of length slightly larger than 1 . in this paper
, we prove a discrete analogue of theorem [ thm : fmp ] by using theorem [ thm : uniform - cover - stability ] .
we prove the following stability result for the edge - isoperimetric inequality in @xmath95 .
[ thm : edge - iso - stability ] let @xmath109 with @xmath110 .
if @xmath24 with @xmath41 and with @xmath111 then there exists a cube @xmath112 such that @xmath113 theorem [ thm : edge - iso - stability ] has the best possible dependence on @xmath46 , as can be seen by taking a ` cuboid ' @xmath47^{d-1 } \times [ b]$ ] , where @xmath114 is slightly larger than @xmath115 ( see remark [ remark : iso - sharpness ] for details ) .
we conjecture that the dependence on @xmath0 could be improved , to @xmath116 ( see section [ sec : conclusion ] ) .
in section [ sec : prelim ] , we cover some background and introduce some notation . in subsection [ lemmas ] , we present our main lemmas and prove that they imply our main theorem . in subsection [ proofsoflemmas ] , we prove the main lemmas . in section [ sec : edgeiso ] , we prove theorem [ thm : edge - iso - stability ] , our stability result for the edge - isoperimetric inequality in the lattice @xmath95 .
finally , in section [ sec : conclusion ] , we conclude with some open problems .
in this section , we state some definitions and known results from probability theory and information theory , and we describe some of our notation . for background concerning the information - theoretic results , and for proofs ,
the reader is referred to @xcite . throughout this paper ,
@xmath117 means @xmath118 , and we use the convention @xmath119 . let @xmath120 be a probability distribution on a finite or countable set @xmath121 .
the _ entropy _ of @xmath120 is defined by @xmath122 the entropy of a random variable is the entropy of its distribution .
( by a slight abuse of terminology , we will often identify a random variable with its distribution . ) intuitively , the entropy of a random variable measures the ` amount of information ' one has from knowing the value of the random variable .
let @xmath123 denote the support of the distribution @xmath120 , i.e. @xmath124 . the convexity of @xmath125 implies that @xmath126 note that equality holds in ( [ eq : uniform - bound ] ) iff @xmath120 is uniformly distributed on its support . for two random variables @xmath127 taking values in a set @xmath121 , the _
conditional entropy of @xmath87 given @xmath93 _ is defined by @xmath128 the _ chain rule _ for entropy follows immediately : @xmath129 it is easy to prove that _ conditioning does not increase entropy _ : for any two random variables @xmath127 , @xmath130 for three random variables @xmath131 , we denote by @xmath132 the probability of the event @xmath133 , we denote by @xmath134 the probability of the event @xmath135 , and if @xmath136 , we denote by @xmath137 the probability of @xmath135 given @xmath138 . if @xmath120 and @xmath139 are two distributions on a finite or countable set @xmath121 , with @xmath140 , the _ kullback - leibler divergence _ between @xmath141 is defined by @xmath142 if @xmath143 , we define @xmath144 .
the kullback - leibler divergence is non - negative , it is zero if and only if @xmath145 , but it is not symmetric , i.e. in general , @xmath146 , even when @xmath147
. the _ mutual information of @xmath87 and @xmath93 _ is defined by @xmath148 note that this is also the kullback - leibler divergence between the joint distribution of @xmath87 and @xmath93 and the product of the marginals .
the _ mutual information of @xmath87 and @xmath93 , conditioned on @xmath149 _ , is defined by @xmath150 mutual information _ is _ symmetric under interchanging @xmath87 and @xmath93 , i.e. @xmath151 another ( intuitively plausible ) property of mutual information is that if @xmath149 is a function of @xmath93 , then @xmath152 we refer to this property as the ` monotonicity of mutual information ' .
[ [ marginal - distributions . ] ] marginal distributions .
+ + + + + + + + + + + + + + + + + + + + + + + let @xmath120 be a probability distribution on @xmath153 .
for a subset @xmath8 $ ] , we denote by @xmath154 the marginal distribution of @xmath120 on the set of coordinates @xmath10 , i.e. @xmath155 \setminus g}).\ ] ] if @xmath156 is a partition of @xmath7 $ ] , we denote by @xmath157 the obvious product - distribution on @xmath153 .
we will need the following equation relating the divergences between various products of marginals of @xmath120 .
@xmath158 } \mid \mid p_{[i-1]}p_i\right).\ ] ] this is easily verified , using the definition of @xmath159 and expanding the logarithms on the right - hand side .
note that if @xmath160 is a random variable with probability distribution @xmath120 , then the left - hand side is precisely the _ total correlation _ of the set of random variables @xmath161 .
in this subsection , we present several statements that , when put together , easily imply our main theorem .
first , we would like to show that if the uniform cover inequality is close to being tight for a set @xmath3 , then there is not much mutual information between any 1-dimensional marginal of the uniform distribution on @xmath3 , and the complementary @xmath2-dimensional marginal .
[ lem : tighttoinfo ] let @xmath109 with @xmath110 .
let @xmath25)$ ] with @xmath162 .
let @xmath163 .
let @xmath69 with @xmath41 and with @xmath164 let @xmath120 denote the uniform distribution on @xmath3 .
then for all @xmath19 $ ] , we have @xmath165 \setminus \{i\ } } ) \leq 2 \rho({{\cal g } } ) { \epsilon}.\ ] ] note that , by monotonicity of mutual information , lemma [ lem : tighttoinfo ] implies that for any @xmath166 $ ] and any @xmath167 , we have @xmath168 \setminus \{i\ } } ) \leq 2 \rho({{\cal g } } ) { \epsilon}.\ ] ] given a set @xmath69 , we want to measure how far the uniform distribution @xmath120 on @xmath3 is from the product of some of its marginals .
it turns out that a useful measure for us ( which we call the ` hole - weight ' ) is the sum of the product of the marginals over all points _ not _ in @xmath3 ( ` holes ' ) .
formally , if @xmath169 is a partition of @xmath7 $ ] , we define the _ hole - weight _ of @xmath3 with respect to @xmath170 by @xmath171 in all but one case below , the partition of @xmath7 $ ] which defines the hole - weight will be of the form @xmath172\setminus\{i\})$ ] , so for brevity , we write @xmath173\setminus\{i\}}(s)$ ] .
also , when stating and proving lemmas , instead of considering @xmath174 \setminus \{i\}}$ ] , we will sometimes consider the general ` two - dimensional ' setting @xmath70 , i.e. @xmath3 is simply a subset of a product of two sets . if @xmath175 and @xmath176 are sets , and @xmath70 , we will write @xmath177 . the following claim bounds from above the hole - weight of @xmath3 by the kullback - leibler divergence between the uniform distribution on @xmath3 , and the product of its marginals .
[ clm : dhvsdd ] let @xmath69 with @xmath41 , let @xmath120 be the uniform distribution on @xmath3 , and let @xmath178 be a partition of @xmath7 $ ] .
let @xmath120 be the uniform distribution on @xmath3 , and let @xmath179 denote the corresponding marginals .
then @xmath180 of course , another very natural way of measuring how far @xmath120 is from the product of its marginals is to simply use the @xmath82-distance @xmath181 ( recall that if @xmath120 and @xmath139 are probability distributions on a finite or countable set @xmath121 , then the _ @xmath82-distance _ between @xmath120 and @xmath139 is defined by @xmath182 the quantity @xmath183 is often called the _ total variation distance _ between @xmath120 and @xmath139 . )
one can bound the @xmath82-distance ( [ eq : l1 ] ) in terms of the divergence on the right - hand side of ( [ eq : hole - div ] ) , using pinsker s inequality .
pinsker s inequality states that if @xmath120 and @xmath139 are two probability distributions on a finite or countable set @xmath121 , then @xmath184 ( note that pinkser originally proved with a worse constant .
the above form , in which the constant is sharp , is due independently to kullback , csiszr and kemperman . )
applying this yields @xmath185 unfortunately , this application of pinsker s inequality introduces @xmath67-dependence in the conclusion of theorem [ thm : uniform - cover - stability ] ( this was our original approach ) .
we obtain @xmath68-dependence by relying only on the hole - weight .
we now need a lemma saying that if the hole - weight of a two - dimensional set @xmath3 is small , then @xmath3 is close to a 2-dimensional box .
[ lem : findbox1 ] let @xmath175 and @xmath176 be sets .
let @xmath70 with @xmath41 .
let @xmath120 denote the uniform distribution on @xmath3 , and let @xmath186 denote its marginals .
let @xmath187 .
then there exists @xmath188 such that @xmath189 and such that for every @xmath190 , we have @xmath191 the idea behind lemma [ lem : findbox1 ] is that the set @xmath188 is a good candidate to be one of the multiplicands ( ` edges ' ) of a box approximating @xmath3 : on the one hand , it captures most of @xmath192 , and on the other hand , @xmath193 restricted to @xmath76 is close ( in a sense ) to being uniform .
indeed , the main step in the proof of theorem [ thm : uniform - cover - stability ] below is to apply lemma [ lem : findbox1 ] , with an appropriate value of @xmath194 , yielding ( for each @xmath19)$ ] a set @xmath80 , and then to show that @xmath3 is close in symmetric difference to the cartesian product of the @xmath195 s .
the above lemmas now yield the proof of our main theorem .
let @xmath25)$ ] with @xmath196 .
we may and shall assume that @xmath197 , as otherwise the conclusion of the theorem holds trivially .
given a set @xmath24 with @xmath41 , and with @xmath42 we apply lemma [ lem : tighttoinfo ] to deduce that for all @xmath19 $ ] , @xmath165\setminus \{i\ } } ) \leq 2 \rho({{\cal g } } ) { \epsilon}.\ ] ] this implies , via ( [ eq : smalld ] ) and claim [ clm : dhvsdd ] , that for every @xmath34 in @xmath7 $ ] , @xmath198 next , for each @xmath19 $ ] , we apply lemma [ lem : findbox1 ] to @xmath174\setminus \{i\}}$ ] ( i.e. , we take @xmath199 and @xmath200\setminus \{i\}}$ ] ) , so that @xmath201 ; we take @xmath202 .
this yields ( for every @xmath19 $ ] ) a set @xmath80 such that @xmath203 and such that for any @xmath204 , @xmath205 let @xmath206 by the union bound , @xmath207 i.e. @xmath208 on the other hand , for every @xmath209 , we have @xmath210 hence , @xmath211 applying claim [ clm : dhvsdd ] and equation ( [ telescope ] ) , it follows that @xmath212}\mid \mid p_{[i-1]}p_i).\ ] ] applying the bound ( [ eq : smalld ] ) gives @xmath213 which implies that @xmath214 .
hence , @xmath215 , and so @xmath216 combining ( [ eq : s - r ] ) and ( [ eq : r - s ] ) gives @xmath217 so we may take @xmath218 , completing the proof . for completeness , we now present the deduction of corollary [ corr : uniform - cover - stability ] from theorem [ thm : uniform - cover - stability ] , using the approximation argument outlined in @xcite .
let @xmath22 .
let @xmath25)$ ] be a uniform @xmath18-cover .
let @xmath12 be a body such that @xmath219 fix @xmath220 .
since @xmath221 is compact and @xmath222 , there exists a compact set @xmath223 such that @xmath224 .
choose an open cover @xmath225 of @xmath226 by open cubes with corners at rational coordinates , such that all the cubes in @xmath225 are contained within @xmath3 . since @xmath226 is compact
, we may choose a finite subset @xmath227 such that @xmath228 is a cover of @xmath226 .
choose an axis - parallel grid of some side - length @xmath229 , which is a common refinement of all the sets in @xmath228 ( meaning that all the open cubes in @xmath228 are unions of open grid - cubes ) .
let @xmath230 be the union of all the open grid - cubes which are contained in @xmath3 .
then @xmath231 . let @xmath232 be the number of open grid - cubes in @xmath230 , and for each @xmath8 $ ] , let @xmath233 be the number of ( lower - dimensional ) open grid - cubes in the projection of @xmath230 onto the plane @xmath234 .
then we have @xmath235 hence , cancelling the common factor of @xmath236 and rearranging , we obtain @xmath237 therefore , by theorem [ thm : uniform - cover - stability ] , there exists a box @xmath238 ( which is a union of open grid - cubes ) , such that @xmath239 hence , @xmath240 since @xmath220 was arbitrary , it follows by a compactness argument that there exists a box @xmath15 such that @xmath241 we now present the proofs of our main lemmas .
our proof is information - theoretic , inspired by the technique of radhakrishnan from @xcite .
let @xmath242 and @xmath120 be as in the statement of the lemma .
let @xmath121 be a random variable uniformly distributed on @xmath3 .
assume without loss of generality that @xmath243 .
recall that @xmath244 and that by ( [ eq : uniform - bound ] ) , for all @xmath8 $ ] , we have @xmath245 .
set @xmath246 , @xmath247 and @xmath248 .
since @xmath249 , we have @xmath250 hence , @xmath251}\mid x_d)\tag{by the chain rule } \\ & = \frac{\sum_{g \in { { \cal g}}:\atop d \in g } ( h(x_d)+h(x_{g \setminus \{d\}}\mid x_d))+ \sum_{g \in { { \cal g}}:\atop d \not \in g}h(x_g)}{m } - h(x_d ) -h(x_{[d-1]}\mid x_d ) \\ & \geq \frac{\sum_{g \in { { \cal g}}:\atop d \in g } h(x_{g \setminus \{d\}}\mid x_d)+ \sum_{g \in { { \cal g}}:\atop d \not \in g}h(x_g)}{m } -h(x_{[d-1]}\mid x_d ) \tag{since , by definition of $ m = m({{\cal g}})$ , there are at least $ m$ sets $ g \in { { \cal g}}$ with $ d \in g$ } \\ & = \sum _ { j < d}\left ( \frac{\sum_{g \in { { \cal g}}:\atop d , j \in g } h(x_j\mid x_{g \cap [ j-1]},x_d)+ \sum_{g \in { { \cal g}}:\atop g \cap \{d , j\}=j}h(x_j\mid x_{g \cap [ j-1]})}{m } -h(x_{j}\mid x_{[j-1]},x_d ) \right ) \tag{by the chain rule } \\ & \geq \sum _ { j < d}\left ( \frac{\sum_{g \in { { \cal g}}:\atop d , j \in g } h(x_j\mid x_{[j-1]},x_d)+ \sum_{g \in { { \cal g}}:\atop g \cap \{d , j\}=j}h(x_j\mid x_{[j-1]})}{m } -h(x_{j}\mid x_{[j-1]},x_d ) \right ) \tag{since extra conditioning does not increase entropy}\\ & \geq \frac{\sigma}{m}\sum _ { j < d } \left ( h(x_j\mid x_{[j-1 ] } ) - h(x_j\mid x_{[j-1]},x_d ) \right ) \tag{using the definitions of $ m$ and $ \sigma$}\\ & = \rho^{-1 } \left(h(x_{[d-1 ] } ) - h(x_{[d-1 ] } \mid x_d)\right ) \tag{by the chain rule}\\ & = \rho^{-1 } i(x_{[d-1]};x_d),\end{aligned}\ ] ] as required .
as mentioned in the introduction , lemma [ lem : tighttoinfo ] can also be proved by appealing to an entropy inequality in the paper @xcite of balister and bollobs .
indeed , without loss of generality taking @xmath243 , it suffices to prove the inequality @xmath252 } ) + ( m-\sigma ) h(x ) \leq \sum_{g \in { { \cal g } } } h(x_g),\ ] ] which follows from two applications of theorem 6 in @xcite , the first application being with @xmath253 .
we have opted to give the self - contained proof above , as it is not much longer .
recall that @xmath254 for all @xmath255 , since @xmath256 . using this , and the convexity of @xmath257
, we have @xmath258 in this subsection , we deal with the most combinatorially - flavoured ingredient in our paper , the proof of lemma [ lem : findbox1 ] . given a set @xmath70 , we proceed to trim away some of its ` vertical fibres ' ( sets of the form @xmath259 , for some @xmath260 ) , in such a way that none of the remaining fibres is very short compared to the others .
this leaves us with a subset @xmath188 , such that weight of each remaining vertical fibre of @xmath76 is at least a constant fraction of the average weight of the remaining fibres .
the total amount of mass trimmed is bounded from above in terms of @xmath261 . applying this procedure with @xmath262 , and then repeating it to find a similar set @xmath263 , yields the following .
[ lem : risz2 ] let @xmath175 and @xmath176 be sets . if @xmath70 with @xmath41 , then there exist @xmath188 and @xmath263 such that @xmath264 we will not use this lemma directly in our proof of our main theorem , but its proof is included for the reader s interest at the end of this subsection . for @xmath265 ,
let @xmath266 denote the natural projection of @xmath3 onto @xmath267 .
we find an appropriate set @xmath76 by iteratively removing from @xmath268 those points @xmath269 for which @xmath270 is too ` small ' , namely , those points whose fibre has size less than @xmath271 times the average size of a fibre .
to this end , we define recursively : @xmath272 and @xmath273 , and for each @xmath274 , @xmath275 define the limit objects : @xmath276 since @xmath3 is finite , this process stabilizes after finitely many steps .
if we can show that @xmath277 then by definition , for every @xmath190 , we indeed have @xmath278 proving lemma [ lem : findbox1 ] . to obtain ( [ eq : epsilon1bound ] ) , it suffices to prove the following claim .
[ claim : smallb1 ] @xmath279 indeed , claim [ claim : smallb1 ] implies @xmath280 implying ( [ eq : epsilon1bound ] ) .
it remains , therefore , to prove claim [ claim : smallb1 ] .
_ proof of claim [ claim : smallb1 ] . _ fix @xmath274 and let @xmath281 .
let @xmath282 be the fibre of @xmath269 , and let @xmath283 .
we first bound the mass of the set of elements of @xmath3 whose projection lies in @xmath284 . by definition
, we have @xmath285 and @xmath286 so @xmath287 we therefore have @xmath288 it follows that @xmath289 now summing over @xmath269 and using , we have @xmath290 finally , summing over all @xmath35 , and recalling that if @xmath291 then @xmath292 , we have @xmath293 this completes the proof of claim [ claim : smallb1 ] , and thus of lemma [ lem : findbox1 ] .
we may assume that @xmath294 , otherwise the conclusion of the lemma holds with @xmath295 .
let @xmath188 be the set given by lemma [ lem : findbox1 ] with @xmath296 .
let @xmath263 be the set given by lemma [ lem : findbox1 ] with @xmath262 , when coordinates 1 and 2 are interchanged .
let @xmath297 . by the union bound , we have @xmath298 on the other hand , for every @xmath299 , we have @xmath300 therefore , @xmath301 which implies that @xmath302 .
it follows that @xmath303 and therefore @xmath304 , completing the proof .
in this section , we prove theorem [ thm : edge - iso - stability ] , our stability result for the edge - isoperimetric inequality in the lattice @xmath95 , using theorem [ thm : uniform - cover - stability ] and some additional combinatorial arguments . we start with a short proof of the isoperimetric inequality ( theorem [ thm : edge - iso ] ) , as it will be useful to refer to it later .
we write @xmath305 for the set of edges in @xmath98 of the form @xmath306 , i.e. the set of all direction-@xmath34 edges of @xmath98 .
since @xmath41 , for each @xmath307 , there are at least two direction-@xmath34 edges of @xmath98 which project to @xmath308 \setminus \{i\}}(x)$ ] , and therefore @xmath309 \setminus \{i\ } } ( s)|.\ ] ] summing over all @xmath34 , we obtain @xmath310 \setminus \{i\ } } ( s)|.\ ] ] the am - gm inequality and the loomis - whitney inequality yield @xmath311 . [
prop : am - gm - stability ] let @xmath312 .
let @xmath313 be such that @xmath314 let @xmath315 .
then @xmath316,\qquad ( 1 - 2d\sqrt{d\epsilon } ) g \leq z_i \leq ( 1 + 2\sqrt{d\epsilon } ) g.\ ] ] first , we assert that @xmath317 let @xmath318 be such that @xmath319 , and let @xmath320 ; then @xmath321 .
assume for a contradiction that @xmath322 .
then @xmath323 the function @xmath324 has @xmath325 for all positive @xmath326 , and is therefore strictly increasing on @xmath327 . hence , @xmath328 contradicting the definition of @xmath329 .
it follows from ( [ eq : assertion ] ) that @xmath330 so @xmath331 completing the proof .
let @xmath69 with @xmath41 and @xmath332 we may assume that @xmath333 otherwise the conclusion of the theorem holds trivially with @xmath334 . by and ( [ eq : small - boundary ] ) , we have @xmath335 \setminus \{i\ } } s|\right)^{1/(d-1 ) } \leq \left(\frac{|\partial s|}{2d}\right)^{d/(d-1 ) } \leq ( 1+\epsilon)^{d/(d-1 ) } |s| \leq \frac{1}{1 - 2\epsilon}|s|.\end{aligned}\ ] ] corollary [ cor : lw - stability ] now implies that there exists a box @xmath336 such that @xmath337 where @xmath338 our aim is to show that @xmath339 is close in symmetric difference to a cube .
clearly , by ( [ eq : close ] ) , we have @xmath340 let @xmath341 } |\pi_{[d ] \setminus \{i\ } } s| \right)^{1/d } \quad \text{and } \quad a : = |s|^{1/d}.\ ] ] note that , by ( [ eq : svsg ] ) and the loomis - whitney inequality , we have @xmath342 by and ( [ eq : small - boundary ] ) , we have @xmath343 \setminus \{i\ } } ( s)| \leq \frac{|\partial s|}{2d } \leq ( 1+\epsilon ) |s|^{(d-1)/d } \leq ( 1+\epsilon)\left ( \prod_{i \in [ d ] } |\pi_{[d ] \setminus \{i\ } } ( s)| \right)^{1/d}.\ ] ] hence , by proposition [ prop : am - gm - stability ] , all the @xmath2-dimensional projections of @xmath3 are of roughly equal size : @xmath344 , \qquad ( 1 - 2 d \delta ) g \leq |\pi_{[d ] \setminus \{i\ } } ( s)| \leq ( 1 + 2\delta ) g.\end{aligned}\ ] ] we now show that all the 1-dimensional projections of @xmath339 are of roughly equal size . [
claim : approx - size1 ] for every @xmath19 $ ] , we have @xmath345 . without loss of generality , we prove the claim for @xmath346 .
observe that @xmath347 indeed , if this does not hold , then , using ( [ eq : sizes - close ] ) , we have @xmath348 contradicting ( [ eq : close ] ) .
therefore , using ( [ eq : sizes - close ] ) , ( [ eq : projection - sizes ] ) , and ( [ eq : avsg ] ) successively , we have @xmath349 similarly , we have @xmath350,\qquad |r_i| \geq
( 1 - 7 \delta)a.\ ] ] hence , using ( [ eq : sizes - close ] ) and ( [ eq : rlarge ] ) we obtain @xmath351 next , for each @xmath19 $ ] we throw away the elements @xmath352 such that @xmath353 contains few elements of @xmath3 , producing a slightly smaller box @xmath354 .
fix @xmath19 $ ] ; without loss of generality , @xmath346 .
define @xmath355 for each @xmath356 , call @xmath357 _ heavy _ if @xmath358 , and call @xmath357 _ light _ otherwise .
we assert that there are at most @xmath359 light elements in @xmath76 . indeed ,
if the number of light elements is larger than this , then @xmath360 contradicting ( [ eq : close ] ) .
. then @xmath362 .
define @xmath363 similarly for each @xmath364 , and let @xmath365 by the union bound , we have @xmath366 and therefore @xmath367 our next step is to show that , for each @xmath19 $ ] , @xmath363 occupies most of the interval in which it is contained , as if not , @xmath368 would be too large .
[ claim : close - to - interval1 ] for each @xmath19 $ ] , we have @xmath369 . without loss of generality , we prove the claim for @xmath346 .
let @xmath370 suppose for a contradiction that @xmath371 define @xmath372 note that @xmath373 .
since @xmath41 , for each @xmath374 , there is at least one edge in @xmath375 of the form @xmath376 , where @xmath377 .
similarly , for each @xmath378 , there is at least one edge in @xmath375 of the form @xmath376 , where @xmath379 .
moreover , for each @xmath380 , there is at least one edge in @xmath375 of the form @xmath376 where @xmath381 .
it follows that @xmath382 we now assert that @xmath383 indeed , suppose on the contrary that does not hold . by ( [ eq : too - wide ] ) , the interval @xmath384 contains more than @xmath385 elements not in @xmath76 .
therefore , @xmath386 contradicting ( [ eq : close ] ) . using claim [ claim : approx - size1 ] and ( [ eq : close ] )
, we have @xmath387 but we also have @xmath388\setminus \{i\}}(s)| \geq 2(1 - 2d\delta)g \geq 2(1 - 2d\delta)a^{d-1},\ ] ] using ( [ eq : avsg ] ) and ( [ eqn : concstabamgm ] ) . hence , using ( [ eq : epsilon - assumption ] ) , we have @xmath389 contradicting ( [ eq : small - boundary ] ) .
we can now complete the proof of the theorem .
let @xmath390\}.\ ] ] assume without loss of generality that @xmath391 define @xmath392 the set @xmath149 is a cube of side - length @xmath393 containing @xmath354 .
using claims [ claim : approx - size1 ] and [ claim : close - to - interval1 ] , we have @xmath394 therefore , using ( [ eq : rlarge ] ) and ( [ eq : sizes - close ] ) , we have @xmath395 finally , using ( [ eq : rr ] ) , ( [ eq : close ] ) and ( [ eq : sizes - close ] ) , we obtain @xmath396 proving the theorem . [
remark : iso - sharpness ] as observed in the section [ sec : intro ] , theorem [ thm : edge - iso - stability ] is sharp up to a constant factor depending upon @xmath0 alone . to see this , take @xmath47^{d-1 } \times [ b]$ ] , where @xmath397 , @xmath398 and @xmath399 , i.e. @xmath3 is a ` cuboid '
then @xmath400 and @xmath401 for all cubes @xmath112 . on the other hand
, we have @xmath402 ) = \tfrac{d-1}{d } \phi^2 a^{d-1}.\end{aligned}\ ] ] hence , we have @xmath403 , where @xmath404 but @xmath3 is @xmath405-far from any cube , where @xmath406
we have proved stability results for the uniform cover inequality ( theorem [ thm : uniform - cover - stability ] ) , and the edge - isoperimetric inequality in the @xmath0-dimensional integer lattice ( theorem [ thm : edge - iso - stability ] ) . | the loomis - whitney inequality , and the more general uniform cover inequality , bound the volume of a body in terms of a product of the volumes of lower - dimensional projections of the body . in this paper , we prove stability versions of these inequalities , showing that when they are close to being tight , the body in question is close in symmetric difference to a box .
our results are best possible up to a constant factor depending upon the dimension alone .
our approach is information theoretic .
we use our stability result for the loomis - whitney inequality to obtain a stability result for the edge - isoperimetric inequality in the infinite @xmath0-dimensional lattice .
namely , we prove that a subset of @xmath1 with small edge - boundary must be close in symmetric difference to a @xmath0-dimensional cube .
our bound is , again , best possible up to a constant factor depending upon @xmath0 alone . |
throughout this paper @xmath5 denotes an algebraically closed field and @xmath6 $ ] denotes the ring of polynomials ( in one variable ) over @xmath7 . for a given vector @xmath8 in @xmath9
, @xmath10 always denotes the @xmath11th coordinate of @xmath8 .
in particular , if @xmath12 is a map with values in @xmath9 , then @xmath13 denotes the @xmath11th coordinate of the vector @xmath14 .
we endow the vector space @xmath9 with a norm @xmath15 defined as usual : @xmath16 and with coordinate - wise ordering @xmath17 defined by @xmath18 for @xmath19 .
then @xmath20 is a solid vector space .
also , we endow @xmath9 with the cone norm @xmath21 defined by @xmath22 then @xmath23 is a cone normed space over @xmath24 ( see , e.g. , proinov @xcite ) .
let @xmath25}$ ] be a polynomial of degree @xmath26 .
a vector @xmath27 is said to be a root - vector of @xmath28 if @xmath29 for all @xmath30 , where @xmath31 .
we denote with @xmath32 the _ separation number _ of @xmath28 which is defined to be the minimum distance between two distinct zeros of @xmath28 . in the literature
, there are a lot of iterative methods for finding all zeros of @xmath28 simultaneously ( see , e.g. , the monographs of sendov , andreev and kjurkchiev @xcite , kjurkchiev @xcite , mcnamee @xcite and petkovi @xcite and references given therein ) . in 1891 , weierstrass @xcite published his famous iterative method for simultaneous computation of all zeros of @xmath28 .
the _ weierstrass method _ is defined by the following iteration @xmath33 where the operator @xmath34 is defined by @xmath35 where @xmath31 is the leading coefficient of @xmath28 and the domain @xmath36 of @xmath37 is the set of all vectors in @xmath9 with distinct components .
the weierstrass method has second - order of convergence provided that all zeros of @xmath28 are simple .
the operator @xmath38 is called _
weierstrass correction_. we should note that @xmath38 plays an important role in many semilocal convergence theorems for simultaneous methods . another famous iterative method for finding simultaneously all zeros of a polynomial @xmath28 was introduced by ehrlich @xcite in 1967 .
the _ ehrlich method _ is defined by the following fixed point iteration : @xmath39 where the operator @xmath40 is defined by @xmath41 with @xmath42 and the domain of @xmath43 is the set @xmath44 here and throughout the paper , we denote by @xmath45 the set of indices @xmath46 , that is @xmath47 .
the ehrlich method has third - order of convergence if all zeros of @xmath28 are simple .
the ehrlich method was rediscovered by abert @xcite in 1973 . in 1970
, brsch - supan @xcite introduced another third - order method for numerical computation of all zeros of a polynomial simultaneously . in 1982 , werner @xcite has proved that the both methods are identical . the ehrlich method is known also as `` ehrlich - abert method '' , `` brsch - supan method '' and `` abert method '' .
recently , proinov @xcite obtained two local convergence theorems for ehrlich method under different types of initial conditions . the first one generalizes and improves the results of kyurkchiev and tashev @xcite and wang and zhao ( * ? ? ?
* theorem 2.1 ) .
the second one generalizes and improves the results of wang and zhao ( * ? ? ?
* theorem 2.2 ) and tilli ( * ? ? ? * theorem 3.3 ) .
before we state the two results of @xcite , we need some notations which will be used throughout the paper . for given vectors @xmath48 and @xmath49
, we define in @xmath50 the vector @xmath51 provided that @xmath52 has no zero components .
given @xmath53 such that @xmath54 , we always denote by @xmath55 the conjugate exponent of @xmath53 , i.e. @xmath55 is defined by means of @xmath56 in the sequel , we use the function @xmath57 defined by @xmath58 with @xmath59 let @xmath60 and @xmath61 .
we define the real function @xmath62 by @xmath63 and the real number @xmath64 as follows @xmath65 [ thm : proinov - ehrlich - local - first ] let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros , @xmath27 be a root - vector of @xmath28 and @xmath54 .
suppose @xmath66 is an initial guess satisfying @xmath67 where @xmath68 and @xmath69 .
then ehrlich iteration is well - defined and converges cubically to @xmath70 with error estimates @xmath71 for all @xmath72 , where @xmath73 and the function @xmath62 is defined by .
[ thm : proinov - ehrlich - local - second ] let @xmath25}$ ] be a polynomial of degree @xmath26 , @xmath27 be a root - vector of @xmath28 and @xmath54 .
suppose @xmath66 is a vector with distinct components satisfying @xmath74 where @xmath68 and @xmath69 .
then @xmath28 has only simple zeros in @xmath7 and ehrlich iteration is well - defined and converges to @xmath70 with error estimates @xmath75 for all @xmath72 , where @xmath73 , @xmath76 and the function @xmath62 is defined by and the function @xmath77 by @xmath78 moreover , the method converges cubically to @xmath70 provided that @xmath79 . in the following definition ,
we define a sequence @xmath80 of iteration functions in the vector space @xmath9 . in what follows
, we define the binary relation @xmath81 on @xmath9 by @xmath82 [ def : iteration - functions ] let @xmath25}$ ] be a polynomial of degree @xmath26 .
define the sequence @xmath83 of functions @xmath84 recursively by setting @xmath85 and @xmath86 where the sequence of domains @xmath87 is also defined recursively by setting @xmath88 and @xmath89 given @xmath90 , the @xmath91th method of kjurkchiev - andreev s family can be defined by the following fixed - point iteration : @xmath92 it is easy to see that in the case @xmath93 the ehrlich - type method coincides with the classical ehrlich method .
the order of convergence of the ehrlich - type method is @xmath4 .
kjurkchiev and andreev @xcite established the following convergence result for the ehrlich - type methods .
this result and its proof can also be found in the monographs of sendov , andreev and kjurkchiev ( * ? ? ?
* section 19 ) and kjurkchiev ( * ? ? ? * chapter 9.2 ) ) .
[ thm : kjurkchiev - andreev ] let @xmath94}$ ] be a polynomial of degree @xmath26 which has only simple zeros , @xmath95 be a root - vector of @xmath28 and @xmath96 .
let @xmath97 and @xmath98 be such that @xmath99 where @xmath100 .
suppose @xmath101 is an initial guess satisfying the condition @xmath102 then the ehrlich - type method converges to @xmath70 with error estimate @xmath103 in this paper , we present two new local convergence theorems as well as a semilocal convergence theorem ( under computationally verifiable initial conditions and with a posteriori error estimate ) for ehrlich - type methods . our first local convergence result ( theorem [ thm : first - local - ehrlich ] ) generalizes theorem [ thm : proinov - ehrlich - local - first ] ( proinov @xcite ) and improves theorem [ thm : kjurkchiev - andreev ] ( kjurkchiev and andreev @xcite ) .
our second local convergence result ( theorem [ thm : second - local - ehrlich ] ) generalizes theorem [ thm : proinov - ehrlich - local - second ] ( proinov @xcite ) , but only in the case @xmath104 . furthermore ,
several numerical examples are provided to show some practical applications of our semilocal convergence result .
recently , proinov @xcite has developed a general convergence theory for iterative processes of the type @xmath105 where @xmath106 is an iteration function in a cone metric space @xmath107 . in order to make this paper
self - contained , we briefly review some basic definitions and results from this theory . throughout this paper @xmath108
denotes an interval on @xmath109 containing @xmath110 . for an integer @xmath111 ,
we denote by @xmath112 the following polynomial : @xmath113
if @xmath114 we assume that @xmath115 . throughout the paper we assume by definition that @xmath116 .
[ def : quasi - homogeneous - function ] a function @xmath117 is called _ quasi - homogeneous _ of degree @xmath118 on @xmath108 if it satisfies the following condition : @xmath119 \text { and } t \in j.\ ] ] if @xmath120 functions @xmath121 are quasi - homogeneous on @xmath108 of degree @xmath122 , then their product @xmath123 is a quasi - homogeneous function of degree @xmath124 on @xmath108 . note also that a function @xmath125 is quasi - homogeneous of degree @xmath110 on @xmath108 if and only it is nondecreasing on @xmath108 .
[ def : gauge - function - high - order ] a function @xmath126 is said to be a _ gauge function of order _
@xmath127 on @xmath108 if it satisfies the following conditions : 1 .
@xmath125 is quasi - homogeneous of degree @xmath128 on @xmath108 ; 2 .
@xmath129 for all @xmath130 .
a gauge function @xmath125 of order @xmath128 on @xmath108 is said to be a _ strict gauge function _ if the inequality in ( ii ) holds strictly whenever @xmath131 .
the following is a sufficient condition for a gauge function of order @xmath128 .
[ lem : gauge - function - condition ] if @xmath117 is a quasi - homogeneous function of degree @xmath127 on an interval @xmath108 and @xmath132 is a fixed point of @xmath125 in @xmath108 , then @xmath125 is a gauge function of order @xmath128 on @xmath133}$ ] .
moreover , if @xmath134 , then function @xmath125 is a strict gauge of order @xmath128 on @xmath135 .
[ def : function - of - initial - conditions ] let @xmath136 be a map on an arbitrary set @xmath107 .
a function @xmath137 is said to be a _ function of initial conditions _ of @xmath43 ( with a gauge function @xmath125 on @xmath108 ) if there exist a function @xmath126 such that @xmath138 [ def : initial - points ] let @xmath106 be a map on an arbitrary set @xmath107 , and let @xmath137 be a function of initial conditions of @xmath43 with a gauge function on @xmath108 .
then a point @xmath139 is said to be an _ initial point _ of @xmath43 ( with respect to @xmath140 ) if @xmath141 and all of the iterates @xmath142 @xmath143 are well - defined and belong to @xmath144 .
the following is a simple sufficient condition for initial points .
[ thm : initial - point - test ] let @xmath136 be a map on an arbitrary set @xmath107 and @xmath137 be a function of initial conditions of @xmath43 with a gauge function @xmath125 on @xmath108 .
suppose that @xmath139 with @xmath141 implies @xmath145 .
then every point @xmath146 such that @xmath147 is an initial points of @xmath43 .
[ def : iterated - contraction - map ] let @xmath106 be an operator in a cone normed space @xmath148 over a solid vector space @xmath149 , and let @xmath137 be a function of initial conditions of @xmath43 with a gauge function on an interval @xmath108 . then the operator @xmath43 is said to be _ an iterated contraction at a point _ @xmath150 ( with respect to @xmath140 ) if @xmath151 and @xmath152 where the control function @xmath153 is nondecreasing .
the following fixed point theorem plays an important role in our paper .
[ proinov @xcite ] [ thm : general - convergence - theorem ] let @xmath106 be an operator of a cone normed space @xmath148 over a solid vector space @xmath149 , and let @xmath137 be a function of initial conditions of @xmath43 with a gauge function @xmath125 of order @xmath127 on an interval @xmath108 .
suppose @xmath43 is an iterated contraction at a point @xmath70 with respect to @xmath140 with control function @xmath154 such that @xmath155 and there exist a function @xmath156 such that @xmath157 where @xmath158 is a nondecreasing function satisfying @xmath159 then the following statements hold true . 1 . the point @xmath70 is a unique fixed point of @xmath43 in the set @xmath160 .
2 . starting from each initial point @xmath161 of @xmath43 , picard iteration remains in the set @xmath162 and converges to @xmath70 with error estimates @xmath163 for all @xmath72 , where @xmath73 and @xmath76 .
in the case @xmath164 , theorem [ thm : general - convergence - theorem ] reduces to the following result .
[ @xcite ] [ cor : general - convergence - theorem ] let @xmath106 be an operator in a cone normed space @xmath148 over a solid vector space @xmath149 , and let @xmath137 be a function of initial conditions of @xmath43 with a strict gauge function @xmath125 of order @xmath127 on an interval @xmath108 .
suppose that @xmath43 is an iterated contraction at a point @xmath70 with respect to @xmath140 and with control function @xmath62 satisfying .
then the following statements hold true . 1 .
the point @xmath70 is a unique fixed point of @xmath43 in the set @xmath160 .
starting from each initial point @xmath161 of @xmath43 , picard iteration remains in @xmath162 and converges to @xmath70 with order @xmath128 and error estimates @xmath165 for all @xmath72 , where @xmath73 .
in this section , we present some useful inequalities in @xmath9 which play an important role in the paper .
[ @xcite ] [ lem : u - v - inequalities - basic ] let @xmath166 , @xmath167 be a vector with distinct components and @xmath54 .
then for all @xmath168 , @xmath169 @xmath170 [ lem : u - v - inequality - distinct - components ] let @xmath166 and @xmath54 .
if the vector @xmath167 has distinct components and @xmath171 then the vector @xmath172 also has distinct components .
[ @xcite ] [ lem : u - v - inequalities-1 ] let @xmath173 , @xmath70 be a vector with distinct components , @xmath54 and @xmath174 then for all @xmath168 , @xmath175 [ lem : u - v - inequalities-2 ] let @xmath173 , @xmath176 and @xmath54 .
if @xmath167 is a vector with distinct components such that @xmath177 then for all @xmath168 , @xmath178 by the triangle inequality of cone norm in @xmath9 and , we obtain @xmath179 which yields @xmath180 taking the @xmath53-norm , we get @xmath181 from and , we obtain which completes the proof .
[ lem : u - v - inequalities-3 ] let @xmath173 , @xmath176 and @xmath54 .
if @xmath167 is a vector with distinct components such that holds , then for all @xmath168 , @xmath182 from and , we get which completes the proof .
let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros in @xmath7 , and let @xmath27 be a root - vector of @xmath28 .
in this section we study the convergence of the ehrlich - type methods with respect to the function of initial conditions @xmath183 defined as follows @xmath184 let @xmath60 and @xmath61 . throughout this section , we define the function @xmath62 and the real number @xmath64 by and , respectively .
it is easy to show that @xmath64 is the unique solution of the equation @xmath185 in the interval @xmath186 , where @xmath187 .
note that @xmath62 is an increasing function which maps @xmath188 $ ] onto [ 0,1 ] .
besides , @xmath62 is quasi - homogeneous of degree @xmath189 on @xmath190}$ ] . in the next definition ,
we introduce a sequence of such functions .
[ def : phi - n ] we define the sequence @xmath191 of nondecreasing functions @xmath192 \to [ 0,1]$ ] recursively by setting @xmath193 and @xmath194 where @xmath60 and @xmath61 are constants .
we prove the correctness of the definition by induction .
for @xmath195 it is obvious .
assume that for some @xmath196 the function @xmath197 is well - defined and nondecreasing on @xmath190}$ ] and @xmath198 .
we shall prove the same for @xmath199 .
it follows from the induction hypothesis that @xmath200\ ] ] which means that the function @xmath201 is well - defined on @xmath133}$ ] . from and the induction hypothesis , we deduce that @xmath201 is nondecreasing on [ 0,r ] . from and @xmath198 ,
we obtain @xmath202 this completes the induction and the proof of the correctness of definition [ def : phi - n ] .
[ def : varphi - n ] for any integer @xmath196 , we define the function @xmath203 \to [ 0,r]}$ ] as follows @xmath204 where the function @xmath205 is defined by definition [ def : phi - n ] . in the next lemma
, we present some properties of the functions @xmath197 and @xmath206 .
[ lem : phi - n - properties ] let @xmath196 .
then : 1 .
@xmath197 is a quasi - homogeneous function of degree @xmath207 on @xmath190}$ ] ; 2 .
@xmath208 for every @xmath209}$ ] ; 3 .
@xmath210 for every @xmath209}$ ] ; 4 .
@xmath211 for every @xmath209}$ ] ; 5 .
@xmath206 is a gauge function of order @xmath4 on @xmath190}$ ] .
claim ( i ) can easily be proved by induction . from and
, we get @xmath212 which proves ( ii ) . claim ( iii )
is a trivial consequence from ( ii ) . claim ( iv ) follows from ( ii ) by induction .
claim ( v ) follows from ( i ) and the definition of @xmath206 .
[ lem : ehrlich - high - order - map ] let @xmath25}$ ] be a polynomial of degree @xmath26 , @xmath27 be a root - vector of @xmath28 and @xmath196 .
suppose @xmath213 is a vector such that @xmath214 for some @xmath215 . _
( i ) _ if @xmath216 , then @xmath217 where @xmath218 is defined by @xmath219 _ ( ii ) _ if @xmath220 , then @xmath221 \(i ) taking into account that @xmath70 is a root - vector of @xmath28 , we get @xmath222 which proves .
\(ii ) it follows from @xmath220 that @xmath223 then from and , we obtain @xmath224 which completes the proof .
[ lem : ehrlich - iterated - contraction ] let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros in @xmath7 , @xmath27 be a root - vector of @xmath28 , @xmath196 and @xmath54 .
suppose @xmath48 is a vector satisfying the following condition @xmath225 where the function @xmath183 is defined by , @xmath68 and @xmath69 .
then @xmath226 we shall prove statements by induction on @xmath91 .
if @xmath227 , then holds trivially .
assume that holds for some @xmath196 .
first , we show that @xmath220 , i.e. @xmath228 and holds for every @xmath215 .
it follows from the first inequality in that the inequality is satisfied with @xmath229 and @xmath230 . then by lemma [ lem : u - v - inequalities-1 ] and
, we obtain @xmath231 for every @xmath232 .
consequently , @xmath228 .
it remains to prove for every @xmath215 .
let @xmath215 be fixed .
we shall consider only the non - trivial case @xmath214 . in this case
is equivalent to @xmath233 we define @xmath234 by .
it follows from lemma [ lem : ehrlich - high - order - map](i ) that is equivalent to @xmath235 . by lemma [ lem : u - v - inequalities - basic ] with @xmath229 and @xmath236 and
, we get @xmath237 for every @xmath232 . from the triangle inequality in @xmath7 , , , induction hypothesis and hlder s inequality
, we get @xmath238 from this , @xmath239 and , we obtain @xmath240 which yields @xmath235 and so holds .
hence , @xmath241 .
second , we show that the inequalities in hold for @xmath242 . the first inequality for @xmath242 is equivalent to @xmath243 let @xmath215 be fixed . if @xmath244 , then @xmath245 and so becomes an equality .
suppose @xmath246 .
by lemma [ lem : ehrlich - high - order - map](ii ) , the triangle inequality in @xmath7 and the estimate , we get @xmath247 which proves . dividing both sides of the inequality by @xmath248 and taking the @xmath53-norm
, we obtain @xmath249 which proves that the second inequality in holds for @xmath242 .
this completes the induction and the proof of the lemma .
now we are ready to state the main result of this section . in the case
@xmath93 this result coincides with theorem [ thm : proinov - ehrlich - local - first ] .
[ thm : first - local - ehrlich ] let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros in @xmath7 , @xmath27 be a root - vector of @xmath28 , @xmath96 and @xmath54 .
suppose @xmath66 is an initial guess satisfying @xmath250 where the function @xmath183 is defined by , @xmath68 and @xmath69 .
then the ehrlich - type iteration is well - defined and converges to @xmath70 with error estimates @xmath251 for all @xmath72 , where @xmath252 and the function @xmath197 is defined by definition [ def : phi - n ] .
we apply corollary [ cor : general - convergence - theorem ] to the iteration function @xmath84 defined by definition [ def : iteration - functions ] and to the function @xmath183 defined by .
let @xmath135 .
it follows from the lemma [ lem : ehrlich - iterated - contraction ] , lemma [ lem : phi - n - properties](v ) and lemma [ lem : gauge - function - condition ] that @xmath140 is a function of initial conditions of @xmath3 with a strict gauge function @xmath206 of order @xmath253 on @xmath108 . since @xmath70 is a root - vector of @xmath28 , then @xmath254 .
it follows from lemma [ lem : ehrlich - iterated - contraction ] , that @xmath255 is an iterated contraction at a point @xmath70 with respect to @xmath140 and with control function @xmath197 .
the fact that @xmath161 is an initial point of @xmath255 follows from lemma [ lem : ehrlich - iterated - contraction ] and theorem [ thm : initial - point - test ] .
hence , all the assumptions of corollary [ cor : general - convergence - theorem ] are satisfied , and the statement of theorem [ thm : first - local - ehrlich ] follows from it .
[ cor : local - ehrlich ] let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros in @xmath7 , @xmath27 be a root - vector of @xmath28 , @xmath96 and @xmath54 .
suppose @xmath66 is an initial guess satisfying .
then the ehrlich - type iteration is well - defined and converges to @xmath70 with error estimates @xmath256 for all @xmath72 , where @xmath73 and @xmath62 is a real function defined by .
it follows from theorem [ thm : first - local - ehrlich ] and lemma [ lem : phi - n - properties](iv ) .
let @xmath97 be a given number . solving the equation @xmath257 in the interval @xmath258
, we can reformulate corollary [ cor : local - ehrlich ] in the following equivalent form .
[ cor : first - local - ehrlich - var ] let @xmath25}$ ] be a polynomial of degree @xmath259 which has @xmath260 simple zeros in @xmath7 , @xmath27 be a root - vector of @xmath28 , @xmath96 , @xmath54 and @xmath97 .
suppose @xmath66 is an initial guess which satisfies @xmath261 where @xmath68 and @xmath69 .
then the ehrlich - type method is well - defined and converges to @xmath70 with error estimates @xmath262 for all @xmath72 .
corollary [ cor : first - local - ehrlich - var ] is an improvement of the result of kjurkchiev and andreev @xcite ( see theorem [ thm : kjurkchiev - andreev ] above ) .
suppose that a vector @xmath66 satisfies .
it is easy to show that condition is equivalent to the following one : @xmath263 from this , the initial condition and @xmath97 , we obtain @xmath264 therefore , @xmath265 satisfies with @xmath266 .
then it follows from corollary [ cor : first - local - ehrlich - var ] that the ehrlich - type method is well - defined and converges to @xmath70 with error estimates . from the second estimate in and
, we get the estimate which completes the proof .
let @xmath25}$ ] be a polynomial of degree @xmath259 .
we study the convergence of the ehrlich - type method with respect to the function of initial conditions @xmath267 defined by @xmath268 in the previous section , we introduce the functions @xmath197 , @xmath206 and the real number @xmath64 with two parameters @xmath60 and @xmath61 . in this section
, we consider a special case of @xmath197 , @xmath206 and @xmath64 when @xmath269 .
in other words , now we define @xmath64 by @xmath270 furthermore , we define the functions @xmath197 and @xmath206 by definitions [ def : phi - n ] and [ def : varphi - n ] , respectively , but with @xmath271 instead of , where @xmath60 is a constant .
[ def : beta - psi - n ] for a given integer @xmath2 , we define the increasing function @xmath272 \to [ 0,1)}$ ] by @xmath273 and we define the decreasing function @xmath274 \to ( 0,1]}$ ] as follows @xmath275 the functions @xmath276 and @xmath277 are well - defined on @xmath133}$ ] since @xmath278.\ ] ] the monotonicity of @xmath276 and @xmath277 is obvious . it remains to prove that @xmath279 and @xmath280 .
since @xmath281 , we obtain @xmath282 which completes the proof of the correctness of definition [ def : beta - psi - n ] [ lem : beta - psi - n - properties ] let @xmath96 .
then : 1 .
@xmath283 is a quasi - homogeneous of degree @xmath207 on @xmath190}$ ] ; 2 .
@xmath284 for every @xmath209}$ ] ; 3 .
@xmath285 for every @xmath209}$ ] ; 4 .
@xmath286 for every @xmath209}$ ] .
the function @xmath276 can be presented in the form @xmath287 , where @xmath288 .
therefore , @xmath276 is quasi - homogeneous of degree @xmath207 on @xmath133}$ ] since it is a product of three quasi - homogeneous functions on @xmath133}$ ] of degree @xmath189 , @xmath289 and @xmath110 . from the definitions of the functions @xmath197 , @xmath290 and @xmath283
, we get @xmath291 claim ( iii ) follows from lemma [ lem : phi - n - properties](iii ) and .
claim ( iv ) follows from ( iii ) and .
[ lem : ehrlich - iterated - contraction - second - fic ] let @xmath25}$ ] be a polynomial of degree @xmath26 which has only simple zeros in @xmath7 , @xmath27 a root - vector of @xmath28 , @xmath96 and @xmath54 .
suppose @xmath48 is a vector with distinct components such that @xmath292 where the function @xmath267 is defined by and @xmath68 .
then @xmath28 has only simple zeros in @xmath7 , @xmath293 besides , the vector @xmath294 has pairwise distinct components .
it follows from and @xmath295 that @xmath296 .
then it follows from lemma [ lem : u - v - inequality - distinct - components ] that the vector @xmath70 has distinct components , which means that @xmath28 has only simple zeros in @xmath7 .
we divide the proof into two steps . in this step ,
we prove @xmath297 and the first inequality in by induction on @xmath91 . if @xmath93 , the proof of the claims can be found in @xcite .
assume that @xmath297 and the first inequality in hold for some @xmath96 .
first we show that @xmath220 i.e. @xmath228 and holds for every @xmath215 .
it follows from the first inequality in that holds with @xmath298 , @xmath299 and @xmath300 .
therefore by lemma [ lem : u - v - inequalities-2 ] , and @xmath301 , we obtain @xmath302 for every @xmath232 . consequently , @xmath228 .
it remains to prove for every @xmath215 .
let @xmath215 be fixed .
we shall consider only the non - trivial case @xmath214 . in this case
is equivalent to . on the other hand
, it follows from lemma [ lem : ehrlich - high - order - map](i ) that is equivalent to @xmath235 , where @xmath303 is defined by . by lemma [ lem : u - v - inequalities - basic ] with @xmath304 and @xmath298 and
, we get @xmath305 for every @xmath232 .
hence , we obtain @xmath306 . from induction hypothesis , we get @xmath307 combining the triangle inequality in @xmath7 , , and , we obtain @xmath308 which , using hlder s inequality , yields @xmath309 which yields @xmath235 , and so holds . thus we prove that @xmath241 .
now we have to prove that the first inequality in is satisfied for @xmath242 , which is equivalent to @xmath310 let @xmath215 be fixed . if @xmath244 , then @xmath245 and the inequality becomes an equality .
suppose @xmath246 .
it follows from lemma [ lem : ehrlich - high - order - map](ii ) , the triangle inequality in @xmath7 and the estimate that @xmath311 from this inequality , lemma [ lem : beta - psi - n - properties](ii ) , @xmath312 , and lemma [ lem : beta - psi - n - properties](iv ) , we obtain @xmath313 which proves .
this completes the induction . in this step
we prove the second inequality in and that @xmath294 has distinct components .
first inequality in allow us to apply lemma [ lem : u - v - inequalities-3 ] with @xmath299 , @xmath298 and @xmath314 . by lemma
[ lem : u - v - inequalities-3 ] and , we deduce @xmath315 by taking the minimum over all @xmath316 such that @xmath232 , we obtain @xmath317 which implies that @xmath294 has distinct components .
it follows from , and lemma [ lem : beta - psi - n - properties](ii ) that @xmath318 by taking the @xmath53-norm , we obtain @xmath319 which proves the second inequality in .
this completes the proof .
now we are able to state the main result of this section . in the case when @xmath93 and @xmath104 this result reduces to theorem [ thm : proinov - ehrlich - local - second ] .
[ thm : second - local - ehrlich ] let @xmath25}$ ] be a polynomial of degree @xmath26 which splits over @xmath7 , @xmath27 be a root - vector of @xmath28 , @xmath96 and @xmath54 .
suppose @xmath66 is an initial guess with distinct components such that @xmath320 where the function @xmath140 is defined by and @xmath68 .
then @xmath28 has only simple zeros in @xmath7 and the ehrlich - type iteration is well - defined and converges to @xmath70 with error estimates @xmath321 for all @xmath72 , where @xmath252 , @xmath322 .
moreover , the method is convergent with order @xmath323 provided that @xmath79 .
we apply theorem [ thm : general - convergence - theorem ] to the iteration function @xmath84 together with the function @xmath324 defined by .
it follows from lemma [ lem : ehrlich - iterated - contraction - second - fic ] and lemma [ lem : phi - n - properties](v ) that @xmath140 is a function of initial conditions of @xmath255 with gauge function @xmath206 of order @xmath253 on the interval @xmath325}$ ] . from lemma [
lem : ehrlich - iterated - contraction - second - fic ] , we get that @xmath255 is an iterated contraction at @xmath70 with respect to @xmath140 and with control function @xmath276 . also , it is easy to see that the functions @xmath283 , @xmath197 , @xmath290 and @xmath206 have the properties , and .
it follows from lemma [ lem : ehrlich - iterated - contraction - second - fic ] that @xmath326 . according to theorem [ thm : initial - point - test ] to prove that @xmath161 is an initial point of @xmath255 it is sufficient to prove that @xmath327 from @xmath213
, we have @xmath328 . by lemma [ lem : ehrlich - iterated - contraction - second - fic ]
, @xmath255 has distinct components and @xmath329 .
the last inequality yields @xmath330 since @xmath126 and @xmath141 .
thus we have both @xmath331 and @xmath330 .
applying lemma [ lem : ehrlich - iterated - contraction - second - fic ] to the vector @xmath332 , we get @xmath333 which proves .
therefore , @xmath161 is an initial point of @xmath255 .
now the statement of theorem [ thm : second - local - ehrlich ] follows from theorem [ thm : general - convergence - theorem ] .
in this section we establish semilocal convergence theorems for ehrlich - type methods for finding all zeros of a polynomial simultaneously .
we study the convergence of these methods with respect to the function of initial conditions @xmath267 defined by @xmath334 recently proinov @xcite has shown that there is a relationship between local and semilocal theorems for simultaneous root - finding methods .
it turns out that from any local convergence theorem for a simultaneous method one can obtain as a consequence a semilocal theorem for the same method .
in particular , from theorem [ thm : first - local - ehrlich ] we can obtain a semilocal convergence theorem for ehrlich - type methods under computationally verifiable initial conditions .
for this purpose we need the following result .
[ prop : semilocal - theorem - second - kind ] let @xmath25}$ ] be a polynomial of degree @xmath26 .
suppose @xmath48 is an initial guess with distinct components such that @xmath335 for some @xmath54 and @xmath336 , where @xmath337 .
in the case @xmath338 and @xmath104 we assume that inequality in is strict .
then @xmath28 has only simple zeros in @xmath7 and there exists a root - vector @xmath27 of @xmath28 such that @xmath339 where the real function @xmath340 is defined by @xmath341 if the inequality is strict , then the second inequality in is strict too .
now , we are ready to state and prove the main result of this paper .
[ thm : semilocal - theorem - ehrlich - second ] let @xmath25}$ ] be a polynomial of degree @xmath26 , @xmath96 , @xmath54 .
suppose @xmath66 is an initial guess with distinct components such that @xmath342 where the function @xmath343 is defined by and @xmath337 .
then @xmath28 has only simple zeros in @xmath7 and the ehrlich - type iteration is well - defined and converges to a root - vector @xmath70 of @xmath28 with order of convergence @xmath323 and with a posteriori error estimate @xmath344 for all @xmath72 such that @xmath345 , where the function @xmath340 is defined by .
let us define @xmath64 by .
it is easy to calculate that @xmath346 and @xmath347 therefore , can be written in the form @xmath348 then it follows from theorem [ prop : semilocal - theorem - second - kind ] that @xmath28 has only simple zeros in @xmath7 and there exists a root - vector @xmath27 of @xmath28 such that @xmath349 now theorem [ thm : second - local - ehrlich ] implies that the ehrlich - type iteration converges to @xmath70 with order of convergence @xmath323 .
it remains to prove the error estimate .
suppose that for some @xmath72 , @xmath350 then it follows from theorem [ prop : semilocal - theorem - second - kind ] that there exists a root - vector @xmath351 of @xmath28 such that @xmath352 from the second inequality in and theorem [ thm : second - local - ehrlich ] , we conclude that the ehrlich - type iteration converges to @xmath353 . by the uniqueness of the limit , we get @xmath354
. therefore , the error estimate follows from the first inequality in .
this completes the proof . setting @xmath104 in theorem [ thm : semilocal - theorem - ehrlich - second ]
, we obtain the following result .
[ cor : semilocal - theorem - ehrlich - infinity ] let @xmath25}$ ] be a polynomial of degree @xmath26 and @xmath96 .
suppose @xmath355 is an initial guess with distinct components such that @xmath356 then @xmath28 has only simple zeros in @xmath7 and the ehrlich - type iteration is well - defined and converges to a root - vector @xmath70 of @xmath28 with order of convergence @xmath323 and with error estimate for @xmath104 .
setting @xmath357 in theorem [ thm : semilocal - theorem - ehrlich - second ] we obtain the following result .
[ cor : semilocal - theorem - ehrlich-1 ] let @xmath25}$ ] be a polynomial of degree @xmath26 and @xmath96 .
suppose @xmath66 is an initial guess with distinct components such that @xmath358 then @xmath28 has only simple zeros in @xmath7 and the ehrlich - type iteration is well - defined and converges with order @xmath323 to a root - vector @xmath70 of @xmath28 with error estimate for @xmath357 .
in this section , we present several numerical examples to show some applications of theorem [ thm : semilocal - theorem - ehrlich - second ] .
let @xmath94}$ ] be a polynomial of degree @xmath26 and let @xmath359 be an initial guess .
we show that theorem [ thm : semilocal - theorem - ehrlich - second ] can be used : * to prove numerically that @xmath28 has only simple zeros ; * to prove numerically that @xmath91th ehrlich - type iteration starting from @xmath161 is well - defined and converges with order @xmath323 to a root - vector of @xmath28 ; * to guarantee the desired accuracy when calculating the roots of @xmath28 via @xmath91th ehrlich - type method . in the examples below
, we use the function of initial conditions @xmath360 defined by @xmath361 where @xmath38 is the weierstrass correction defined by .
we consider only the case @xmath104 since the other cases are similar .
also , we use the real function @xmath340 defined by @xmath362 it follows from theorem [ thm : semilocal - theorem - ehrlich - second ] that if there exists an integer @xmath363 such that @xmath364 then @xmath28 has only simple zeros and the ehrlich - type iteration is well - defined and converges to a root - vector @xmath70 of @xmath28 with order of convergence @xmath323 . besides , for all @xmath365 such that @xmath366 the following a posteriori error estimate holds : @xmath367 in the examples , we apply the ehrlich - type methods for some @xmath96 using the following stopping criterion : @xmath368 for given @xmath91 we calculate the smallest @xmath369 which satisfies the convergence condition , the smallest @xmath370 for which the stopping criterion is satisfied , as well as the value of @xmath371 for the last @xmath372 . in table 2
the values of iterations are given to 15 decimal places .
the values of other quantities ( @xmath373 , @xmath374 , etc . ) are given to 6 decimal places .
[ exmp : zhph-06 ] we consider the polynomial @xmath375 and the initial guess @xmath376 which are taken from zhang et al .
we have @xmath377 and @xmath378 .
the results for this example are presented in table [ tab : zhph - example ] .
for example , we can see that for @xmath379 at the first iteration we have proved that the ehrlich - type method converges with order of convergence @xmath380 and that at the second iteration we have calculated the zeros @xmath28 with accuracy less than @xmath381 .
moreover , at the next iteration we obtain the zeros of @xmath28 with accuracy less than @xmath382 .
also , we can see that for @xmath383 at the second iteration we have obtained the zeros of @xmath28 with accuracy less than @xmath384 .
c c c c|c l l @xmath91 & @xmath120 & @xmath385 & @xmath386 & @xmath372 & & + + [ -2.0ex ] 1 & @xmath189 & @xmath387 & @xmath388 & @xmath389 & @xmath390 & @xmath391 + 2 & @xmath392 & @xmath393 & @xmath394 & @xmath395 & @xmath396 & @xmath397 + 3 & @xmath392 & @xmath398 & @xmath399 & @xmath395 & @xmath400 & @xmath401 + 4 & @xmath392 & @xmath402 & @xmath403 & @xmath189 & @xmath404 & @xmath405 + 5 & @xmath392 & @xmath406 & @xmath407 & @xmath189 & @xmath408 & @xmath409 + 6 & @xmath392 & @xmath410 & @xmath411 & @xmath189 & @xmath412 & @xmath413 + 7 & @xmath392 & @xmath414 & @xmath415 & @xmath189 & @xmath416 & @xmath417 + 8 & @xmath392 & @xmath418 & @xmath419 & @xmath189 & @xmath420 & @xmath421 + 9 & @xmath392 & @xmath422 & @xmath423 & @xmath189 & @xmath424 & @xmath425 + 10 & @xmath392 & @xmath426 & @xmath427 & @xmath189 & @xmath428 & @xmath429 + 100 & @xmath392 & @xmath426 & @xmath430 & @xmath392 & @xmath431 & @xmath432 + in table [ tab : zhph - example - weierstrass - numerical - results - n=10 ] , we present numerical results for example [ exmp : zhph-06 ] in the case @xmath379 .
l l l @xmath372 & & + @xmath110 & @xmath433 & @xmath434 + @xmath392 & @xmath435 & @xmath436 + @xmath189 & @xmath437 & @xmath438 + @xmath372 & & + @xmath110 & @xmath439 & @xmath440 + @xmath392 & @xmath441 & @xmath442 + @xmath189 & @xmath443 & @xmath444 + [ exmp : petilpet ] we consider the polynomial @xmath445 and aberth s initial approximation @xmath101 given by ( see aberth @xcite and petkovi et al .
@xcite ) : @xmath446 where @xmath447 , @xmath448 and @xmath449 .
we have @xmath450 and @xmath451 .
the results for this example are presented in table [ tab : petilpet - example ] .
for example , we can see that for @xmath452 at the third iteration we have obtained the zeros of @xmath28 with accuracy less than @xmath453 .
moreover , at the next iteration we get the zeros of @xmath28 with accuracy less than @xmath454 .
l l @xmath91 & @xmath120 & @xmath385 & @xmath386 & @xmath372 & & + + [ -2.0ex ] 1 & @xmath455 & @xmath456 & @xmath457 & @xmath458 & @xmath459 & @xmath460 + 2 & @xmath461 & @xmath462 & @xmath463 & @xmath455 & @xmath464 & @xmath465 + 3 & @xmath389 & @xmath466 & @xmath467 & @xmath461 & @xmath468 & @xmath469 + 4 & @xmath389 & @xmath426 & @xmath470 & @xmath461 & @xmath471 & @xmath472 + 5 & @xmath395 & @xmath473 & @xmath474 & @xmath389 & @xmath475 & @xmath476 + 6 & @xmath395 & @xmath477 & @xmath478 & @xmath389 & @xmath479 & @xmath480 + 7 & @xmath395 & @xmath422 & @xmath481 & @xmath389 & @xmath482 & @xmath483 + 8 & @xmath395 & @xmath426 & @xmath484 & @xmath389 & @xmath485 & @xmath486 + 9 & @xmath395 & @xmath426 & @xmath487 & @xmath389 & @xmath488 & @xmath489 + 10 & @xmath395 & @xmath426 & @xmath490 & @xmath389 & @xmath491 & @xmath492 + 30 & @xmath189 & @xmath493 & @xmath494 & @xmath395 & @xmath495 & @xmath496 + [ exmp : wilkinson - polynomial ] we consider the wilkinson polynomial ( @xcite ) @xmath497 and abert s initial approximation with @xmath498 , @xmath499 and @xmath500 .
we have @xmath501 and @xmath502 .
the results foe example [ exmp : wilkinson - polynomial ] are shown in table [ tab : wilkinson - aberth - example ] .
for example ,
we for @xmath383 at the seventh iteration we get the zeros of @xmath28 with accuracy less than @xmath503 .
c c c c|c l l @xmath91 & @xmath120 & @xmath385 & @xmath386 & @xmath372 & & + + [ -2.0ex ] 1 & @xmath504 & @xmath505 & @xmath506 & @xmath507 & @xmath508 & @xmath509 + 2 & @xmath510 & @xmath511 & @xmath512 & @xmath513 & @xmath514 & @xmath515 + 3 & @xmath516 & @xmath517 & @xmath518 & @xmath510 & @xmath519 & @xmath520 + 4 & @xmath458 & @xmath521 & @xmath522 & @xmath516 & @xmath523 & @xmath524 + 5 & @xmath458 & @xmath426 & @xmath525 & @xmath516 & @xmath526 & @xmath527 + 6 & @xmath528 & @xmath529 & @xmath530 & @xmath458 & @xmath531 & @xmath532 + 7 & @xmath528 & @xmath426 & @xmath533 & @xmath528 & @xmath534 & @xmath535 + 8 & @xmath536 & @xmath537 & @xmath538 & @xmath528 & @xmath539 & @xmath540 + 9 & @xmath536 & @xmath541 & @xmath542 & @xmath528 & @xmath543 & @xmath544 + 10 & @xmath536 & @xmath426 & @xmath545 & @xmath528 & @xmath546 & @xmath547 + 30 & @xmath461 & @xmath548 & @xmath549 & @xmath455 & @xmath550 & @xmath551 + in the figure [ fig : wilkinson - example - ehrlich ] , we present the trajectories of approximations generated by the method for @xmath452 after @xmath455 iterations .
[ exmp : wang - zhao-40 ] we consider the polynomial @xmath552 in this example we use abert s initial approximation with @xmath553 , @xmath448 and @xmath554 .
we have @xmath555 , @xmath556 . the results for example [ exmp : wang - zhao-40 ]
can be seen in table [ tab : wang - zhao - example - abert - power-40 ] .
c c c c|c l l @xmath91 & @xmath120 & @xmath385 & @xmath386 & @xmath372 & & + + [ -2.0ex ] 1 & @xmath557 & @xmath558 & @xmath559 & @xmath560 & @xmath561 & @xmath562 + 2 & @xmath563 & @xmath564 & @xmath565 & @xmath510 & @xmath566 & @xmath567 + 3 & @xmath458 & @xmath568 & @xmath569 & @xmath516 & @xmath570 & @xmath571 + 4 & @xmath528 & @xmath572 & @xmath573 & @xmath458 & @xmath574 & @xmath575 + 5 & @xmath536 & @xmath576 & @xmath577 & @xmath528 & @xmath578 & @xmath579 + 6 & @xmath536 & @xmath580 & @xmath581 & @xmath528 & @xmath582 & @xmath583 + 7 & @xmath536 & @xmath426 & @xmath584 & @xmath528 & @xmath585 & @xmath586 + 8 & @xmath455 & @xmath587 & @xmath588 & @xmath536 & @xmath589 & @xmath590 + 9 & @xmath455 & @xmath591 & @xmath592 & @xmath536 & @xmath593 & @xmath594 + 10 & @xmath455 & @xmath426 & @xmath595 & @xmath536 & @xmath596 & @xmath597 + 30 & @xmath461 & @xmath426 & @xmath598 & @xmath461 & @xmath599 & @xmath600 + in the figure [ fig : wang - zhao - example - ehrlich ] , we present the trajectories of approximations generated by the method for @xmath452 after @xmath461 iterations .
the authors declare that they have no competing interests .
both authors contributed equally and significantly in writing this paper .
both authors read and approved the final manuscript .
this research is supported by the project ni15-fmi-004 of plovdiv university .
: neuer beweis des satzes , dass jede ganze rationale function einer vernderlichen dargestellt werden kann als ein product aus linearen functionen derselben vernderlichen .
berlin , 10851101 ( 1891 ) | we study a family of high order ehrlich - type methods for approximating all zeros of a polynomial simultaneously .
let us denote by @xmath0 the famous ehrlich method ( 1967 ) .
starting from @xmath0 , kjurkchiev and andreev ( 1987 ) have introduced recursively a sequence @xmath1 of iterative methods for simultaneous finding polynomial zeros . for given @xmath2 ,
the ehrlich - type method @xmath3 has the order of convergence @xmath4 . in this paper , we establish two new local convergence theorems as well as a semilocal convergence theorem ( under computationally verifiable initial conditions and with a posteriori error estimate ) for the ehrlich - type methods @xmath3 .
our first local convergence theorem generalizes a result of proinov ( 2015 ) and improves the result of kjurkchiev and andreev ( 1987 ) .
the second local convergence theorem generalizes another recent result of proinov ( 2015 ) , but only in the case of maximum - norm .
our semilocal convergence theorem is the first result in this direction . |
many quantum algorithms may be modelled as a quantum process occurring on a graph . in @xcite
, childs shows that any quantum computation can be encoded as a quantum walk in some graph and thus quantum walks can be regarded as a quantum computation primitive .
we model a network of @xmath6 interacting qubits by a simple graph @xmath7 on @xmath6 vertices , where vertices correspond to qubits and edges to interactions .
the interactions are defined by a time - independent hamiltonian ; that is , a symmetric matrix that acts on the hilbert space of dimension @xmath8 . in this paper , we are concerned with the @xmath9-hamiltonian , whose action on the @xmath10-excitation subspaces is equivalent to the action of the @xmath11-symmetric adjacency matrix of @xmath7 on @xmath12 .
assume all qubits are initialized at the state @xmath13 , except for the qubit corresponding to vertex @xmath3 , which is initialized at state @xmath14 .
let @xmath15 be the adjacency matrix of @xmath7 . according to the schrdinger equation
, @xmath16 indicates the probability that the state @xmath17 is measured at @xmath18 after time @xmath1 .
whenever there is a time @xmath1 such that this probability is equal to @xmath10 , we say that there is _ perfect state transfer _ between @xmath3 and @xmath18 .
perfect state transfer has been studied in many families of graphs , including circulants @xcite , cubelike graphs @xcite and distance regular graphs @xcite .
see @xcite for a survey of recent results . for paths ( or linear chains ) with @xmath6 vertices , perfect state transfer occurs if and only if @xmath6 is @xmath19 or @xmath20 . in general ,
perfect state transfer is understood to be a rare phenomenon , which motivates the definition of _ pretty good state transfer _ ( or _ almost state transfer _ ) , which is said to occur if , for any @xmath21 , there is a time @xmath1 such that @xmath22 pretty good state transfer was first studied for paths in @xcite and @xcite .
proofs and methods in this area are a mix of interesting applications of number theory and algebraic graph theory .
pretty good state transfer between the end vertices in paths with @xmath6 elements occurs if and only if @xmath23 is prime , twice a prime or a power of @xmath19 ( see @xcite ) .
moreover , in these cases , it occurs between any pair of vertices equally distant from the centre of the path .
in contrast , the question of whether pretty good state transfer is possible between internal vertices in paths when it does not occur between the end vertices gives rise to a more interesting story and does not appear to have a simple answer .
this question was raised in @xcite . in this paper
, we exhibit an infinite family of paths that admit pretty good state transfer between inner vertices but not between the two end vertices .
these paths have @xmath24 vertices where @xmath0 is an odd prime and @xmath25 .
we will make use of the following definitions and preliminary results .
given a symmetric matrix @xmath26 with @xmath27 distinct eigenvalues @xmath28 , we may write the spectral decomposition of @xmath26 as follows : @xmath29 where @xmath30 denotes the idempotent projection onto the eigenspace corresponding to @xmath31 .
note that @xmath32 . given a vertex @xmath33 ,
let @xmath34 denote the @xmath11-vector that is @xmath10 at the entry corresponding to @xmath3 , and @xmath35 elsewhere .
we define the _ eigenvalue support _ of @xmath3 as a subset of the distinct eigenvalues as follows : @xmath36 we say that vertices @xmath3 and @xmath18 are _ strongly cospectral _ if @xmath37 for all idempotents @xmath38 in the spectral decomposition .
this property is necessary for both perfect and pretty good state transfer ( see ( * ? ? ?
* lemma 3 ) ) .
the spectrum of the adjacency matrix of the path ( see @xcite for example ) on @xmath6 vertices is @xmath39 note that they are indexed such that @xmath40 . for each @xmath41 ,
the eigenvector corresponding to @xmath42 is given by @xmath43 where @xmath44 .
the following lemma immediately follows .
[ eq:3 ] vertices @xmath3 and @xmath18 of @xmath45 are strongly cospectral if and only if @xmath46 .
we will make use of the following result , which is derived from ( * ? ? ? *
theorem 2 ) , and completely characterizes pretty good state transfer .
the second condition is an immediate consequence of kronecker s theorem ( see ( * ? ? ?
* theorem 4 ) ) .
@xcite [ thm:1 ] let @xmath3 and @xmath18 be vertices in a path with @xmath6 vertices .
then pretty good state transfer happens between @xmath3 and @xmath18 if and only if both conditions below hold .
a. @xmath47 . in this case , for all @xmath48 , define @xmath49 if @xmath50 , and @xmath51 if @xmath52 . b. for any set of integers @xmath53 such that @xmath54 then @xmath55
the path graph @xmath45 is the graph on vertices @xmath56 where vertex @xmath3 is adjacent to @xmath57 for all @xmath58 .
[ thm : main ] given any odd prime @xmath0 and positive integer @xmath1 , there is pretty good state transfer in @xmath59 between vertices @xmath3 and @xmath60 , whenever @xmath3 is a multiple of @xmath5 . for simplicity ,
let @xmath61 . for vertices @xmath3 and @xmath4 , condition ( i ) of theorem [ thm:1 ] is satisfied with @xmath62 , by lemma [ eq:3 ] .
the eigenvalues of the path @xmath45 belong to the cyclotomic field @xmath63 $ ] , where @xmath64 .
more precisely , @xmath65 if @xmath66 , then the cyclotomic polynomial is @xmath67 we will proceed by showing that part ( ii ) of theorem [ thm:1 ] holds .
if @xmath3 is a multiple of @xmath5 , suppose there is a linear combination of the eigenvalues in @xmath68 , satisfying @xmath69 where we make @xmath70 if @xmath71 .
recall that the @xmath3th entry of the @xmath42-eigenvector is @xmath72 .
we see that @xmath42 belongs to @xmath73 if and only if @xmath74 does not divide @xmath41 .
we define the polynomial @xmath75 as follows : @xmath76 we see that @xmath77 is a root of @xmath75 and , since @xmath78 is the minimal polynomial of @xmath77 , we see that @xmath78 divides @xmath75 .
let @xmath79 be the following polynomial : @xmath80 consider @xmath81 \phi_{2m}(x ) q(x)$ ] .
it is easy to see that @xmath81 \phi_{2m}(x ) q(x ) = [ x^k ] p(x)$ ] for @xmath82 .
since the degree of @xmath79 is @xmath83 , we may conclude that @xmath79 is the unique polynomial of degree @xmath83 such that @xmath84 \phi_{2m}(x ) q(x ) = [ x^k ] p(x)\ ] ] for @xmath82 .
in particular , the quotient @xmath85 is a polynomial of degree @xmath83 such that ( [ eq : eq1 ] ) holds , therefore @xmath86 from the coefficients of @xmath87 for @xmath88 , it follows that , for @xmath89 and @xmath90 , @xmath91 recall that @xmath92 for any integer @xmath93 .
given @xmath94 , note that @xmath95 , and since @xmath96 , there is @xmath97 such that @xmath98 . if @xmath99 , then @xmath100 , and if @xmath101 , then @xmath102 .
in either case , it follows that @xmath103 .
therefore @xmath104 for all even @xmath41 .
thus , we see that @xmath105 which concludes the proof .
pretty good state transfer between end vertices of paths was classified in @xcite . in this paper , we have given an infinite family of paths where pretty good state transfer occurs between internal vertices , but not between the end vertices .
a full classification of the orders @xmath6 where this occurs would be interesting and would complete the classification of pretty good state transfer in paths . in @xcite ,
the authors study pretty good state transfer between the end vertices in a heisenberg chain ; they study the hamiltonian @xmath106 , where @xmath107 denotes the laplacian adjacency matrix . in this case , they show that pretty good state transfer occurs between the end vertices of @xmath45 if and only if @xmath6 is a prime congruent to @xmath10 modulo @xmath108 or @xmath6 is a power of @xmath19 .
it is an open problem to find examples in which pretty good state transfer occurs between internal vertices while it does not occur between the end vertices .
gc acknowledges the support of grants fapesp 15/16339 - 2 and fapesp 13/03447 - 6 .
cv acknowledges the support of an nserc canada graduate scholarship .
the three authors thank chris godsil for generously inviting gabriel coutinho to visit waterloo , as well as for his general advice , support and for promoting a fruitful and positive research environment . | in this paper , we show that , for any odd prime @xmath0 and positive integer @xmath1 , the path on @xmath2 vertices admits pretty good state transfer between vertices @xmath3 and @xmath4 for each @xmath3 that is a multiple of @xmath5 with respect to the quantum walk model determined by the xy - hamiltonian .
this gives the first examples of pretty good state transfer occurring between internal vertices on a path , when it does not occur on the extremal vertices .
keywords : quantum walks , state transfer , graph eigenvalues mathematical subject classification : 81p68 , 15a16 , 05c50 |
measurements of type ia supernovae have shown that the expansion of the universe is accelerating @xcite , suggesting that most of the energy density of the universe consists of some form of dark energy with negative pressure @xcite . combining measurements of the cosmic microwave background anisotropy and observations of large - scale structure provides important corroborating evidence.@xcite two candidates for the dark energy are a cosmological constant ( or vacuum density ) and quintessence,@xcite a time - varying , spatially inhomogeneous component . in a previous paper @xcite ( paper i ) , we addressed the question of whether supernova measurements can be used to measure the equation of state ( eos ) of the negative pressure component , the ratio @xmath0 of the pressure to the energy density .
the issue is important because @xmath6 for a cosmological constant whereas @xmath0 takes on different values and can be significantly time - varying in the case of quintessence.@xcite under the assumption that @xmath0 is constant , its value can be determined to better than 5 per cent by measuring several thousand supernovae distributed equally between red shift @xmath7 and @xmath8 .
however , we showed that a degeneracy opens up if @xmath0 is time - dependent which makes it impossible to determine accurately the current value of @xmath0 or its time - derivative .
the cause of the degeneracy is that supernovae measure luminosity distance , which is related by a multi - integral expression to the eos as a function of red shift , @xmath9 .
widely different @xmath9 can have the same multi - integral value .
the purpose of this paper is to explore some pitfalls and challenges in determining the eos and its time variation using supernovae .
for example , we shall show how the standard practice of considering only models with @xmath10 or only models with constant @xmath0 when doing likelihood analyses can lead to grossly incorrect results .
for example , we will illustrate cases where the standard practice will suggest that @xmath0 is near -1 or much more negative than @xmath1 when , in fact , @xmath0 is significantly greater than -1 and rapidly time varying .
we shall show that a non - zero value of @xmath11 is more easily detected if @xmath12 than if @xmath13 .
we shall also contrast measuring @xmath0 for the negative pressure component alone ( @xmath14 ) versus the mean value for the total energy density ( including ordinary and dark matter ) , @xmath15 .
the degeneracy problem is less severe for @xmath15 , but this parameter provides less useful information .
we consider possibilities of breaking the degeneracy between @xmath0 and @xmath16 by combining supernovae results with either cosmic microwave background anisotropy measurements and/or the alcock - paczynski test .
we shall show that neither additional test significantly improves the measurement of the time - variation of @xmath0 , although optimistic assumptions about the alcock - paczynski test suggest that the current value of @xmath0 can be measured to within 20 percent or so .
we conclude that a new , yet to be found test has to be devised to resolve well the cosmic eos and its time variation . we stress that with current data it is possible to determine the eos to about a factor of two . for a future experiment to significantly enhance the determination of the eos , and enable the distinction between a constant eos and a time - dependent one
, it needs to resolve the equation of state at the 10% level or better .
the results of our analysis agree with many other analyses @xcite , although not always with their interpretation , and can be used to explain why some other analyses seem to indicate a superior resolving power of sn measurements alone @xcite , or in combination with other measurements @xcite .
some of the latter analyses implicitly assume unrealistic accuracy in independent determination of cosmological parameters , by not including self - consistently the uncertainty in @xmath9 in all measurements .
for example , assume a reported resolution of matter energy density which was based on assuming @xmath17 .
type ia supernovae have intrinsic variability of about 0.15 in absolute magnitude , but currently the errors in measuring distant sn are above this .
there are ongoing programs to extend the search to deeper redshifts and improve measurement quality .
the proposed snap satellite plans to measure 2000 sn per year , mostly in the range of redshifts @xmath18 , and some as far as @xmath19.@xcite the anticipated error in individual magnitudes is @xmath20 statistical and 0.02 systematic error , which yields about 1% relative error in luminosity distance @xmath21 . for our numerical estimates
, we have generated 50 sn magnitudes randomly chosen from a uniform distribution in @xmath22 values , between @xmath23 and @xmath8 .
magnitudes were generated from a gaussian distribution with mean value @xmath24 calculated using fiducial models .
we have used our 50 points to simulate approximately 2000 sn by reducing their magnitude error by a factor @xmath25 to 0.03 from the minimal 0.15 magnitude .
thus each generated point corresponds to 40 snap - like points , binned together .
this corresponds to @xmath26 relative error in @xmath21 .
hence , our analysis is based on a sn search more extensive than the actual snap proposal . to obtain a quantitative estimate of how well models are resolved
, we use one of two procedures .
first , we can find the maximum likelihood contours of the various models for each of the fiducial models and explore the degeneracy in parameter space .
alternatively , we can assume that all models which predict @xmath27 within 1% of the fiducial cosmological model for all @xmath22 between 0 and 2 are deemed indistinguishable .
we find that both approaches give comparable results .
that is , the @xmath28 cl likelihood contours using the first procedure are roughly equivalent to the indistinguishability region of the second .
luminosity distance is defined to be the ratio of luminosity @xmath29 to flux @xmath30 @xmath31 where the present value of the scale factor @xmath32 is normalized to unity ( throughout subscript 0 denotes present values ) ; @xmath33 is the coordinate distance @xmath34 and @xmath35 is the hubble parameter @xmath36 .
the observed sn magnitudes are related to @xmath21 , @xmath37 , \ ] ] @xmath38 being the sn absolute magnitude . for a flat universe with two energy sources , matter ( including dark matter ) and
a dark q - component , there are two equivalent routes to computing @xmath21 without assuming any prior about the time dependence of @xmath14 . one way is to use the algebraic relation between the total energy density and the hubble parameter @xmath35 . using conservation equations , the energy densities of the dark component @xmath39 and that of ordinary matter @xmath40 , are given by @xmath41 \\
\rho_m(z ) & = & ( \rho_m)_0(1+z)^3.\end{aligned}\ ] ] since @xmath42\end{aligned}\ ] ] where @xmath43 denotes the present ratio of matter to dark energy densities @xmath44 , @xmath35 can be expressed as @xmath45 \right)^{1/2}. \label{ein}\ ] ] substituting this into @xmath21 gives @xmath46^{-1/2 } \label{dl1}\end{aligned}\ ] ] an equivalent approach is to treat the sum of dark matter and the dark energy component as a single cosmic fluid with average equation of state @xmath47 , where @xmath48}\end{aligned}\ ] ] since @xmath49 is proportional to the total energy density in the universe , it can be expressed in terms of @xmath15 as follows @xmath50.\ ] ] here we have used the conservation equation for the total energy density . using ( [ seven ] ) we can express @xmath21 in terms of @xmath14 as follows , @xmath51= \nonumber \\ & = & \frac{(1+z)}{h_0}\int_1^{1+z}\frac{dx}{x^{3/2 } } exp\left[-\frac{3}{2}\int\limits_1^{x}\left(\frac{w_q(y)}{1+g\ exp\left[-3\int_1^{y}w_q(u)\frac{du}{u } \right ] } \right)\frac{dy}{y}\right].\end{aligned}\ ] ] this expression for @xmath21 as a function of @xmath14 has one more integral than the relation in eq .
( [ dl1 ] ) , but it is pedagogically useful in demonstrating that @xmath21 is sensitive only to a weighted average of @xmath15 or @xmath14 and not to their detailed time dependence .
based on the previous sections , a number of lessons can be learned about measuring the eos .
first , the relation between @xmath15 and @xmath21 in ( [ eight ] ) involves an integral , so we do expect some degeneracy in the determination of @xmath52 from sn measurements . to determine the eos of the dark energy itself , @xmath14 ,
the total energy density must be resolved into a matter component and a dark energy component .
hence , if @xmath53 and @xmath54 are not known from independent measurements , determining @xmath14 entails an additional uncertainty . for example , consider a flat universe with @xmath55 .
since the matter eos is @xmath56 , it follows that @xmath57 .
two models with different values of @xmath14 may produce the same value of @xmath15 and , consequently , @xmath21 , due to offsetting differences in the value of @xmath53 .
the more negative @xmath15 is , the faster is the expansion .
therefore , a more negative ( positive ) @xmath15 will make @xmath21 larger ( smaller ) .
in addition , light from earlier times ( emitted at higher @xmath22 values ) must pass through the low @xmath22 universe to reach us .
this means that changes in @xmath15 at lower @xmath22 affect @xmath21 at higher @xmath22 .
the converse is not true .
changes in @xmath15 at high @xmath22 do not affect @xmath21 at lower @xmath22 .
= 3.3 in consequently , it is not surprising that sn measurements of @xmath21 provide stronger constraints on @xmath52 at low @xmath22 than at high @xmath22 . in particular , if all cosmic parameters other than @xmath58 are fixed , there is a particular , relatively low value of @xmath59 for which @xmath52 is most tightly constrained .
this value of @xmath60 is clearly seen in our numerical results and was noted independently by @xcite .
for example , figure [ qln8tot ] shows the eos @xmath52 for three models each of which is obtained by best fit to @xmath27 for a fiducial model @xmath61 using one of three fitting assumptions : ( 1 ) that the dark eos is constant ; ( 2 ) that the dark eos is linear @xmath62 ; and , ( 3 ) that the total eos is quadratic @xmath63 .
while the real degeneracy is stronger than what is seen in the figure , we have chosen three examples to illustrate the existence of @xmath64 and @xmath60 .
as can be seen , the fits disagree significantly for @xmath22 far from @xmath65 , but all fits agree near @xmath60 .
= 3.3 in unfortunately , the resolution of @xmath66 , the quantity which most interests us , is degraded when we do not fix @xmath53 but , instead , allow for the current uncertainty in its value . in fig . [ evolution ] , we show some linear fits to simulated data generated from the fiducial model @xmath67 .
the fits are representative examples which fit the fiducial model to within the 95% confidence region .
the upper plot shows that @xmath52 ( with @xmath53 fixed at 0.3 ) is relatively well resolved , and particularly well resolved at around redshift @xmath68 .
the resolution is not that sharp in the middle plot which shows the corresponding @xmath14 , but a special point of enhanced resolution around @xmath69 is still clearly seen . if one lets @xmath53 vary in the realistic range of @xmath70 , then @xmath14 becomes poorly resolved and the spread at @xmath60 increases significantly .
similarly the spread in @xmath71 and @xmath60 increases significantly if more general functional forms of the eos are considered .
we would like to stress that the constant or linear forms of @xmath14 that we use are not meant to be anything more than simple concrete examples to highlight the fact that we are dealing with a degenerate parameter space . showing that if one assumes a linear @xmath58 , then it can be resolved to , say
, 50% does not logically mean that it can be measured to 50% accuracy generally since @xmath58 is resolved with different accuracy depending on its functional form .
this can be illustrated with the following examples . in fig .
[ sciam ] , the difference in magnitude ( @xmath72 ) for models with various eos is shown .
there are three clusters of points , each of which corresponds to a simulation of sn data for pair of different models .
each pair consists of a constant and linear @xmath14 .
each pair can be clearly separated from other pairs but the constant and linear members " of a pair can not be distinguished by sn data .
the examples chosen for fig . [ sciam ] have unrealistic large derivatives ( of order unity ) and therefore start to diverge from their constant partners for large @xmath22 .
more realistic examples with smaller derivatives or oscillatory behaviour will be much harder to distinguish from a constant eos . clearly
the treatment of the sn analysis is important .
if it is assumed that @xmath14 is constant , the figure shows that different values can be resolved to high accuracy , but if the assumption of constancy is relaxed and a linear @xmath22 dependence is allowed it becomes clear that the data can determine well only a single relation between @xmath73 and @xmath74 and that @xmath58 is poorly resolved .
= 3.3 in the degree of degeneracy exhibited in the @xmath75 fits depends on whether @xmath14 is positive or negative .
recall that if different models yield a total eos @xmath76 that is approximately equal , they are degenerate , and therefore changes in @xmath14 can be compensated by changes in @xmath54 ( or equivalently , in @xmath53 ) .
the difference between the case where @xmath14 is positive is due to the specific way in which this compensation mechanism operates .
if @xmath14 is positive , the curvature of degeneracy lines in @xmath77 plane is positive , as shown in the right panel of fig .
[ qc5qc ] for a fiducial model with @xmath78 .
conversely , if @xmath14 is negative , the curvature of the degeneracy line in @xmath77 plane is negative , as demonstrated in the left panel of figure [ qc5qc ] .
we have found that this result is unaffected by the value of the derivative of @xmath14 , even if it is quiet large .
the previous section ( and paper i ) show that the determination of @xmath58 from sn data is a more delicate process than it would seem .
if we know _ a priori _ that @xmath14 is constant , then its value can be determined quite accurately .
however , without this assumption , @xmath14 is poorly determined , and matters much get much worse if @xmath53 is uncertain .
the analysis can be further confounded if certain common practices are followed .
for example , many analyses assume that @xmath14 is constant and presume that , even if @xmath14 is time - varying , the constant- @xmath14 fit will provide the mean value over recent epochs . another common practice is to impose the condition that @xmath58 be limited to @xmath79 , based on the positivity and stability conditions that apply to most ( but not all ) forms of dark energy .
we shall see that both practices can produce enormous distortions of the likelihood surface that lead to grossly incorrect conclusions .
for example , we have tried to fit data generated from a fiducial model with @xmath80 , and @xmath81 over a redshift range @xmath82 .
note that the fiducial model has @xmath83 for all @xmath22 . yet , if we do a best - fit assuming the prior that @xmath14 is constant , we find it to be @xmath84 .
not only does the best - fit have @xmath85 , but the whole 95% confidence contour lies in a region where @xmath85 .
the results are the elongated contours in the lower part of fig .
[ cutqlp8 ] .
the reason for such a strange result can be understood from the functional dependence of @xmath21 and @xmath14 . assuming that @xmath86 , the energy density of the dark component is given in eq.([four ] ) , @xmath87 , \label{rhoq}\ ] ] so increasing @xmath74 and decreasing @xmath73 have opposite and compensating effects , which tend to cancel each other s influence on @xmath21 .
the exponential factor determines the quantitative details of this compensation since changes in @xmath73 need to compensate also for changes in the exponential factor .
it is therefore clear that the compensation can not be perfect over a range of @xmath22 s .
if we insist on the prior of constant eos ( _ i.e. _ , @xmath88 ) , the fitting procedure will pick out a value of @xmath73 which is much more negative than the fiducial value . in the figure we have picked a fiducial with a large positive derivative to illustrate our point , but it is clear from our discussion that the same problem arises when time dependence is weaker , or in cases that the eos has a more general functional form . introducing a prior that @xmath83 can give a very misleading impression of how well @xmath73 is resolved .
for example , suppose that we assume the priors that @xmath14 is constant and @xmath83 , as is standard practice .
the results are shown in fig .
[ cutqlp8 ] , the small contours truncated at @xmath17 .
they seem to suggest that the data supports the conclusion that @xmath17 with a high level of confidence .
the best fit is @xmath89 . yet , this is not related in any obvious way to the fiducial model , @xmath90 and @xmath91 .
the values of @xmath92 per degree of freedom for the best fit models of fig .
[ cutqlp8 ] are reasonable , .95 for the unconstrained fit , and 1.39 for the constrained fit .
so , what appears to be a compelling result is actually a total distortion .
of course , it is also conceivable that the actual @xmath58 is less than @xmath1 , in which case the same procedure of introducing a prior would falsely suggest that @xmath6 fits well .
an eos in which @xmath9 has a large positive time derivative is easier to detect than one which has a large negative time derivative . in either case
, the derivative must be large to be detected , as pointed out in paper i , but here we are demonstrating that the challenge is asymmetric .
the point is illustrated in fig .
the middle panel shows a fiducial model with a modest value of @xmath93 , and , as can be seen , this case can not be distinguished from a model in which @xmath88 .
that is , the 95% cl contours overlap the line @xmath88 , corresponding to no time variation .
the left and right panels show cases in which @xmath94 and @xmath95 , respectively .
the contours for the @xmath94 case ( left ) lie far from the @xmath88 line , so the time variation is detectable in this example . on the other hand , the contours for the @xmath96 case ( right ) overlay the @xmath88 ,
so the time - variation is not resolved .
this effect can be explained by considering the variation of the total average eos @xmath15 with respect to @xmath74 , @xmath97 : @xmath98 .
\label{erreq}\end{aligned}\ ] ] we consider @xmath15 because the measurements of @xmath21 are directly sensitive to @xmath15 , so that models can only be distinguished if they have different @xmath15 . as can be seen from eq.([erreq ] ) , @xmath15 is much less sensitive to changes in @xmath74 when it is negative than when it is positive , mainly due to the value of @xmath54 being larger for positive values of @xmath74 , we conclude that , in order to detect that @xmath14 is time - dependent , it must be that the time - variation is large , roughly @xmath99 , and it helps if @xmath74 is positive .
this corresponds to the case where acceleration is becoming stronger as time evolves .
measurements to determine the eos of the dark energy can be direct or indirect . direct methods , such as snia observations , the alcock - paczynski ( ap ) test @xcite , and the cosmic microwave background ( cmb ) attempt to measure the hubble parameter @xmath35 , its derivative @xmath100 and @xmath53 , or some function of them .
indirect methods , such as structure formation aspects of the cmb and measurements of large scale structure ( lss ) try to infer @xmath58 from its effects on structure evolution .
= 4.0 in an example of a complementary observation is the the alcock - pazcynski ( ap ) test .
the physical transverse size of an object is given by @xmath101 , @xmath102 being the angular distance and @xmath103 the observed angular size .
the physical radial size is @xmath104 . for a population of spherical objects ,
the ap test is given by equating the transverse and radial sizes : @xmath105 the ap test on its own is not expected to improve the resolution of the dark eos since it has a more complex dependence on @xmath14 than @xmath21 .
what does seem promising , as pointed out by mcdonald @xcite , is that the ap test can further constrain the range of @xmath53 .
[ contour ] shows the likelihood contours assuming optimistic anticipated errors over a continuous range between @xmath7 and @xmath8 of 1.4% for @xmath21 , and , for the ap test , 50 bins between redshift @xmath106 and @xmath107 measured with 3% error per bin.@xcite both simulations represent highly optimistic assumptions about future measurements .
the results are interesting .
the better constraint on @xmath53 from the ap test reduces the uncertainty in @xmath73 , but does not significantly change the uncertainty in the time - variation , @xmath74 .
this is not surprising since even a perfect determination of @xmath53 would leave a considerable uncertainty in @xmath74 , as shown in paper i. to be sure , the alcock - paczynski test is useful and worth pursuing , and a highly precise measurement combined with a highly precise measurement of sne could determine the present value of @xmath0 to within 15 or 20 percent .
however , it does not help significantly with the particular problem of pinning down the time - variation of the equation - of - state .
measurements of the cmb anisotropy provide an additional probe of @xmath9 .
this probe also suffers from a degeneracy problem , even in the case where @xmath0 is constant .
the positions of the acoustic peaks in the temperature anisotropy power spectrum depend on the angular distance ( @xmath102 ) to the last scattering surface which , just like the luminosity distance for supernovae , depends on a multi - integral over @xmath9 , @xmath108 .
in addition , the heights of the peaks depend on @xmath54 and the hubble parameter , @xmath109 . when all effects are considered , then , as shown by huey , et al .
@xcite , the power spectrum is unchanged as certain combinations of @xmath53 , @xmath110 , and @xmath0 are varied . consequently , none of these parameters can be determined well by the cmb data alone . instead
, measurements can only constrain these parameters to a thin two - dimensional surface in this three - dimensional parameter subspace .
the reason why one might be optimistic about combining cmb anisotropy and sn measurements is that the degeneracy surface for the cmb anisotropy measurements is nearly orthogonal to the degeneracy surface for the sn measurements for the case of constant @xmath0 .
figure 8 illustrates the small overlap between the sn and cmb degeneracy regions in the @xmath53-@xmath0 plane .
other authors have considered adding the cmb contribution@xcite but they have not included the degeneracy aspect . as we shall show below , introducing time - varying @xmath9 introduces additional degeneracy that spoils the resolution even when the sn and cmb anisotropy measurements are combined .
= 4 in rather than do another complete survey , which is a major technical challenge on its own , we illustrate the degeneracy in parameter space with a simple example in which we consider the family of @xmath9 of the form : @xmath111 this form was chosen to allow significant time variation recently when @xmath112 is large and , in particular , to be similar to the models considered in paper i for @xmath113 . for @xmath113 , the degeneracy problem with respect to sn data
was already demonstrated and the @xmath73-@xmath74 degeneracy region was characterized .
however , we could not simply maintain the linear change in @xmath9 with respect to @xmath22 out to the last scattering surface at @xmath114 because the value of @xmath0 would be ridiculously non - physical .
hence , we cutoff the @xmath22-dependence at a value of @xmath22 where @xmath54 is negligible and @xmath9 is physically plausible .
we , then , maintain that condition back to the last scattering surface .
for example , for @xmath115 and @xmath116 , at @xmath8 the dark energy contribution to the total energy density is less than 15% , which makes the details of the @xmath22-dependence cutoff unimportant . from @xmath8 until last scattering surface
, this model will have @xmath6 .
the value of @xmath73 in our time - varying examples is fixed to be @xmath117 except where otherwise stated . in each of these models , we also have @xmath118 , @xmath119 , @xmath120 , and @xmath121 . here
@xmath122 is the baryon density and @xmath53 is the total matter density ( baryonic plus non - baryonic ) .
note that luminosity distance - red shift measurements are not sensitive to @xmath123 , but the cmb measurements are .
the time - varying models were treated as the fiducial model , and then a numerical search was performed for a constant eos model that is indistinguishable from the fiducial model based on the combined measurement of the cmb and of supernovae .
models were considered degenerate under the combined tests if : ( 1 ) the percent difference between the luminosity distance - redshift predictions for the two models is less than one percent out to @xmath8 ( the same criterion as in paper i ) ; and , ( 2 ) the cmb predictions for the two models assuming a full - sky cosmic - variance limited measurement ( no experimental error ) can not be distinguished to better than 3@xmath124 .
both criteria are based on optimistic predictions of what will be realistically possible .
= 5.0 in for the cmb , distinguishability between a model with a constant @xmath0 and a fiducial with a time - dependent @xmath0 was determined by a log - likelihood analysis .
the log - likelihood was calculated according to the log - likelihood formula obtained by huey , _
et al._@xcite ; @xmath125 the coefficients @xmath126 and @xmath127 are the cmb multiple moments corresponding to the fiducial and constant equation of state models , respectively .
8 illustrates the problem that arises if one assumes @xmath9 is constant in the fitting procedure .
we have already observed that this distorts results for the case of sn data alone .
here we show that the problem remains if cmb data is co - added . assuming @xmath0 is constant ( @xmath128 ) , both measurements produce a thin degeneracy region in the @xmath53-@xmath0 plane . based on the small overlap , one
is tempted to conclude that constancy of @xmath0 is well established and its value is well determined .
however , this conclusion is absolutely wrong . in this example
, the fiducial model actually has a rapidly time - varying eos @xmath129 for @xmath113 ( which produces a change in @xmath0 of 50% over this range ) , and @xmath130 for @xmath131 .
the degeneracy regions were computed assuming @xmath88 .
if @xmath74 were to be fixed at a different value , once again the two measurements will give two narrow degeneracy regions with a small overlap , but the value of @xmath73 in this overlap region is significantly shifted .
for example , as shown in fig 8 , fixing @xmath88 results in @xmath132 with a few percent error .
however , if @xmath74 were to be fixed at its correct value @xmath133 , the result would have been @xmath134 , again with a few percent error .
but the central values differ by more than 10% .
it is clear , then , that the two measurements produce two degeneracy surfaces which intersect along a degenerate curve which passes through a range of models with varying values for @xmath73 and @xmath74 , that remain degenerate under the combined observations . for more complicated functional forms of @xmath58
the degeneracy curve becomes a more complicated higher dimensional surface , and the range of degeneracy in parameter space ( say , for @xmath73 ) increases . the complications in the process of extracting the eos from both measurements are further illustrated in fig .
there we show two time - varying models with slopes @xmath135 , one of which is degenerate ( by the log - likelihood test ) with a constant @xmath0 model with @xmath136 and @xmath121 , @xmath137 , @xmath138 , and @xmath139 .
the other can be barely resolved making the most optimistic estimates about cosmic variance . a slight decrease in @xmath74 , or
a slight decrease in experimental sensitivity would render the second model degenerate .
the lower two plots magnify the differences between the predictions of the models .
for the case of the cmb , we have also shown the envelope based on the constant @xmath0 model corresponding to the full - sky cosmic variance limit .
for the sn , we have constrained the limits to lie between @xmath140% .
if @xmath141 is larger than 1/6 for our particular form of @xmath9 , we find that there is no overlap between the degeneracy curve picked out by cmb measurements and the degeneracy contour picked out by sn measurements ( where both fits assuming @xmath0 is constant ) .
an example is shown in fig .
10 . in the case of negative ( positive ) @xmath74
the cmb measurements that fit best suggest low ( high ) @xmath53 whereas the sn measurements suggest high ( low ) @xmath53 .
if this absence of overlap were to be found in the real data , an interpretation to pursue is that @xmath0 is rapidly time - varying .
yet such an extreme scenario is not favoured by most theoretical models , most of which predict a moderately time - varying @xmath0 . for the more likely case , in which the two measurements do overlap ,
combining them reduces degeneracy by only a modest amount , generally not even enough to decide whether the dark energy in the universe has a time varying equation of state or not .
co - adding the cmb to the sn data represents an improvement in the sense that @xmath74 and @xmath73 are more constrained than with sn data alone based on our earlier analysis or in paper i. the improvement is by a factor of four or so _ assuming a linear form for @xmath9 _ , which is significant .
however , there remains a large uncertainty in the eos .
furthermore , we would stress once again that the accuracy in determining @xmath0 strongly depends on its assumed functional form .
the range of degeneracy obtained for @xmath74 ( a bit more than @xmath142 ) in our example underestimates the degeneracy for general @xmath9 .
for example , for parabolic forms , the uncertainty in @xmath74 blows up to @xmath143 .
given the extraordinarily precise data that has been brought to bear , the allowed variation in @xmath73 , @xmath74 and @xmath53 is disappointing .
an important challenge for observational cosmology is to measure the equation of state of the dark energy , @xmath58 .
this can provide important information about the fundamental physics that is responsible for the accelerated expansion of the universe .
measurements of the distance - red shift relation using supernovae , perhaps combined with other direct methods such as the alcock - paczyinski test or the cosmic microwave background , would appear to be promising methods .
indeed , analyses based on the _ a priori _ assumption that @xmath58 is constant suggest that @xmath14 can be resolved to 5% accuracy or better . in this paper and paper i ,
though , we have uncovered a number of problems and pitfalls that arise when trying to determine @xmath58 without making prior assumptions .
our lessons may be summarized as follows : * because measures of luminosity or angular distance depend on integrals over @xmath58 , a first degeneracy problem arises in which neither the current value and nor its time - variation can be resolved to any useful accuracy .
ii ) * since the effect of dark energy on the luminosity distance depends on the combination @xmath144 rather than @xmath14 itself , a second degeneracy problem arises in which @xmath14 and @xmath54 are changed simultaneously so as to keep @xmath144 fixed .
iii ) * although sn measurements may extend to @xmath8 , they are most sensitive to the behavior of @xmath58 at a modest value of @xmath145 .
iii ) * consequently , if there were only the first degeneracy problem , @xmath58 could be well - resolved at @xmath59 even though it is not well - resolved for other values of @xmath22 .
unfortunately , the resolution of @xmath66 is totally degraded when one includes uncertainty in @xmath54 and the second degeneracy problem .
iii , especially fig .
2 ) * the common practice of fitting data assuming that @xmath58 is constant can lead to grossly distorted results . similarly , the common practice of assuming @xmath146 can lead to grossly distorted results .
5 shows a dramatic example in which these practices lead to the conclusion that @xmath17 and is well - resolved when , in reality , @xmath83 and rapidly increasing .
iv ) * time - variation of @xmath14 is more easily detected if @xmath58 is an increasing function of @xmath22 rather than decreasing .
v ) * to resolve @xmath58 with supernova data , an additional test is needed .
given optimistic estimates of experimental uncertainties , the alcock - paczynski test combined with the supernovae measurements can constraint the current value of @xmath0 to within 20 percent or so .
however , neither the alcock - paczynski test nor microwave background anisotropy measurements provide the needed resolution to constrain the time - variation .
vi ) our principal conclusion is that a new test is required to achieve the goal of measuring @xmath58 . in devising a new test
, the two considerations must be precision and model dependence .
thus far , among the measurements that we have considered , the measurements which are precise give constraints on @xmath14 that are highly model dependent , leading to degeneracy problems .
tests which are not model dependent turn out to be difficult to measure precisely .
so , there lies the challenge . in considering alternatives
, it is critical to include practical estimates of their uncertainties .
furthermore , one must consider how the new tests , themselves , depend on @xmath58 .
for example , claims have been made that @xmath53 and @xmath54 have been or will be measured very accurately by measurements of the cosmic microwave background @xcite .
however , those estimates are based on assuming that @xmath17 . making no prior assumption about @xmath58 , a degeneracy problem once again arises@xcite that spoils the resolution of @xmath54 and @xmath58 , as discussed in sec .
vi . while trying to devise a new test to determine @xmath14 , it is worth mentioning that a precise measurement of @xmath100 will be extremely useful .
the dependence of @xmath14 on @xmath35 , @xmath100 ( prime denotes a derivative with respect to @xmath147 ) and @xmath53 is given by @xmath148 a good measurement of @xmath100 is clearly crucial to the resolution of @xmath14 , but current tests do not probe @xmath100 directly . the next best option is to measure @xmath149 , and then estimate @xmath100 by calculating its derivative .
obviously , this worsens the resolution for @xmath100 and increases the uncertainty in @xmath14 .
three additional approaches that we have not tried yet are measuring the time dependence of structure growth on @xmath9;@xcite gravitational lensing ; and direct measurements of @xmath150 ( to be discussed elsewhere@xcite ) .
we thank a. albrecht , g. efstathiou , d. eichler , p. mcdonald , and b. paczynski for helpful comments , and d. oaknin for valuable programming assistance .
this research was supported by grant no .
1999071 from the united states - israel binational science foundation ( bsf ) ( i m and rb ) and by the us department of energy grant de - fg02 - 91er40671 ( jm and pjs ) .
n. weiss , _ phys .
b _ * 197 * , 42 ( 1987 ) ; b. ratra and j.p.e .
peebles , _ ap . j. _ , 325 , l17 ( 1988 ) ; c.wetterich , _ nucl .
b_*302 * , 668 ( 1988 ) , and _ astron .
astrophys . _ * 301 * , 32 ( 1995 ) ; j.a .
frieman , _ et al .
lett . _ * 75 * , 2 077 ( 1995 ) ; k. coble , s. dodelson , and j. frieman , _ phys .
d _ * 55 * , 1851 ( 1995 ) ; p.g .
ferreira and m. joyce , _ phys .
lett . _ * 79 * , 4740 ( 1997 ) ; _ phys .
d _ * 58 * , 023503 ( 1998 ) ; c. armendariz - picon , v. mukhanov , and p.j .
steinhardt , _ phys .
* 85 * , 4438 - 41 ( 2000 ) .
s. podariu , p. nugent and b. ratra , astrophys .
j. * 553 * , 39 ( 2001 ) .
p. astier , astro - ph/0008306 .
j. weller and a. albrecht , phys .
lett . * 86 * , 1939 ( 2001 ) .
t. chiba and t. nakamura , phys .
d * 62 * , 121301 ( 2000 ) . v. barger and d. marfatia , phys .
b * 498 * , 67 ( 2001 ) .
m. chevallier and d. polarski , int .
j. mod .
d * 10 * , 213 ( 2001 ) .
m. goliath , r. amanullah , p. astier , a. goobar and r. pain , astro - ph/0104009 .
n. trentham , astro - ph/0105404 .
e. h. gudmundsson and g. bjornsson , astro - ph/0105547 . j. weller and a. albrecht , astro - ph/0106079 . | we explore various pitfalls and challenges in determining the equation - of - state ( @xmath0 ) of dark energy component that dominates the universe and causes the current accelerated expansion .
we demonstrated in an earlier paper the existence of a degeneracy that makes it impossible to resolve well the value of @xmath0 or its time - derivative with supernovae data .
here we consider standard practices , such as assuming priors that @xmath0 is constant or greater than @xmath1 , and show that they also can lead to gross errors in estimating the true equation - of - state .
we further consider combining measurements of the cosmic microwave background anisotropy and the alcock - paczynski test with supernovae data and find that the improvement in resolving the time - derivative of @xmath0 is marginal , although the combination can constrain its present value perhaps to 20 percent uncertainty .
l epsf irit maor@xmath2 , ram brustein@xmath2 , jeff mcmahon@xmath3 , paul j. steinhardt@xmath3 + _ @xmath4department of physics , ben gurion university , beer - sheva 84105 , israel + @xmath5 department of physics , princeton university , princeton , nj 08540 usa _
pacs number(s ) : 98.62.py , 98.80.es , 98.80.-k |
let @xmath4 be a ( possibly infinite ) network rooted at node @xmath5 .
assume that independent and identically distributed noisy observations of an hidden random variable @xmath6 are available at a subset @xmath7 of the vertices .
explicitly , each @xmath8 has access to a private signal @xmath9 where @xmath10 are independent and identically distributed , conditional on @xmath2 .
the ` state of the world ' @xmath2 is drawn from a prior probability distribution @xmath11 .
the objective is to aggregate information about @xmath2 at the root node under communication constraints encoded by the network structure , while minimizing the error probability at @xmath5 .
we ask the following question : _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ how much does the error probability at the root node @xmath5 increase due to these communication constraints ? _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ in order to address this question , consider a sequence of information aggregation problems indexed by @xmath0 .
information is revealed in a subset of the vertices @xmath12 .
there are @xmath0 rounds in which information aggregation occurs . in each round , a subset of the nodes in @xmath13 make ` decisions ' that are broadcasted to their neighbors . in the initial round , nodes @xmath14 with distance @xmath15 ( with @xmath16 being the graph distance ) broadcast a decision @xmath17 to their neighbors , with @xmath3 a finite alphabet . in the next round ,
nodes @xmath18 with distance @xmath19 broadcast a decision @xmath17 to their neighbors . and
so on , until the neighbors of @xmath5 announce their decisions in round @xmath0 .
finally , the root makes its decision .
the decision of any node @xmath20 is a function of decisions of @xmath20 s neighbors in earlier rounds , and , if @xmath8 , on the private signal @xmath21 received by @xmath20 .
clearly , the root can possibly access only the private information available at nodes @xmath22 with @xmath23 ( with @xmath16 the graph distance ) .
we can therefore assume , without loss of generality , that @xmath24 .
it is convenient to think of @xmath25 as the _ information horizon _ at time @xmath0 .
consider first the case in which communication is unconstrained .
this can be modeled by considering the graph with vertices @xmath26 and edges @xmath27 .
in other words , this is a star network , with the root at the center . without loss of generality , we take @xmath28 , with @xmath29 as @xmath30 . a simple procedure for information aggregation would work as follows .
each node @xmath20 computes the log - likelihood ratio ( llr ) @xmath31 corresponding to the observed signal , and quantizes it to a value @xmath32 .
the root adds up the quantized llrs and decides on the basis of this sum .
it follows from basic large deviation theory @xcite that , under mild regularity assumptions , the error probability decreases exponentially in the number of observations @xmath33 this result is extremely robust : * @xmath34 * it holds for any non - trivial alphabet @xmath35 ; * @xmath36 * using concentration - of - measure arguments @xcite it is easy to generalize it to families of weakly dependent observations @xcite ; * @xmath37 * it can be generalized to network structures @xmath38 with weak communications constrains .
for instance , @xcite proved that the error probability decays exponentially in the number of observations for trees of bounded depth .
the crucial observation here is that such networks have large degree diverging with the number of vertices .
in particular , for a tree of depth @xmath0 , the maximum degree is at least @xmath39 . at the other extreme ,
hellmann and cover @xcite considered the case of a line network . in our notations , we have @xmath40 , @xmath41 , and @xmath42 .
in @xcite they proved that , as long as the llrs are bounded ( namely @xmath43 almost surely for some constant @xmath44 ) , and the decision rule is independent of the node , the error probability remains bounded away from @xmath45 as @xmath30 . if the decision rule is allowed to depend on the node , the error probability can vanish as @xmath30 provided @xmath46 @xcite . despite this , even if the probability of error decays to @xmath45 , it does so much more slowly than for highly connected networks .
namely , tay , tsitsiklis and win @xcite proved that @xmath47 for some @xmath48 .
in other words , the communication constraint is so severe that , after @xmath0 steps , the amount of information effectively used by the root is equivalent to a vanishingly small fraction of the one within the ` information horizon ' .
these limit cases naturally lead to the general question : given a rooted network @xmath49 , a sequence of information horizons @xmath50 and a finite alphabet @xmath3 , can information be aggregated at the root in such a way that the error probability decays exponentially in @xmath51
? the question is wide open , in particular for networks of with average degree bounded or increasing slowly ( e.g. logarithmically ) with the system size .
networks with moderate degree arise in a number of practical situations . within decentralized detection applications ,
moderate degree is a natural assumption for interference - limited wireless networks . in particular , systems in which a single root node communicates with a significant fraction of the sensors
are likely to scale poorly because of interference at the root .
standard models for wireless ad hoc networks @xcite are indeed based on random geometric graphs whereby each node is connected to a logarithmic number of neighbors .
a different domain of applications for models of decentralized decision making is social learning @xcite . in this case
, each node corresponds to an agent , and the underlying graph is the social network across which information is exchanged .
also in this case , it is reasonable to assume that each agent has a number of neighbors which is bounded , or diverges slowly as the total number of agents grows . in many graph - theoretic models of social networks
@xcite , although a small number of nodes can have large degree , the average degree is bounded or grows logarithmically with the network size .
given the slow progress with extreme network structures ( line networks and highly - connected networks ) , the study of general moderate degree networks appears extremely challenging . in this paper
we focus on regular trees .
more precisely , we let @xmath38 be the ( infinite ) regular tree with branching factor @xmath52 , rooted at @xmath5 ( each node has @xmath52 descendants and , with the exception of the root , one parent ) . the information horizon @xmath25 is formed by all the nodes at distance @xmath0 from the root , hence @xmath53 . under a broad set of assumptions , we prove that the probability of error decays subexponentially in the size of the information set , cf .
( [ eq : subexp ] ) , where @xmath54 depends on the size of the alphabet @xmath55 . more precisely , we establish subexponential convergence in the following cases : 1 . for binary messages @xmath56 and any choice of the decision rule .
in fact , we obtain a precise characterization of the smallest possible error probability in this case .
2 . for general message alphabet @xmath57
provided the decision rule does not depend on the node , and satisfies a mild ` irreducibility ' condition ( see section [ subsec : subexp_decay_general ] for a definition ) . in the latter case
, one expects that exponential convergence is recovered as the message set gets large .
indeed we prove that the optimal exponent in eq .
( [ eq : subexp ] ) obeys @xmath58 the upper bound follows from our general proof for irreducible decision rules , while the lower bound is obtained by constructing an explicit decision rule that achieves it .
our investigation leaves several interesting open problems .
first , it would be interesting to compute the optimal exponent @xmath59 for given degree of the tree and size of the alphabet .
even the behavior of the exponent for large alphabet sizes is unknown at the moment ( cf .
second , the question of characterizing the performance limits of general , node - dependent decision rules remains open for @xmath60 .
third , it would be interesting to understand the case where non - leaf nodes also get private signals , e.g. , @xmath61 .
finally , this paper focuses on tree of bounded degree .
it would be important to explore generalization to other graph structures , namely trees with slowly diverging degrees ( which could be natural models for the local structure of preferential attachment graphs @xcite ) , and loopy graphs .
our current results can be extended to trees of diverging degree only in the case of binary signals . in this case
we obtain that the probability of error is subexponential @xmath62 as soon as the degree is sub - polynomial , i.e. @xmath63 for all @xmath64 .
the rest of the paper is organized as follows : section [ sec : model ] defines formally the model for information aggregation .
section [ sec : binary ] presents our results for binary messages @xmath56 .
section [ sec : nodeoblivious ] treats the case of decision rules that do not depend on the node , with general @xmath3 .
as mentioned in the introduction , we assume the network @xmath65 to be an ( infinite ) rooted @xmath52-ary tree , i.e. a tree whereby each node has @xmath52 descendants and one parent ( with the exception of the root , that has no parent ) .
independent noisy observations ( ` private signals ' ) of the state of the world @xmath2 are provided to the nodes at all the nodes at @xmath0-th generation @xmath66 .
these will be also referred to as the ` leaves ' .
define @xmath67 .
formally , the state of the world @xmath68 is drawn according to the prior @xmath69 and for each @xmath70 an independent observation @xmath71 is drawn with probability distribution @xmath72 ( if @xmath73 ) or @xmath74 ( if @xmath75 ) . for notational simplicity
we assume that @xmath76 is finite , and that @xmath77 , @xmath78 for all @xmath79 . also , we exclude degenerate cases by taking @xmath80 .
we refer to the refer to the two events @xmath81 and @xmath82 as the hypotheses @xmath83 and @xmath84 .
in round 0 , each leaf @xmath20 sends a message @xmath17 to its parent at level 1 . in round 1
, the each node @xmath85 at level 1 sends a message @xmath86 to its parent at level 2 . similarly up to round @xmath0 .
finally , the root node @xmath5 makes a decision @xmath87 based on the @xmath52 messages it receives .
the objective is to minimize @xmath88 .
we call a set of decision rules _ optimal _ if it minimizes @xmath89 . we will denote by @xmath90 the set of children of node @xmath20 .
we denote the probability of events under @xmath83 by @xmath91 , and the probability of events under @xmath84 by @xmath92 .
finally , we denote by @xmath93 the decision rule at node @xmath20 in the tree . if @xmath20 is not a leaf node and @xmath94 , then @xmath95 .
the root makes a binary decision @xmath96 .
if @xmath20 is a leaf node , it maps its private signal to a message , @xmath97 . in general , @xmath93 s can be randomized .
in this section , we consider the case @xmath98 , i.e. , the case of binary messages .
consider the case @xmath99 , @xmath100 and @xmath101 for @xmath102 ; where @xmath103 .
define the majority decision rule at non - leaf node @xmath20 as follows : @xmath32 takes the value of the majority of @xmath104 ( ties are broken uniformly at random ) .
it is not hard to see that if we implement majority updates at all non - leaf nodes , we achieve @xmath105 note that this is an upper bound on error probability under majority updates .
our main result shows that , in fact , this is essentially the best that can be achieved .
[ thm : binary_lower_bound ] fix the private signal distribution , i.e. , fix @xmath106 and @xmath107 .
there exists @xmath108 such that for all @xmath109 and @xmath110 , for any combination of decision rules at the nodes , we have @xmath111 in particular , the error probability decays subexponentially in the number of private signals @xmath112 , even with the optimal protocol .
we prove the theorem for the case @xmath99 , @xmath100 and @xmath101 for @xmath102 ; where @xmath103 .
the proof easily generalizes to arbitrary @xmath113 and @xmath114 .
also , without loss of generality we can assume that , for every node @xmath20 , @xmath115 ( otherwise simply exchange the symbols and modify the decision rules accordingly ) .
denote by @xmath116 the ( negative ) logarithm of the ` type i error ' in @xmath32 , i.e. @xmath117 . denote by @xmath118 the ( negative ) logarithm of the ` type ii error ' in @xmath32 , i.e. @xmath119 . the following is the key lemma in our proof of theorem [ thm : binary_lower_bound ] .
[ lemma : binary_key_lemma ] given @xmath120 , there exists @xmath121 such that for any @xmath52 we have the following : there exists an optimal set of decision rules such that for any node @xmath20 at level @xmath122 , @xmath123 applying lemma [ lemma : binary_key_lemma ] to the root @xmath5 , we see that @xmath124 .
the result follows immediately .
lemma [ lemma : binary_key_lemma ] is proved using the fact that there is an optimal set of decision rules that correspond to deterministic likelihood ratio tests ( lrts ) at the non - leaf nodes .
choose a node @xmath20 .
fix the decision functions of all descendants of @xmath20 .
define @xmath125 . + a ) the decision function @xmath93 is a _ monotone deterministic likelihood ratio test _ if : + ( i ) it is deterministic . + ( ii ) there is a threshold @xmath126 such that @xmath127 \b ) the decision function @xmath93 is a
_ deterministic likelihood ratio test _ if either @xmath93 or @xmath128 is a monotone deterministic likelihood ratio test .
here @xmath128 is the boolean complement of @xmath93 .
the next lemma is an easy consequence of a beautiful result of tsitsiklis @xcite . though we state it here only for binary message alphabet , it easily generalizes to arbitrary finite @xmath3 .
[ lemma : lrtoptimal ] there is a set of monotone deterministic likelihood ratio tests at the nodes that achieve the minimum possible @xmath129 .
consider a set of decision rules that minimize @xmath129 .
fix the rule at every node except node @xmath20 to the optimal one .
now , the distributions @xmath130 and @xmath131 are fixed .
moreover , @xmath129 is a linear function of @xmath132 , where @xmath133 denotes the distribution of @xmath32 under hypothesis @xmath134 .
the set @xmath135 of achievable @xmath136 s is clearly convex , since randomized @xmath93 is allowed . from (
* proposition 3.1 ) , we also know that @xmath135 is compact
. thus , there exists an extreme point of @xmath135 that minimizes @xmath129 .
now ( * ? ? ?
* proposition 3.2 ) tells us that any extreme point of @xmath135 can be achieved by a deterministic lrt .
thus , we can change @xmath93 to a deterministic lrt without increasing @xmath129 .
if @xmath93 is not monotone ( we know that @xmath94 in this case ) , then we do @xmath137 and @xmath138 .
clearly , @xmath129 is unaffected by this transformation , and @xmath93 is now a monotone rule .
we do this at each of the nodes sequentially , starting at level @xmath45 , then covering level @xmath139 and so on until the root @xmath5 .
thus , we change ( if required ) each decision rule to a monotone deterministic lrt without increasing @xmath129 .
the result follows . clearly , if @xmath93 is a monotone lrt , eq
. holds .
in fact , we argue that there is a set of deterministic monotone lrts with strict inequality in eq . , i.e. , such that @xmath140 holds for all @xmath20 , that are optimal .
eq . can only be written when @xmath141 and @xmath142 .
consider a leaf node @xmath20 . without loss of generality we can take @xmath143 for each leaf node @xmath20 ( since any other rule can be ` simulated ' by the concerned level 1 node ) .
so we have @xmath141 and @xmath142 , eq
. holds and @xmath93 is a deterministic lrt .
we can ensure these properties inductively at all levels of the tree by moving from the leaves towards the root .
consider any node @xmath20 .
if @xmath144 , then @xmath94 ( else @xmath145 ) and the parent of @xmath20 is ignoring the constant message received from @xmath20 .
we can do at least as well by using any non - trivial monotone deterministic lrt at @xmath20 .
similarly , we can eliminate @xmath146 . if @xmath141 and @xmath142 , then eq . must hold for any monotone deterministic lrt @xmath93 , using the inductive hypothesis .
let @xmath147 and @xmath148 be binary vectors of the same length @xmath149 .
we say @xmath150 if @xmath151 for all @xmath152 .
we now prove lemma [ lemma : binary_key_lemma ] . from lemma [ lemma : lrtoptimal ] and eq .
, we can restrict attention to monotone deterministic lrts satisfying eq . .
we proceed via induction on level @xmath149 . for any leaf node @xmath20 ,
we know that @xmath153 . choosing @xmath154 , eq
. clearly holds for all nodes at level @xmath45 .
. holds for all nodes at level @xmath149 .
let @xmath20 be a node at level @xmath155 .
let its children be @xmath156 .
without loss of generality , assume @xmath157 * claim : * we can also assume @xmath158 proof of claim : suppose , instead , @xmath159 ( so @xmath160 is doing better than @xmath161 on both types of error ) . we can use the protocol on the subtree of @xmath160 also on the subtree of @xmath161 .
call the message of @xmath161 under this modified protocol @xmath162 . since , @xmath163 and @xmath164 ( both types of error
have only become less frequent ) , there exists a randomized function @xmath165 , such that @xmath166 for @xmath167 .
thus , node @xmath20 can use @xmath168 to achieve the original values of @xmath169 and @xmath170 , where @xmath93 is decision rule being used at @xmath20 before . clearly , the error probabilities at @xmath20 , and hence at the root , stay unchanged with this .
thus , we can safely assume @xmath171 .
similarly , we can assume @xmath172 for @xmath173 .
clearly , our transformations retained the property that nodes at levels @xmath155 and below use deterministic lrts satisfying eq . .
similar to our argument for eq .
above , we can make appropriate changes in the decision rules at levels above @xmath155 so that they also use deterministic lrts satisfying eq .
, without increasing error probability .
this proves the claim .
recall that @xmath174 is the decision rule at node @xmath20 .
assume the first bit in the input corresponds to @xmath175 , the second corresponds to @xmath176 , and so on . using lemma [ lemma : lrtoptimal ] , we can assume that @xmath93 implements a deterministic likelihood ratio test .
define the @xmath52-bit binary vectors @xmath177 , @xmath178 , , @xmath179 . from lemma [ lemma : lrtoptimal ] and
, it follows that @xmath180 for some @xmath181 .
* claim : * without loss of generality , we can assume that @xmath182 and @xmath183 .
proof of claim : suppose @xmath184 .
it follows from lemma [ lemma : lrtoptimal ] and eq . that @xmath185 for every possible @xmath104 . if @xmath186 then we have @xmath187 .
suppose @xmath188 .
then @xmath32 is a constant and is ignored by the parent of @xmath20 .
we can not do worse by using an arbitrary non - trivial decision rule at @xmath20 instead .
( the parent can always continue to ignore @xmath189 . )
the case @xmath190 can be similarly eliminated .
this proves the claim .
thus , we can assume @xmath191 without loss of generality .
now @xmath192 contribute to type i error and @xmath193 contribute to type ii error .
it follows that @xmath194 where we have used the ordering on the error exponents ( eqs . and ) .
eqs . and lead immediately to @xmath195 now , for any @xmath196 , we have @xmath197 . plugging @xmath198 and @xmath199
, we obtain from eq .
@xmath200 by our induction hypothesis @xmath201 .
thus , @xmath202 as required .
induction completes the proof .
in this section we allow a general finite message alphabet @xmath3 that need not be binary .
however , we restrict attention to the case of _ node - oblivious _ rules : the decision rules @xmath93 at all nodes in the tree , except the leafs and the root , must be the same .
we denote this ` internal node ' decision rule by @xmath203 .
also , the decision rules used at each of the leaf nodes should be same .
we denote the leaf decision rule by @xmath204 .
the decision rule at the root is denoted by @xmath205 .
we call such @xmath206 a node - oblivious decision rule vector . define @xmath207 . in section [ subsec : node - oblivious_efficient_scheme ]
, we present a scheme that achieves @xmath208 when the error probability in the private signals is sufficiently small . next , under appropriate assumptions , we show that the decay of error probability must be sub - exponential in the number of private signals @xmath209 . for convenience ,
we label the messages as @xmath210 the labels have been chosen so as to be suggestive ( in a quantitative sense , see below ) of the inferred log - likelihood ratio .
further , we allow the messages to be treated as real numbers ( corresponding to their respective labels ) that can be operated on . in particular , the quantity @xmath211 is well defined for a non - leaf node @xmath20 .
the node - oblivious decision rule we employ at a non - leaf node @xmath188 is @xmath212 \left \lfloor \frac{s_i / k - ( m-1)/2}{1 - 1/m}\right \rfloor + \frac{m-1}{2}\ , , & \textup{if } s_i > 0 \end{array } \right .
\label{eq : good_decision_rule}\end{aligned}\ ] ] note that the rule is symmetric with respect to a inversion of sign , except that @xmath213 is mapped to the message @xmath214 when @xmath215 is even .
the rule @xmath216 used at the leafs is simply @xmath217 and @xmath218 .
the decision rule at the root is @xmath219 if we associate @xmath83 with negative quantities , and @xmath84 with positive quantities , then again , the rule at the leafs is symmetric , and the rule at the root is essentially symmetric ( except for the case @xmath220 ) .
assume @xmath83 .
define @xmath229 and @xmath230 .
we show that , in fact , for suitable choice of @xmath231 the following holds : if @xmath224 , then for any node @xmath20 at any level @xmath225 , @xmath226 \geq \nonumber\\ ( l / m ) \gamma^\tau + c \label{eq : letter_prob_decay}\end{aligned}\ ] ] we proceed by induction on @xmath149 .
consider @xmath20 at level @xmath232 .
we have @xmath233 = 0 $ ] for @xmath234 and @xmath235 = \delta$ ] . choosing @xmath236 , we can ensure that eq
. holds at level @xmath45 .
note that for @xmath237 , we have @xmath238 .
now suppose eq
. holds at level @xmath149 .
consider node @xmath20 at level @xmath155 . from eq .
, for @xmath239 we need @xmath240 \label{eq : si_lb}\end{aligned}\ ] ] for every @xmath241 such that eq . holds , we have @xmath242 .
thus , @xmath243 obviously , there are at most @xmath244 such @xmath104 .
thus , @xmath245 \\ \leq \ ; & m^k \exp\left(- kc - ( 1/m)l \gamma^{\tau+1 } \right ) \\
= \ ; & \exp\left(- c - ( 1/m)l \gamma^{\tau+1 } \right ) \end{aligned}\ ] ] thus , eq . holds at level @xmath155 .
induction completes the proof . for @xmath221 and @xmath222
, there exists @xmath246 , and a node - oblivious decision rule vector , such that the following is true : for any @xmath224 , we have @xmath247 & \leq \exp \left \ { - \frac{m-1}{2 m } \big \{k \left ( 1 - 1/m\right ) \big \}^t \right \ } \nonumber\\ & = \exp \left \{- \frac{m-1}{2m}\ , n^\rho \right \}\end{aligned}\ ] ] with @xmath248 .
[ thm : good_decision_rule_exists ] assume @xmath83 .
for every @xmath249 such that @xmath250 , we have @xmath251 . from lemma [ lemma : good_decision_decay](i ) , @xmath252 where @xmath229 and @xmath230 .
obviously , there are at most @xmath244 such @xmath253 .
it follows that @xmath254 define @xmath258 , i.e. , @xmath259 is the number of private signals received , one at each leaf .
the scheme presented in the previous section allows us to achieve error probability that decays like @xmath260 , where @xmath261 for @xmath262 . in this section
we show that under appropriate assumptions , error probability that decays exponentially in @xmath259 , i.e. , @xmath263 , is not achievable with node - oblivious rules .
in this section we call the letters of the message alphabet @xmath264 . for simplicity , we consider only deterministic node - oblivious rules , though our results and proofs extend easily to randomized rules .
we define here a directed graph @xmath265 with vertex set @xmath3 and edge set @xmath266 that we define below .
we emphasize that @xmath265 is distinct from the tree on which information aggregation is occurring .
there is a directed edge from node @xmath267 to node @xmath268 in @xmath265 if there exists @xmath269 such that @xmath270 appears at least once in @xmath147 and @xmath271 .
informally , @xmath272 if @xmath273 can be ` caused ' by a message vector received from children that includes @xmath270 .
we call @xmath265 the _ dependence graph_. [ ass : strongly_connected ] the dependence graph @xmath265 is strongly connected . in other words , for any @xmath267 and @xmath268 such that @xmath274 , there is a directed path from @xmath273 to @xmath270 in @xmath265
. there exists @xmath282 , @xmath283 , @xmath284 and @xmath285 such that , for all @xmath286 the following holds : for node @xmath20 at level @xmath149 , we have @xmath287 and @xmath288 .
[ ass : one_letter_dominant ] it is not hard to verify that for @xmath221 , @xmath222 and @xmath289 ( where @xmath231 is same as in lemma [ lemma : good_decision_decay ] and theorem [ thm : good_decision_rule_exists ] ) , the scheme presented in the previous section satisfies all four of our assumptions .
in other words , the assumptions are all satisfied in the regime where our scheme has provably good performance .
consider a directed graph @xmath290 that is strongly connected .
for @xmath291 , let @xmath292 be the length of the shortest path from @xmath293 to @xmath294 .
then the _ diameter _ of @xmath265 is defined as -15pt @xmath295 [ thm : subexp_decay ] fix @xmath215 and @xmath52 .
consider any node - oblivious decision rule vector @xmath206 such that assumptions [ ass : strongly_connected ] , [ ass : all_letters_+ve_prob ] and [ ass : one_letter_dominant ] are satisfied .
let @xmath296 be the diameter of the dependence graph @xmath265 .
then , there exists @xmath297 such that we have @xmath298 \geq \exp \left \{-
c n^{\brho } \right \}\ , , \end{aligned}\ ] ] where @xmath299 .
[ coro : subexp_decay ] fix @xmath215 and @xmath52 .
consider any node - oblivious decision rule vector @xmath206 such that assumptions [ ass : strongly_connected ] , [ ass : all_letters_+ve_prob ] and [ ass : one_letter_dominant ] are satisfied .
then , there exists @xmath297 such that we have @xmath298 \geq \exp \left \{-
c n^{\rho } \right \}\ , , \end{aligned}\ ] ] where @xmath301 .
[ rem : mus_vec_maps_to_itself ] we have @xmath302 .
it follows that we must have @xmath303 ( else the probability of error is bounded below by @xmath304 for any @xmath0 ) .
similarly , we must have @xmath305 .
in particular , @xmath306 .
it follows from assumption [ ass : all_letters_+ve_prob ] that for any @xmath278 , there is some @xmath269 such that @xmath308 .
we prove the lemma by induction on the level @xmath149 .
let @xmath309 by assumption , @xmath310 holds .
suppose @xmath311 holds .
consider node @xmath20 at level @xmath155 .
consider any @xmath278 . by inductive hypothesis
, we have @xmath312 .
it follows that @xmath277 .
thus , @xmath313 holds .
lemma [ lemma : subexp_decay ] can be thought of as a quantitative version of lemma [ lemma : all_letters_+ve_prob_after_tau0 ] , showing that the probability of the least frequent message decays subexponentially .
suppose assumptions [ ass : strongly_connected ] , [ ass : all_letters_+ve_prob ] and [ ass : one_letter_dominant ] are satisfied .
fix @xmath314 . consider a node @xmath20 at level @xmath149 .
define @xmath315 .
let @xmath316 ( cf .
assumptions [ ass : all_letters_+ve_prob ] , [ ass : one_letter_dominant ] ) .
let @xmath317 .
there exists @xmath318 such that for any @xmath319 and @xmath320 , we have , @xmath321 [ lemma : subexp_decay ] we proceed via induction on @xmath323 .
first consider @xmath324 .
consider a node @xmath20 at level @xmath325 for @xmath320 .
consider the descendants of node @xmath20 at level @xmath326 .
for any @xmath278 , we know from lemma [ lemma : all_letters_+ve_prob_after_tau0 ] that there must be _ some _ assignment of messages to the descendants , such that @xmath327 .
it follows that @xmath328 thus , choosing @xmath329 , we can ensure that eq .
holds for @xmath324 and all @xmath320 .
now suppose eq
. holds for some @xmath330 .
consider a node @xmath20 at level @xmath331 .
let @xmath332 be the set of descendants of node @xmath20 at level @xmath333 .
note that @xmath334 .
consider any @xmath278 . by assumption [ ass : strongly_connected ] , there is a directed path in @xmath265 of length at most @xmath296 going from @xmath335 to @xmath336 . by remark [ rem : mus_vec_maps_to_itself ] , we know that @xmath337 .
it follows that there is a directed path in @xmath265 of length _ exactly _ @xmath296 going from @xmath335 to @xmath336 .
thus , there must be an assignment of messages @xmath338 to nodes in @xmath332 , including at least one occurrence of @xmath336 , such that @xmath327 .
using assumption [ ass : one_letter_dominant ] , we deduce that @xmath339 rewriting as @xmath340 and combining with eq .
, we obtain @xmath341 induction completes the proof . for the scheme presented in section [ subsec : node - oblivious_efficient_scheme ] , we have @xmath346 , where @xmath347 . for any @xmath348 , theorem [ thm : subexp_decay ]
provides a lower bound on error probability with @xmath349 for some @xmath350 .
this closely matches the @xmath215 dependence of the upper bound on error probability we proved in theorem [ thm : good_decision_rule_exists ] .
we already mentioned that the efficient node - oblivious rule presented in section [ subsec : node - oblivious_efficient_scheme ] satisfies all of assumptions [ ass : strongly_connected ] , [ ass : all_letters_+ve_prob ] and [ ass : one_letter_dominant ] .
moreover , it is natural to expect that similar schemes based on propagation of quantized likelihood ratio estimates should also satisfy our assumptions . in this section
, we further discuss our assumptions taking the cases of binary and ternary messages as examples .
binary messages are not the focus of section [ subsec : subexp_decay_general ] .
however , we present here a short discussion of assumptions 1 , 2 and 3 in the context of binary messages for illustrative purposes .
proof of claim : call the messages @xmath98 .
consider a node - oblivious decision rule vector @xmath206 such that error probability decays to @xmath45 with @xmath0 .
then @xmath352 can not be a constant function ( e.g. , identically @xmath45 ) , since this leads to @xmath353 .
suppose assumption [ ass : strongly_connected ] is violated .
without loss of generality , suppose @xmath354 .
then @xmath355 for all @xmath356 .
it follows that for node @xmath20 at level @xmath149 , we have @xmath357 for both @xmath73 and @xmath75 .
in particular , @xmath89 is bounded away from @xmath45 .
this is a contradiction .
suppose assumption [ ass : all_letters_+ve_prob ] is violated .
then , wlog , all nodes at level @xmath139 transmit the message @xmath139 almost surely , under either hypothesis .
thus , all useful information is lost and @xmath353 .
this is a contradiction .
finally , we show that assumption [ ass : one_letter_dominant ] must hold as well .
define @xmath358 for node @xmath20 at level @xmath0 .
wlog , suppose @xmath359 occurs infinitely often
. then we have @xmath360 , else @xmath361 for infinitely many @xmath0 .
define @xmath362 for node @xmath20 at level @xmath0 . if @xmath363 occurs infinitely often
, then it follows that @xmath364 and hence @xmath365 occur for infinitely many @xmath0 .
so we can have @xmath363 only finitely many times .
also , @xmath366 must hold .
it follows that @xmath367 occurs only finitely many times .
thus , assumption [ ass : one_letter_dominant ] holds with @xmath368 .
we first show that if assumption [ ass : all_letters_+ve_prob ] is violated , then @xmath370 . if assumption [ ass : all_letters_+ve_prob ] does not hold , then only at most two letters are used at each level .
it follows that we can have a ( possibly node - dependent ) scheme with binary messages that is equivalent to the original scheme at levels @xmath139 and higher .
our lower bound on @xmath89 then follows from theorem [ thm : binary_lower_bound ] .
thus , even in the best case , performance is significantly worse than the scheme presented in section [ subsec : node - oblivious_efficient_scheme ] .
so a good scheme for ternary messages must satisfy assumption [ ass : all_letters_+ve_prob ]
. now consider assumption [ ass : strongly_connected ] .
let @xmath371 .
suppose assumption [ ass : strongly_connected ] is violated .
then wlog , there is no path from letter @xmath45 to one of the other letters .
it follows that under either hypothesis , we have @xmath372 for node @xmath20 at level @xmath149 .
thus , the letter @xmath45 occurs with exponentially small probability , irrespective of @xmath2 .
this should essentially reduce , then , to the case of binary messages , and we expect performance to be constrained as above
. finally , consider assumption [ ass : one_letter_dominant ] .
we can not have @xmath373 for all @xmath278 , since that will lead to @xmath374 for all @xmath0 .
similarly , we can also exclude the possibility @xmath375 for all @xmath278 .
wlog , suppose @xmath376 and @xmath377 . now consider the problem of designing a good aggregation protocol . by the above , we must have @xmath378 and @xmath379 , for node @xmath20 at level @xmath149 , to each converge to 0 with increasing @xmath149 .
further , it appears natural to use the message @xmath380 with an interpretation of ` not sure ' in such a situation .
we would then like the probability of this intermediate symbol to decay with @xmath149 , or at least be bounded in the limit , i.e. , @xmath381 for each possible @xmath2 .
if this holds , we immediately have assumption [ ass : one_letter_dominant ] ( with @xmath382 and @xmath383 ) .
we argued above that our irreducibility assumptions are quite reasonable in various circumstances .
in fact , we expect the assumptions to be a proof artifact , and conjecture that a subexponential convergence bound holds for general node - oblivious rules .
a possible approach to eliminate our assumptions would be to prune the message alphabet @xmath3 , discarding letters that never appear , or appear with probability bounded by @xmath384 ( because they require descendants from a strict subset of @xmath3 ) . | we consider the decentralized binary hypothesis testing problem on trees of bounded degree and increasing depth . for a regular tree of depth @xmath0 and branching factor @xmath1 , we assume that the leaves have access to independent and identically distributed noisy observations of the ` state of the world ' @xmath2 . starting with the leaves , each node makes a decision in a finite alphabet @xmath3 , that it sends to its parent in the tree .
finally , the root decides between the two possible states of the world based on the information it receives .
we prove that the error probability vanishes only subexponentially in the number of available observations , under quite general hypotheses .
more precisely the case of binary messages , decay is subexponential for any decision rule . for general ( finite ) message alphabet @xmath3 ,
decay is subexponential for ` node - oblivious ' decision rules , that satisfy a mild irreducibility condition . in the latter case
, we propose a family of decision rules with close - to - optimal asymptotic behavior . |
geometric integrators are numerical methods that preserve geometric structures and properties of the flow of a differential equation .
structure - preserving integrators have attracted considerable interest due to their excellent numerical behavior , especially for long - time integration of equations possessing geometric properties ( see @xcite , @xcite , @xcite ) . an important class of structure - preserving integrators
are _ variational integrators _ ( see @xcite ) .
this type of numerical schemes is based on discrete variational principles and provides a natural framework for the discretization of lagrangian systems , including forced , dissipative , or constrained ones .
variational integrators were introduced in the context of finite - dimensional mechanical systems , but were later generalized to lagrangian field theories ( see @xcite ) and applied in many computations , for example in elasticity ( @xcite ) , electrodynamics ( @xcite ) , or fluid dynamics ( @xcite ) .
theoretical aspects of variational integration are well understood in the case when the lagrangian describing the considered system is regular , that is , when the corresponding legendre transform is ( at least locally ) invertible .
however , the corresponding theory for degenerate lagrangian systems is less developed .
the analysis of degenerate systems becomes a little more cumbersome , because the euler - lagrange equations may cease to be second order , or may not even make any sense at all . in the latter case one needs to determine if there exists a submanifold of the configuration bundle @xmath0 on which consistent equations of motion can be derived
this can be accomplished by applying the dirac theory of constraints or the pre - symplectic constraint algorithm ( see @xcite , @xcite ) .
a particularly simple case of degeneracy occurs when the lagrangian is linear in velocities . in that case
, the dynamics of the system is defined on the configuration manifold @xmath1 itself , rather than its tangent bundle @xmath0 , provided that some regularity conditions are satisfied .
such systems arise in many physical applications , including interacting point vortices in the plane ( see @xcite , @xcite , @xcite ) , or partial differential equations such as the nonlinear schrdinger ( @xcite ) , kdv ( @xcite , @xcite ) or camassa - holm equations ( @xcite , @xcite ) . in section [ sec : numerical experiments ] we show how certain poisson systems can be recast as lagrangian systems whose lagrangians are linear in velocities
. therefore , our approach offers a new perspective on geometric integration of poisson systems , which often arise as semi - discretizations of integrable nonlinear partial differential equations , e.g. , the toda or volterra lattice equations , and play an important role in the modeling of many physical phenomena ( see @xcite , @xcite , @xcite ) .
this paper is organized as follows . in section [ sec : geometric setup ] we introduce a proper geometric setup and discuss the properties of systems that are linear in velocities . in section [ sec : veselov discretization and discrete mechanics ] we analyze the general properties of variational integrators and point out how the relevant theory differs from the non - degenerate case . in section
[ sec : variational partitioned runge - kutta methods ] we introduce variational partitioned runge - kutta methods and discuss their relation to numerical integration of differential - algebraic systems . in section [ sec : numerical experiments ] we present the results of our numerical experiments for kepler s problem , a system of two interacting vortices , and the lotka - volterra model .
we summarize our work and discuss possible extensions in section [ sec : summary ] .
let @xmath1 be the configuration manifold and @xmath0 its tangent bundle . throughout this work
we will assume that the dimension of the configuration manifold @xmath2 is even .
we will further assume @xmath1 is a vector space and by a slight abuse of notation we will denote by @xmath3 both an element of @xmath1 and the vector of its coordinates @xmath4 in a local chart on @xmath1 .
it will be clear from the context which definition is invoked .
consider the lagrangian @xmath5 given by @xmath6 where @xmath7 is a smooth one - form , @xmath8 is the hamiltonian , and @xmath9 .
let @xmath10 denote canonical coordinates on @xmath0 , where @xmath11 . in these coordinates we can consider @xmath12 where summation over repeated greek indices is implied .
the lagrangian is degenerate , since the associated legendre transform @xmath13 is not invertible .
the local representation of the legendre transform is @xmath14 that is , @xmath15 where @xmath16 denote canonical coordinates on @xmath17 .
the dynamics is defined by the action functional @xmath18 = \int_{a}^{b } l\big(q(t),\dot q(t)\big)\,dt\ ] ] and hamilton s principle , which seeks the curves @xmath19 such that the functional @xmath20 $ ] is stationary under variations of @xmath19 with fixed endpoints , i.e. , we seek @xmath19 such that @xmath21 \cdot \delta q(t)=\frac{d}{d\epsilon } \bigg|_{\epsilon=0}s[q_\epsilon(t)]=0\end{aligned}\ ] ] for all @xmath22 with @xmath23 , where @xmath24 is a smooth family of curves satisfying @xmath25 and @xmath26 .
the resulting euler - lagrange equations @xmath27 form a system of _ first - order _ odes , where we assume that the even - dimensional antisymmetric matrix @xmath28 is invertible for all @xmath29 . without loss of generality
we can further assume that the coordinate mapping @xmath30 is invertible and its inverse is smooth : if the jacobian @xmath31 is singular , we can redefine @xmath32 , where @xmath33 are arbitrary functions ; the euler - lagrange equations remain the same , and with the right choice of the functions @xmath33 , the redefined jacobian can be made nonsingular . using @xmath34 , eq .
can be equivalently written as the poisson system @xmath35 the euler - lagrange equations can also be formulated as the implicit hamiltonian system @xmath36 since the lagrangian @xmath37 is degenerate , is an index-1 system of differential - algebraic equations ( dae ) , rather than a hamiltonian ode system : the legendre transform is an algebraic equation and has to be differentiated once with respect to time in order to turn this system into .
this reflects the fact that the evolution of the considered degenerate system takes place on the _ primary constraint _ @xmath38 .
it is easy to see that the primary constraint @xmath39 is ( locally ) diffeomorphic to the configuration manifold @xmath1 , where the diffeomorphism @xmath40 is locally , in the coordinates on @xmath17 , given by @xmath41 where by a slight abuse of notation @xmath42 .
this shows that @xmath43 can also be used as local coordinates on @xmath39 .
note that @xmath44 is simply the restriction of @xmath45 to @xmath39 , i.e. , @xmath46 .
the spaces @xmath1 , @xmath0 , @xmath17 and @xmath39 can be equipped with several symplectic or pre - symplectic forms .
it is instructive to investigate the relationships between them in order to later avoid confusion regarding the sense in which variational integrators for lagrangians linear in velocities are symplectic . on the configuration space
@xmath1 we can define the two - form @xmath47 which in local coordinates can be expressed as @xmath48 the two - form @xmath49 is symplectic if it is nondegenerate , i.e. , if the matrix @xmath50 is invertible for all @xmath3 .
the cotangent bundle @xmath17 is equipped with the canonical cartan one - form @xmath51 , which is intrinsically defined by the formula @xmath52 for any @xmath53 , where @xmath54 is the cotangent bundle projection . in canonical coordinates
we have @xmath55 we further have the canonical symplectic two - form @xmath56 the symplectic forms @xmath49 and @xmath57 are related by @xmath58 this follows from the simple calculation @xmath59 where we used and the fact that @xmath60 .
hence @xmath61 , and taking the exterior derivative on both sides we obtain . using the legendre
transform we can define the lagrangian two - form @xmath62 on @xmath0 by @xmath63 , which in canonical coordinates @xmath64 is given by @xmath65 the lagrangian form @xmath62 is only pre - symplectic , because it is degenerate . noting that @xmath66 , where @xmath67 is the tangent bundle projection , we can relate @xmath49 and @xmath62 through the formula @xmath68 the symplectic structure on @xmath39 can be introduced in two ways : by pushing forward @xmath49 from @xmath1 , or pulling back @xmath57 from @xmath17 .
both ways are equivalent @xmath69 where @xmath70 is the inclusion map .
this follows from the calculation @xmath71 where we used @xmath72 .
if we use @xmath43 as coordinates on @xmath39 , then the local representation of @xmath73 will be given by .
let @xmath74 denote the flow of or .
this flow is symplectic on @xmath1 , that is @xmath75 this fact can be proven by considering the hamiltonian or poisson properties of equation or equation ( see @xcite , @xcite ) .
it also follows directly from the action principle ( see @xcite ) .
since the lagrangian is degenerate , the dynamics of the system is defined on @xmath1 rather than @xmath0 .
however , we can obtain the associated flow on @xmath0 through lifting @xmath76 by its tangent map @xmath77 .
this flow preserves the lagrangian two - form @xmath78 this can be seen from the calculation @xmath79 where we used , , and the property @xmath80 .
the flow @xmath76 induces the flow @xmath81 in a natural way as @xmath82 this flow is symplectic on @xmath39 , i.e. , @xmath83 which can be established through the simple calculation @xmath84 where we used and .
the flow @xmath85 can be interpreted as the symplectic flow for the hamiltonian dae .
for a veselov - type discretization we consider the discrete state space @xmath86 , which serves as a discrete approximation of the tangent bundle ( see @xcite ) .
we define a discrete lagrangian @xmath87 as a smooth map @xmath88 and the corresponding discrete action @xmath89 the variational principle now seeks a sequence @xmath90 , @xmath91 , @xmath92 , @xmath93 that extremizes @xmath94 for variations holding the endpoints @xmath90 and @xmath93 fixed .
the discrete euler - lagrange equations follow @xmath95 assuming that these equations can be solved for @xmath96 , i.e. , @xmath87 is non - degenerate , they implicitly define the discrete lagrangian map @xmath97 such that @xmath98 .
let @xmath99 denote local coordinates on @xmath100 .
we can define the discrete legendre transforms @xmath101 , which in local coordinates on @xmath100 and @xmath17 are respectively given by @xmath102 where @xmath4 and @xmath103 .
the discrete euler - lagrange equations can be equivalently written as @xmath104 using either of the transforms , one can define the discrete lagrange two - form on @xmath86 by @xmath105 , which in coordinates gives @xmath106 it then follows that the discrete flow @xmath107 is symplectic , i.e. , @xmath108 . using the legendre transforms we can pass to the cotangent bundle and define the discrete hamiltonian map @xmath109 by @xmath110 .
this map is also symplectic , i.e. , @xmath111 . to relate discrete and continuous mechanics
it is necessary to introduce a timestep @xmath112 . if the continuous lagrangian @xmath37 is non - degenerate , it is possible to define a particular choice of discrete lagrangian which gives an exact correspondence between discrete and continuous systems ( see @xcite ) , the so - called _ exact discrete lagrangian _ @xmath113 where @xmath114 is the solution to the continuous euler - lagrange equations associated with @xmath37 such that it satisfies the boundary conditions @xmath115 and @xmath116 .
note , however , that in the case of a regular lagrangian the associated euler - lagrange equations are second order , therefore boundary value problems are solvable , at least for sufficiently small @xmath117 and @xmath118 sufficiently close to @xmath3 . in the case of the lagrangian
the associated euler - lagrange equations are first order in time , therefore we have the freedom to choose an initial condition either at @xmath119 or @xmath120 , but not both . an exact discrete lagrangian analogous to can not thus be defined on whole @xmath86
. we will therefore assume the following definition : let @xmath121 be the graph of @xmath122 .
the exact discrete lagrangian @xmath123 for the lagrangian is @xmath124 where @xmath114 is the solution to that satisfies the initial condition @xmath115 .
note that in this definition we automatically have @xmath116 .
as mentioned , the purpose of introducing an exact discrete lagrangian is to establish an exact correspondence between the continuous and discrete systems . for a regular lagrangian @xmath37 and its exact discrete lagrangian @xmath125
, one can show that the exact discrete hamiltonian map @xmath126 is equal to @xmath127 , where @xmath85 is the symplectic flow for the hamiltonian system associated with @xmath37 .
the problem is that the exact discrete lagrangian is not defined on the whole space @xmath100 , so the discrete euler - lagrange equations do not make sense , and it is not entirely clear how to define the associated discrete lagrangian map @xmath128 .
one possible way to deal with this issue is to consider a singular perturbation problem .
assume that @xmath1 is a riemannian manifold equipped with the nondegenerate scalar product @xmath129 .
define the @xmath130-regularized lagrangian @xmath131 or in coordinates @xmath132 where @xmath133 denotes the local coordinates of the metric tensor . without loss of generality
assume that in the chosen coordinates @xmath134 and @xmath135 if @xmath136 . for @xmath137
this lagrangian is nondegenerate and the legendre transform @xmath138 is given by @xmath139 that is , @xmath140 the euler - lagrange equations @xmath141 are second order .
the corresponding hamiltonian equations ( in implicit form ) are @xmath142 there is no reason to expect that the solutions of or unconditionally approximate the solutions of or , respectively .
the equations form a system of first - order ordinary differential equations , and therefore it is possible to specify arbitrary initial conditions @xmath143 and @xmath144 , whereas initial conditions for have to satisfy the algebraic constraint @xmath145 . under certain restrictive analytic assumptions , for some singular perturbation problems it is possible to show that , in order to satisfy the initial conditions , the solutions initially develop a steep boundary layer , but then rapidly converge to the solution of the corresponding dae system ( see @xcite ) . on the other hand , for other singular perturbation problems , when the initial conditions do not satisfy the algebraic constraint
, it may happen that the solutions do not converge to the solution of the dae , but instead rapidly oscillate ( see @xcite , @xcite ) .
we expect the latter behavior for , as will be demonstrated by a simple example in section [ sec : example ] .
since our main goal here is to show how the notion of a discrete legendre transform can be introduced for the exact discrete lagrangian , we will make two intuitive , although nontrivial , assumptions .
we refer the interested reader to @xcite and @xcite for techniques that can be used to prove these statements rigorously .
[ ass : initial value problem assumption ] let @xmath146 and @xmath147 be the unique smooth solutions of and on the interval @xmath148 $ ] satisfying the initial conditions @xmath143 , @xmath149 and @xmath150 , where @xmath151
. then @xmath152 , @xmath153 and @xmath154 , @xmath155 uniformly on @xmath148 $ ] as @xmath156 .
[ ass : boundary value problem assumption ] let @xmath19 be the unique smooth solution of on the interval @xmath148 $ ] satisfying the initial condition @xmath143 and let @xmath157 be the unique smooth solution of on the interval @xmath148 $ ] satisfying the boundary conditions @xmath149 , @xmath158 , where @xmath159
. then @xmath152 and @xmath154 uniformly on @xmath148 $ ] as @xmath156 .
with these assumption one can easily see that @xmath160 where @xmath161 is the exact discrete lagrangian for . since @xmath162 is regular , @xmath161 is properly defined on the whole space @xmath100 ( or at least in a neighborhood of @xmath163 ) and the associated exact discrete legendre transforms satisfy the properties ( see @xcite ) @xmath164 where @xmath165 is the solution to the regularized euler - lagrange equations satisfying the boundary conditions @xmath166 and @xmath167 , and we denoted @xmath168 , @xmath169 . in the spirit of
, we can assume the following definitions of the exact discrete legendre transforms @xmath170 @xmath171 where @xmath172 and @xmath173 by uniform convergence of @xmath174 .
note that @xmath175 , where @xmath176 and @xmath177 are projections ( both @xmath178 are diffeomorphisms ) .
this is a close analogy to @xmath179 ( see section [ sec : geometric setup ] ) .
we also note the property @xmath180 where @xmath114 is the solution of satisfying the initial condition @xmath115 .
this further indicates that our definition of the exact discrete legendre transforms is sensible .
note that @xmath181 .
it is convenient to redefine @xmath182 , that is @xmath183 , so that both transforms are diffeomorphisms between @xmath163 and @xmath39 .
the discrete euler - lagrange equations for @xmath125 can be obtained as the limit of the discrete euler - lagrange equations for @xmath161 , that is , one can substitute @xmath161 in and take the limit @xmath156 on both sides to obtain @xmath184 this equation implicitly defines the exact discrete lagrangian map @xmath185 , which , given our definitions , necessarily takes the form @xmath186 .
using the discrete legendre transforms @xmath187 we can define the corresponding exact discrete hamiltonian map @xmath188 as @xmath189 .
the simple calculation @xmath190 shows that the discrete hamiltonian map associated with the exact discrete lagrangian @xmath125 is equal to the hamiltonian flow @xmath127 for , i.e. , the evolution of the discrete systems described by @xmath125 coincides with the evolution of the continuous system described by @xmath37 at times @xmath191 , @xmath192 let us illustrate these ideas with a very simple example for which analytic solutions are known .
let @xmath193 and let @xmath194 denote local coordinates on @xmath1 .
the tangent bundle is @xmath195 , and the induced local coordinates are @xmath196 .
consider the lagrangian @xmath197 the corresponding euler - lagrange equations are simply @xmath198 so the flow @xmath199 is the identity , i.e. , @xmath200 .
let @xmath201 denote canonical coordinates on the cotangent bundle @xmath202 .
the legendre transform is @xmath203 let @xmath117 be a timestep .
note @xmath204 .
the exact discrete lagrangian is therefore @xmath205 let us now consider the @xmath130-regularized lagrangian @xmath206 the corresponding euler - lagrange equations take the form @xmath207 one can easily verify analytically that @xmath208 & + \frac{1}{2}\bigg [ ( y_f - y_i)+(x_f - x_i ) \cot \frac{t}{2 \epsilon}\bigg ] \sin \frac{t}{\epsilon } \nonumber \\ & - \frac{1}{2}\bigg [ ( x_f - x_i)-(y_f - y_i ) \cot \frac{t}{2 \epsilon}\bigg ] \cos \frac{t}{\epsilon } , \nonumber \\
y^\epsilon(t ) = \frac{1}{2}\bigg [ ( y_i+y_f)+(x_f - x_i ) \cot \frac{t}{2 \epsilon}\bigg ] & -\frac{1}{2}\bigg [ ( x_f - x_i)-(y_f - y_i ) \cot \frac{t}{2 \epsilon}\bigg ] \sin \frac{t}{\epsilon } \nonumber \\ & - \frac{1}{2}\bigg [ ( y_f - y_i)+(x_f - x_i ) \cot \frac{t}{2 \epsilon}\bigg ] \cos \frac{t}{\epsilon},\end{aligned}\ ] ] is the solution to satisfying the boundary conditions @xmath209 and @xmath210 .
note that if @xmath211 or @xmath212 , then as @xmath156 this solution is rapidly oscillatory and not convergent .
however , if @xmath213 ( cf .
assumption [ ass : boundary value problem assumption ] ) then we have @xmath214 and this solution converges uniformly ( in this simple example it is in fact equal ) to the solution of with the same initial condition .
we can also find an analytic expression for the exact discrete lagrangian associated with as @xmath215 restricting the domain to @xmath163 we get @xmath216 , and comparing to we verify that indeed holds .
the discrete legendre transforms associated with @xmath161 take the form @xmath217 restricting the domain to @xmath163 and taking the limit @xmath156 as in , we can define the exact discrete legendre transforms associated with @xmath218 comparing with , we see that the property is satisfied , which replicates the analogous property for regular lagrangians . for a given continuous system described by the lagrangian @xmath37 ,
a variational integrator is constructed by choosing a discrete lagrangian @xmath87 which approximates the exact discrete lagrangian @xmath125 .
we can define the order of accuracy of the discrete lagrangian in a way similar to that for discrete lagrangians resulting from regular continuous lagrangians ( see @xcite ) .
a discrete lagrangian @xmath219 is of order @xmath220 if there exists an open subset @xmath221 with compact closure and constants @xmath222 and @xmath223 such that @xmath224 for all solutions @xmath19 of the euler - lagrange equations with initial conditions @xmath225 and for all @xmath226 .
we will always assume that the discrete lagrangian @xmath87 is non - degenerate , so that the discrete euler - lagrange equations can be solved for @xmath96 .
this defines the discrete lagrangian map @xmath227 and the associated discrete hamiltonian map @xmath228 , as in section [ sec : discrete mechanics ] .
of particular interest is the rate of convergence of @xmath229 to @xmath127 .
one usually considers a _ local error _
( error made after one step ) and a _ global error _ ( error made after many steps )
. we will assume the following definitions , which are appropriate for differential - algebraic systems ( see @xcite , @xcite , @xcite , @xcite ) .
a discrete hamiltonian map @xmath229 is of order @xmath220 if there exists an open set @xmath230 and constants @xmath222 and @xmath231 such that @xmath232 for all @xmath233 and @xmath234 .
[ thm : definition of the order of convergence ] a discrete hamiltonian map @xmath229 is convergent of order @xmath220 if there exists an open set @xmath230 and constants @xmath222 , @xmath231 and @xmath235 such that @xmath236 where @xmath237 , for all @xmath233 , @xmath234 , and @xmath238 . if the lagrangian @xmath37 is regular , then one can show that a discrete lagrangian @xmath87 is of order @xmath220 if and only if the corresponding hamiltonian map @xmath229 is of order @xmath220 ( see @xcite ) .
also , the associated hamiltonian equations are a set of ordinary differential equations , and under some smoothness assumptions one can show that if @xmath229 is of order @xmath220 , then it is also convergent of order @xmath220 ( see @xcite ) . however ,
in the case of the lagrangian it is not true in general both the order of the discrete lagrangian and the local order of the discrete hamiltonian map may be different than the actual global order of convergence ( see @xcite , @xcite ) , as will be demonstrated in section [ sec : variational partitioned runge - kutta methods ] .
[ [ example - midpoint - rule . ] ] example : midpoint rule .
+ + + + + + + + + + + + + + + + + + + + + + + in a simple example we will demonstrate that the variational order of accuracy of a discretization method is unaffected by a degeneracy of a lagrangian @xmath37 . in order to calculate the order of a discrete lagrangian @xmath87
, we can expand @xmath239 in a taylor series in @xmath117 and compare it to the analogous expansion for @xmath125 . if the two expansions agree up to @xmath220 terms , then @xmath87 is of order @xmath220 . expanding @xmath19 in a taylor series about @xmath119 and substituting it in , we get the expression @xmath240 where we denoted @xmath241 , @xmath242 , etc . , and the lagrangian l and its derivatives are computed at @xmath243 . for the lagrangian the values of @xmath244 , @xmath245 , @xmath246
are determined by differentiating sufficiently many times and substituting the initial condition @xmath247 .
note that in case of regular lagrangians the value of @xmath244 is determined by the boundary conditions @xmath247 , @xmath248 , and the higher - order derivatives by differentiating the corresponding euler - lagrange equations , but apart from that the expression remains qualitatively unaffected .
the _ midpoint rule _ is an integrator obtained by defining the discrete lagrangian @xmath249 calculating the expansion in @xmath117 yields @xmath250 comparing this to shows that the discrete lagrangian defined by the midpoint rule is second order regardless of the degeneracy of @xmath37 .
however , as mentioned before , if @xmath37 is degenerate we can not conclude about the global order of convergence of the corresponding discrete hamiltonian map .
the midpoint rule can be formulated as a runge - kutta method , namely the 1-stage gauss method .
we discuss gauss and other runge - kutta methods and their convergence properties in more detail in section [ sec : variational partitioned runge - kutta methods ] .
note that low - order variational integrators for lagrangians based on the midpoint rule have been studied in @xcite and @xcite in the context of the dynamics of point vortices .
to construct higher - order variational integrators one may consider a class of partitioned runge - kutta ( prk ) methods .
variational partitioned runge - kutta ( vprk ) methods for regular lagrangians are described in @xcite and @xcite . in this section
we show how vprk methods can be applied to systems described by lagrangians such as . as in the case of regular lagrangians
, we will construct an @xmath251-stage variational partitioned runge - kutta integrator for the lagrangian by considering the discrete lagrangian @xmath252 where the internal stages @xmath253 , @xmath254 , @xmath255 , satisfy the relation @xmath256 and are chosen so that the right - hand side of is extremized under the constraint @xmath257 a variational integrator is then obtained by forming the corresponding discrete euler - lagrange equations .
the @xmath251-stage variational partitioned runge - kutta method based on the discrete lagrangian with the coefficients @xmath258 and @xmath259 is equivalent to the following partitioned runge - kutta method applied to the hamiltonian dae : [ eq : prk for dae ] @xmath260^t \dot q_i - dh(q_i ) , } \qquad \qquad \qquad i=1,\ldots , s,\\ \label{eq : prk for dae 2 } \dot p^i & = [ d\alpha(q_i)]^t \dot q_i - dh(q_i ) , \qquad \qquad \qquad \ !
i=1,\ldots , s , \\
\label{eq : prk for dae 3 } q_i & = q + h \sum_{j=1}^s a_{ij } \dot q_j , \phantom{-dh(q_i ) , } \qquad \qquad \qquad i=1,\ldots , s , \\
\label{eq : prk for dae 4 } p^i & = p + h \sum_{j=1}^s \bar a_{ij } \dot p_j , \phantom{-dh(q_i ) , } \qquad \qquad \qquad \ , i=1,\ldots , s,\\ \label{eq : prk for dae 5 } \bar q & = q + h \sum_{j=1}^s b_j \dot q_j,\\ \label{eq : prk for dae 6 } \bar p & = p + h \sum_{j=1}^s b_j \dot p_j,\end{aligned}\ ] ] where the coefficients satisfy the condition @xmath261 and @xmath262 denote the current values of position and momentum , @xmath263 denote the respective values at the next time step , @xmath264 , @xmath265 , and @xmath253 , @xmath254 , @xmath266 , @xmath267 are the internal stages , with @xmath268 , and similarly for the others .
see theorem vi.6.4 in @xcite or theorem 2.6.1 in @xcite .
the proof is essentially identical .
the only qualitative difference is the fact that in our case the lagrangian is degenerate , so the corresponding hamiltonian system is in fact the index-1 differential - algebraic system rather than a typical system of ordinary differential equations .
+ [ [ existence - and - uniqueness - of - the - numerical - solution . ] ] existence and uniqueness of the numerical solution .
+ + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + given @xmath3 and @xmath269 , one can use equations to compute the new position @xmath118 and momentum @xmath270 .
first , one needs to solve - for the internal stages @xmath253 , @xmath254 , @xmath266 , and @xmath267 .
this is a system of @xmath271 equations for @xmath271 variables , but one has to make sure these equations are independent , so that a unique solution exists .
one may be tempted to calculate the jacobian of this system for @xmath272 , and then use the implicit function theorem . however , even if we start with consistent initial values @xmath273 , the numerical solution @xmath274 for @xmath275 will only approximately satisfy the algebraic constraint ; so @xmath276 and @xmath277 can not be assumed to be the solution of - for @xmath272 , and consequently , the implicit function theorem will not yield a useful result .
let us therefore regard @xmath3 and @xmath269 as @xmath117-dependent , as they result from the previous iterations of the method with the timestep @xmath117 .
if the method is convergent , it is reasonable to expect that @xmath278 is small and converges to zero as @xmath117 is refined .
the following approach was inspired by theorem 4.1 in @xcite .
[ thm : existence of the numerical solution for prk ] let @xmath279 and @xmath45 be smooth in an @xmath117-independent neighborhood @xmath280 of @xmath3 and let the matrix @xmath281 be invertible with the inverse bounded in @xmath282 , i.e. , there exists @xmath222 such that @xmath283 where @xmath284 , @xmath285 , @xmath286 is the @xmath287 identity matrix , and @xmath288 denotes the block diagonal matrix @xmath289 suppose also that @xmath262 satisfy @xmath290 then there exists @xmath223 such that the nonlinear system - has a solution for @xmath226 .
the solution is locally unique and satisfies @xmath291 substitute and in and to obtain @xmath292 for @xmath255 , where for notational convenience we left the @xmath253 s as arguments of @xmath45 , @xmath293 and @xmath294 , but we keep in mind they are defined by , so that is a nonlinear system for @xmath254 and @xmath267 .
let us consider the homotopy @xmath295 for @xmath255 .
it is easy to see that for @xmath296 the system has the solution @xmath297 and @xmath298 , and for @xmath299 it is equivalent to .
let us treat @xmath254 and @xmath267 as functions of @xmath300 , and differentiate with respect to this parameter .
the resulting ode system can be written as [ eq : ode system for dot qi and dot pi ] @xmath301 where for compactness we introduced the following notations : @xmath302 , similarly for @xmath303 ; @xmath304 is the @xmath251-dimensional vector of ones ; @xmath305 , and similarly , @xmath306 denotes the block diagonal matrix @xmath307 with @xmath308 , where @xmath309 denotes the hessian matrix of the respective function , and summation over @xmath310 is implied .
the system is further simplified if we substitute in . this way we obtain an ode system for the variables @xmath311 of the form @xmath312 \frac{d \dot q}{d\tau } = \nonumber \\ & ( \mathcal{\bar{a } } \mathbbm{1}_s)\otimes dh(q ) - \frac{1}{h}\mathbbm{1}_s\otimes\big(p-\alpha(q)\big).\end{aligned}\ ] ] since @xmath45 is smooth , we have @xmath313_{ij } = a_{ij } d\alpha(q_i ) = a_{ij } d\alpha(q_j ) + o(\delta ) = \big [ ( \mathcal{a}\otimes i_n ) \{d\alpha\ } \big]_{ij } + o(\delta),\ ] ] where @xmath314 for @xmath315 assumed small , but independent of @xmath117 . moreover , since @xmath45 and @xmath279 are smooth , the term @xmath306 , as a function of @xmath311 , is bounded in a neighborhood of 0 .
therefore , we can write as @xmath316 \frac{d \dot q}{d\tau } = ( \mathcal{\bar{a } } \mathbbm{1}_s)\otimes dh(q ) - \frac{1}{h}\mathbbm{1}_s\otimes\big(p-\alpha(q)\big).\end{aligned}\ ] ] by , for sufficiently small @xmath117 and @xmath315 , the matrix @xmath317 has a bounded inverse , provided that @xmath318 remain in @xmath280 .
therefore , the ode with the initial condition @xmath319 has a unique solution @xmath320 on a non - empty interval @xmath321 , which can be extended until any of the corresponding @xmath322 leaves @xmath280 .
let us argue that for a sufficiently small @xmath117 we have @xmath323 . given and
, the ode implies that @xmath324 therefore , we have @xmath325 and further @xmath326 for @xmath327 .
this implies that all @xmath322 remain in @xmath280 for @xmath328 if @xmath117 is sufficiently small .
consequently , the ode has a solution on the interval @xmath329 $ ] .
then @xmath330 and @xmath331 satisfy the estimates , and are a solution to the nonlinear system - .
the corresponding @xmath267 and @xmath266 can be computed using and , and the remaining estimates can be proved using the fact that @xmath45 and @xmath279 are smooth .
this completes the proof of the existence of a numerical solution to - . in order to prove local uniqueness , we substitute the second equation of in the first one to obtain a nonlinear system for @xmath254 , namely @xmath332 for @xmath255 , where we again left the @xmath253 s for notational convenience .
suppose there exists another solution @xmath333 that satisfies the estimates , and denote @xmath334 .
based on the assumptions , we have @xmath335 , i.e. , it is at least bounded as @xmath336 .
we will show that for sufficiently small @xmath117 we in fact have @xmath337
. since @xmath338 satisfy , we have @xmath339 for @xmath255 .
subtract from , and linearize around @xmath254 . based on the fact that @xmath335 , and using the notation introduced before , we get @xmath340 \delta \dot q + o(h^2 \|\delta \dot q \|).\ ] ] by a similar argument as before , for sufficiently small @xmath117 the matrix @xmath341 $ ] has a bounded inverse , therefore implies @xmath342 , that is , @xmath343 for some constant @xmath344 .
note that for @xmath345 we have @xmath346 , and therefore @xmath347 , which completes the proof of the local uniqueness of a numerical solution to - . .
[ [ remarks . ] ] remarks .
+ + + + + + + + the condition may be tedious to verify , especially if one uses a runge - kutta method with many stages . however , this condition is significantly simplified in the following special cases : 1 . for a non - partitioned runge - kutta method
we have @xmath348 , and the condition is satisfied if @xmath349 is invertible , and the mass matrix @xmath350 , as defined in section [ sec : equations of motion ] , is invertible in @xmath280 and its inverse is bounded .
if @xmath351 is antisymmetric , then the condition is satisfied if @xmath352 is invertible , and the matrix @xmath353 is invertible in @xmath280 and its inverse is bounded .
an interesting special case is obtained if we have , in some local chart on @xmath1 , @xmath355 for some constant matrix @xmath356 . without loss of generality
assume that @xmath356 is invertible and antisymmetric .
the lagrangian then takes the form @xmath357 the euler - lagrange equations become @xmath358 and the hamiltonian dae system is @xmath359 let us consider a special case of the method with @xmath360 , i.e. , a non - partitioned runge - kutta method . applying it to we get [ eq : prk for dae for linear alpha ] @xmath361
since @xmath356 is antisymmetric and invertible , then by theorem [ thm : existence of the numerical solution for prk ] the scheme yields a unique numerical solution to if the runge - kutta matrix @xmath362 is invertible .
[ thm : equivalence with a rk for the ode ] suppose @xmath362 is invertible and @xmath363 . then the method is equivalent to the same runge - kutta method applied to .
substitute and in , and use the fact @xmath363 to obtain @xmath364 since @xmath349 is invertible , this implies @xmath365 substituting this in yields @xmath366 together with and , this gives a runge - kutta method for . moreover , substituting and @xmath363 in , and using , one has @xmath367 that is , @xmath368 satisfy the algebraic constraint .
+ [ thm : invariance of the primary constraint ] the numerical flow on @xmath17 defined by leaves the primary constraint @xmath39 invariant , i.e. , if @xmath369 , then @xmath370 .
if the coefficients of the method satisfy the condition , then is a variational integrator and the associated discrete hamiltonian map @xmath229 is symplectic on @xmath17 , as explained in section [ sec : discrete mechanics ] . given corollary [ thm : invariance of the primary constraint ] , we further have : [ thm : symplecticity on the primary constraint ] if the coefficients @xmath258 and @xmath259 in satisfy the condition , then the discrete hamiltonian map @xmath229 associated with is symplectic on the primary constraint @xmath39 , that is , @xmath371 .
[ [ convergence . ] ] convergence .
+ + + + + + + + + + + + various runge - kutta methods and their classical orders of convergence , that is , orders of convergence when applied to ( non - stiff ) ordinary differential equations , are discussed in many textbooks on numerical analysis , for instance @xcite and @xcite . when applied to differential - algebraic equations , the order of convergence of a runge - kutta method may be reduced ( see @xcite , @xcite , @xcite ) . however , in the case of theorem [ thm : equivalence with a rk for the ode ] implies that the classical order of convergence of non - partitioned runge - kutta methods is retained .
[ thm : retention of the classical order of convergence ] a runge - kutta method with the coefficients @xmath258 and @xmath259 applied to the dae system retains its classical order of convergence .
let @xmath220 be the classical order of the considered runge - kutta method , @xmath369 an initial condition , @xmath372 the exact solution to such that @xmath373 , and @xmath274 the numerical solution obtained by applying the method iteratively @xmath374 times with @xmath375 .
theorem [ thm : equivalence with a rk for the ode ] states that the method is equivalent to applying the same runge - kutta method to the ode system .
hence , we obtain convergence of order @xmath220 in the @xmath3 variable , that is , for a fixed time @xmath376 and an integer @xmath377 such that @xmath378 , we have the estimate @xmath379 for some constant @xmath222 ( cf .
definition [ thm : definition of the order of convergence ] ) . by corollary [ thm : invariance of the primary constraint ] we know that @xmath380 , so we have the estimate @xmath381 which completes the proof , since @xmath382 . + of particular interest to us are runge - kutta methods that satisfy the condition , for instance symplectic diagonally - implicit runge - kutta methods ( dirk ) or gauss collocation methods ( see @xcite ) .
the @xmath251-stage gauss method is of classical order @xmath383 , therefore we have : [ thm : order of convergence of gauss methods ] the @xmath251-stage gauss collocation method applied to the dae system is convergent of order @xmath383 .
as mentioned in section [ sec : variational error analysis ] , the midpoint rule is a 1-stage gauss method , therefore it retains its classical second order of convergence .
[ [ backward - error - analysis . ] ] backward error analysis .
+ + + + + + + + + + + + + + + + + + + + + + + + the system can be rewritten as the poisson system @xmath384 with the structure matrix @xmath385 ( see @xcite , @xcite ) .
the flow @xmath76 for this equation is a _ poisson map _ , that is , it satisfies the property @xmath386^{t } = \lambda^{-1},\ ] ] which is in fact equivalent to the symplecticity property or written in local coordinates on @xmath1 or @xmath39 , respectively .
let @xmath387 represent the numerical flow defined by some numerical algorithm applied to .
we say this flow is a _ poisson integrator _ if @xmath388^{t } = \lambda^{-1}.\ ] ] the left - hand side of can be regarded as a quadratic invariant of . by theorem
[ thm : equivalence with a rk for the ode ] the method is equivalent to applying the same runge - kutta method to . if its coefficients also satisfy the condition , then it can be shown that the method preserves quadratic invariants ( see theorem iv.2.2 in @xcite ) .
therefore , we have : [ thm : r - k methods as poisson integrators ] if @xmath362 is invertible , the coefficients @xmath258 and @xmath259 satisfy the condition , and @xmath363 , then the method is a poisson integrator for
. the true power of symplectic integrators for hamiltonian equations is revealed through their backward error analysis : a symplectic integrator for a hamiltonian system with the hamiltonian @xmath389 defines the _ exact _ flow for a nearby hamiltonian system , whose hamiltonian can be expressed as the asymptotic series @xmath390 owing to this fact , under some additional assumptions , symplectic numerical schemes nearly conserve the original hamiltonian @xmath389 over exponentially long time intervals ( see @xcite for details ) .
a similar result holds for poisson integrators for poisson systems : a poisson integrator defines the exact flow for a nearby poisson system , whose structure matrix is the same and whose hamiltonian has the asymptotic expansion ( see theorem ix.3.6 in @xcite ) .
therefore , we expect the non - partitioned runge - kutta schemes satisfying the condition to demonstrate good preservation of the original hamiltonian @xmath279 .
see section [ sec : numerical experiments ] for numerical examples .
partitioned runge - kutta methods do not seem to have special properties when applied to systems with linear @xmath354 , therefore we describe them in the general case in section [ sec : nonlinear alpha ] .
when the coordinates @xmath354 are nonlinear functions of @xmath3 , then the runge - kutta methods discussed in section [ sec : linear alpha ] lose some of their properties : a theorem similar to theorem [ thm : equivalence with a rk for the ode ] can not be proved , most of the runge - kutta methods ( whether non - partitioned or partitioned ) do not preserve the algebraic constraint @xmath391 , i.e. , the numerical solution does not stay on the primary constraint @xmath39 , and therefore their order of convergence is reduced , unless they are _
stiffly accurate_. let us again consider non - partitioned methods with @xmath360 .
convergence results for some classical runge - kutta schemes of interest can be obtained by transforming into a semi - explicit index-2 dae system .
let us briefly review this approach .
more details can be found in @xcite and @xcite .
the system can be written as the quasi - linear dae @xmath392 where @xmath393 and @xmath394^t & -i_n \\ 0 & 0 \\ \end{matrix } \right ) , \qquad\qquad\qquad f(y ) = \left ( \begin{matrix } dh(q ) \\
p-\alpha(q ) \\ \end{matrix } \right),\ ] ] where @xmath286 denotes the @xmath287 identity matrix .
let us introduce a slack variable @xmath395 and rewrite as the index-2 dae system [ eq : index 2 dae form ] @xmath396 this system is of index 2 , because it has @xmath397 dependent variables , but only @xmath398 differential equations , and some components of the algebraic equations have to be differentiated twice with respect to time in order to derive the missing differential equations for @xmath395 .
note that @xmath399 is a singular matrix of constant rank @xmath400 , therefore it can be decomposed ( using gauss elimination or the singular value decomposition ) as @xmath401 for some non - singular matrices @xmath402 and @xmath403 . since @xmath404 is assumed to be smooth , one can choose @xmath94 and @xmath405 so that they are also smooth ( at least in a neighborhood of @xmath406 ) .
premultiplying both sides of by @xmath407 turns the dae into [ eq : index 2 dae block form ] @xmath408 where we introduced the block structure @xmath409 , @xmath410 , and @xmath411 since @xmath403 is invertible , we can assume without loss of generality that the block @xmath412 is invertible , too ( one can always permute the columns of @xmath403 otherwise ) .
let us compute @xmath413 from and substitute it in .
the resulting system , [ eq : index 2 dae final block form ] @xmath414 has the form of a semi - explicit index-2 dae @xmath415 provided that @xmath416 has a bounded inverse .
it is an elementary exercise to show that the partitioned runge - kutta method is invariant under the presented transformation , that is , it defines a numerically equivalent partitioned runge - kutta method for .
runge - kutta methods for semi - explicit index-2 daes have been studied and some convergence results are available .
convergence estimates for the @xmath406 component of can be readily applied to the solution of . as in section
[ sec : linear alpha ] , of particular interest to us are variational runge - kutta methods , i.e. , methods satisfying the condition , for example gauss collocation methods ( see @xcite , @xcite ) .
however , in the case when @xmath404 is a nonlinear function , the solution generated by the gauss methods does not stay on the primary constraint @xmath39 and this affects their rate of convergence , as will be shown below . for comparison
, we will also consider the radau iia methods ( see @xcite ) , which , although not variational / symplectic , are _ stiffly accurate _ , that is , their coefficients satisfy @xmath417 for @xmath418 , so the numerical value of the solution at the new time step is equal to the value of the last internal stage , and therefore the numerical solution stays on the submanifold @xmath39 .
we cite the following convergence rates for the @xmath406 component of after @xcite and @xcite : * @xmath251-stage gauss method
convergent of order @xmath419 , * @xmath251-stage radau iia method
convergent of order @xmath420 . with the exception of the midpoint rule ( @xmath421 )
, we see that the order of convergence of the gauss methods is reduced . on the other hand ,
the radau iia methods retain their classical order @xmath420 .
[ [ symplecticity . ] ] symplecticity .
+ + + + + + + + + + + + + + since the gauss methods satisfy the condition , they generate a flow which preserves the canonical symplectic form @xmath57 on @xmath17 , as explained in section [ sec : discrete mechanics ]
. however , since the primary constraint @xmath39 is not invariant under this flow , a result analogous to corollary [ thm : symplecticity on the primary constraint ] does not hold , i.e. , the flow is not symplectic on @xmath39 . in section [ sec : numerical experiments ]
we present numerical results for the lobatto iiia - iiib methods ( see @xcite ) .
their numerical performance appears rather unattractive , therefore our theoretical results regarding partitioned runge - kutta methods are less complete .
below we summarize the experimental orders of convergence of the lobatto iiia - iiib schemes that we observed in our numerical computations ( see figure [ fig : convergence plots for kepler s problem ] , figure [ fig : convergence plot for point vortices ] , and figure [ fig : convergence plot for the lotka - volterra model ] ) : * @xmath422-stage lobatto iiia - iiib
inconsistent , * @xmath423-stage lobatto iiia - iiib convergent of order 2 , * @xmath424-stage lobatto iiia - iiib convergent of order 2
. comments regarding the symplecticity of these schemes are the same as for the gauss methods mentioned above in section [ sec : runge - kutta methods ] .
in this section we present the results of the numerical experiments we performed to test the methods discussed in section [ sec : variational partitioned runge - kutta methods ] .
we consider kepler s problem , the dynamics of planar point vortices , and the lotka - volterra model , and we show how each of these models can be formulated as a lagrangian system linear in velocities . a particle or a planet moving in a central potential in two dimensions can be described by the hamiltonian @xmath425 where @xmath194 denotes the position of the planet and @xmath426 its momentum ; @xmath427 is an arbitrary constant .
the corresponding lagrangian can be obtained in the usual way as @xmath428 if one performs the standard legendre transform @xmath429 , @xmath430 , then @xmath431 will take the usual nondegenerate form , quadratic in velocities .
however , one can also introduce the variable @xmath432 and view @xmath433 as , that is , a lagrangian linear in velocities ( see @xcite ) . comparing and
we see that the corresponding @xmath356 is singular . without loss of generality
we replace @xmath356 with its antisymmetric part @xmath434 , which is invertible , and consider the lagrangian @xmath435 as a test problem we considered an elliptic orbit with eccentricity @xmath436 and semi - major axis @xmath437 .
we took the initial condition at the pericenter , i.e. , @xmath438 , @xmath439 , @xmath440 , @xmath441 .
this is a periodic orbit with period @xmath442 .
a reference solution was computed by integrating until the time @xmath443 using verner s method ( a 6-th order explicit runge - kutta method ; see @xcite ) with the small time step @xmath444 .
the reference solution is depicted in figure [ fig : reference solution for kepler s problem ] . using verner s method with the time step @xmath444.,scaledwidth=80.0% ]
we solved the same problem using several of the methods discussed in section [ sec : variational partitioned runge - kutta methods ] for a number of time steps ranging from @xmath445 to @xmath446 .
the value of the solutions at @xmath443 was then compared against the reference solution .
the max norm errors are depicted in figure [ fig : convergence plots for kepler s problem ] .
we see that the rates of convergence of the gauss and the 3-stage radau iia methods are consistent with theorem [ thm : retention of the classical order of convergence ] and corollary [ thm : order of convergence of gauss methods ] . for the lobatto iiia - iiib methods we observe a reduction of order .
the 2-stage lobatto iiia - iiib method turns out to be inconsistent and is not depicted in figure [ fig : convergence plots for kepler s problem ] .
both the 3- and 4-stage methods converge only quadratically , while their classical orders of convergence are 4 and 6 , respectively .
we also investigated the long - time behavior of our integrators and conservation of the hamiltonian . for convenience
, we set @xmath447 in , so that @xmath448 on the considered orbit .
we applied the gauss methods with the relatively large time step @xmath449 and computed the numerical solution until the time @xmath450 .
figure [ fig : energy plots for gauss methods for kepler s problem ] shows that the gauss integrators preserve the hamiltonian very well , which is consistent with corollary [ thm : r - k methods as poisson integrators ] .
we performed similar computations for the lobatto iiia - iiib and radau iia methods , also with @xmath449 .
the results are depicted in figure [ fig : energy plots for lobatto and radau methods for kepler s problem ] .
the 3- and 4-stage lobatto iiia - iiib schemes result in instabilities , the planet s trajectory spirals down on the center of gravity , and the computations can not be continued too far in time . the hamiltonian shows major variations whose amplitude grows in time .
the non - variational radau iia scheme yields an accurate solution , but it demonstrates a gradual energy dissipation . over the time interval @xmath451 $ ] ( _ right column _ ) , with a close - up on the initial interval @xmath452 $ ] shown in the _ left column_.,scaledwidth=90.0% ] point vortices in the plane are another interesting example of a system with linear @xmath354 ( see @xcite , @xcite , @xcite ) .
a system of @xmath377 interacting point vortices in two dimensions can be described by the lagrangian @xmath453 with the hamiltonian @xmath454 where @xmath455 denotes the location of the @xmath456-th vortex , @xmath457 is its circulation , and @xmath427 is an arbitrary constant . as a test problem we considered the system of @xmath458 vortices with circulations @xmath459 and @xmath460 , respectively , and distance @xmath461 between them .
the vortices rotate on concentric circles about their center of vorticity at @xmath462 and @xmath463 .
we took the initial condition at @xmath464 , @xmath465 , @xmath466 and @xmath467 .
the analytic solution can be found ( see @xcite ) as @xmath468 where @xmath469 .
this is a periodic solution with period @xmath470 .
see figure [ fig : reference solution for point vortices ] . and
@xmath463.,scaledwidth=80.0% ] we performed similar convergence tests as in section [ sec : kepler s problem ] .
the value of the numerical solutions at time t=7 were compared against the exact solution .
the max norm errors are depicted in figure [ fig : convergence plot for point vortices ] .
the results are qualitatively the same as for kepler s problem . over the time interval @xmath451$].,scaledwidth=90.0% ] over the time interval @xmath451 $ ] ( _ right column _ ) , with a close - up on the initial interval @xmath471 $ ] shown in the _ left column_.,scaledwidth=80.0% ] we set @xmath472 in , so that @xmath448 for the considered solution . figure [ fig : energy plots for gauss methods for point vortices ] and figure [ fig : energy plots for lobatto methods for point vortices ] show the behavior of the numerical hamiltonian over a long integration interval .
the 3- and 4-stage lobatto iiia - iiib integrators performed better than for kepler s problem . in the case of the gauss methods the hamiltonian stayed virtually constant
the visible minor erratic oscillations are the result of round - off errors .
the radau iia scheme demonstrated a slow but systematic drift .
the dynamics of the growth of two interacting species can be modeled by the lotka - volterra equations @xmath473 where @xmath474 denotes the number of predators and @xmath475 the number of prey , and the constants 1 and 2 were chosen arbitrarily .
these equations can be rewritten as the poisson system @xmath476 where the hamiltonian is given by @xmath477 with an arbitraty constant @xmath427 ( see @xcite ) .
using an approach similar to the one presented in section [ sec : kepler s problem ] , one can easily verify that the lagrangian @xmath478 reproduces the same equations of motion , where @xmath479 .
the coordinates @xmath354 ( cf .
equation ) were chosen , so that the assumptions of theorem [ thm : existence of the numerical solution for prk ] are satisfied for the considered runge - kutta methods . as a test problem we considered the solution with the initial condition @xmath480 and @xmath481 (
note that @xmath482 is an equilibrium point ) .
this is a periodic solution with period @xmath483 .
a reference solution was computed by integrating until the time @xmath484 using verner s method with the small time step @xmath485 .
the reference solution is depicted in figure [ fig : reference solution for the lotka - volterra model ] .
using verner s method with the time step @xmath485.,scaledwidth=80.0% ] convergence plots are shown in figure [ fig : convergence plot for the lotka - volterra model ] .
the convergence rates for the gauss and radau iia methods are consistent with the theoretical results presented in section [ sec : runge - kutta methods]we see that the orders of the 2- and 3-stage gauss schemes are reduced .
the 2-stage lobatto iiia - iiib scheme again proves to be inconsistent , and the 3- and 4-stage schemes converge quadratically , just as in section [ sec : kepler s problem ] and section [ sec : point vortices ] .
we performed another series of numerical experiments with the time step @xmath449 to investigate the long time behavior of the considered integrators .
the results are shown in figure [ fig : energy plots for gauss methods for the lotka - volterra model ] and figure [ fig : energy plots for gauss , lobatto and radau methods for the lotka - volterra model ] .
we set @xmath486 in , so that @xmath448 for the considered solution .
the 1- and 3-stage gauss methods again show excellent hamiltonian conservation over a long time interval .
the 2-stage gauss method , however , does not perform equally well the hamiltonian oscillates with an increasing amplitude over time , until the computations finally break down .
the lobatto iiia - iiib methods show similar problems as in section [ sec : kepler s problem ] .
the non - variational radau iia method yields an accurate solution , but demonstrates a steady drift in the hamiltonian . over the time interval @xmath451 $ ] ( _ right column _ ) , with a close - up on the initial interval @xmath487 $ ] shown in the _
left column_.,scaledwidth=90.0% ]
we analyzed a class of degenerate systems described by lagrangians that are linear in velocities , and presented a way to construct appropriate higher - order variational integrators .
we pointed out how the theory underlying variational integration is different from the non - degenerate case and we made a connection with numerical integration of differential - algebraic equations .
we also performed numerical experiments for several example models .
our work can be extended in several ways . in section [ sec : lotka - volterra model ] we presented our numerical results for the lotka - volterra model , which is an example of a system for which the coordinate functions @xmath354 are nonlinear .
the 1- and 3-stage gauss methods performed exceptionally well and preserved the hamiltonian over a very long integration time .
it would be interesting to perform a backward error ( or similar ) analysis to check if this behavior is generic .
if confirmed , our variational approach could provide a new way to construct geometric integrators for a broader class of poisson systems
. it would also be interesting to further consider _ constrained _ systems with lagrangians that are linear in velocities and construct associated higher - order variational integrators .
this would allow to generalize the space - adaptive methods presented in @xcite , @xcite to degenerate field theories , such as the nonlinear schrdinger , kdv or camassa - holm equations .
we would like to thank prof .
ernst hairer and dr .
joris vankerschaver for useful comments and references .
partial funding was provided by nsf grant ccf-1011944 . | in this paper we construct higher - order variational integrators for a class of degenerate systems described by lagrangians that are linear in velocities .
we analyze the geometry underlying such systems and develop the appropriate theory for variational integration .
our main observation is that the evolution takes place on the primary constraint and the hamiltonian equations of motion can be formulated as an index-1 differential - algebraic system .
we also construct variational runge - kutta methods and analyze their properties .
the general properties of runge - kutta methods depend on the velocity part of the lagrangian .
if the velocity part is also linear in the position coordinate , then we show that non - partitioned variational runge - kutta methods are equivalent to integration of the corresponding first - order euler - lagrange equations , which have the form of a poisson system with a constant structure matrix , and the classical properties of the runge - kutta method are retained . if the velocity part is nonlinear in the position coordinate , we observe a reduction of the order of convergence , which is typical of numerical integration of daes .
we verify our results through numerical experiments for various dynamical systems . |
observational and numerical studies suggest that dynamical interactions are frequent occurrences in clusters of galaxies and that these events are the most energetic in the recent history of the universe .
matter is believed to stream along preferred axes ( so - called filaments and sheets ) often accelerating in - falling galaxies to velocities of the order of 10@xmath2 km s@xmath3 .
rich clusters form at the intersection of the filament and sheets , and larger mass congregations , such as superclusters , are often detected as a collection of several clusters and filaments . these dynamical processes cause shocks , turbulence and bulk flows along the filamentary axes resulting in disturbance of the intracluster medium ( icm ) .
such perturbations are frequently detected with x - ray observations which show a variety of features including sharp temperature jumps , distorted x - ray isophotes and regions of enhanced pressure and entropy . at radio wavelengths , cluster mergers
are thought to generate large areas of diffuse synchrotron emission either as a central , unpolarised halo or one or occasionally two peripheral `` radio relics '' .
cluster `` weather '' and bulk flows have also been cited as causes of the bending of the jets in head - tail galaxies ( burns et al .
1998 , johnston - hollitt et al . 2004 ) , although recent results suggest that more local conditions might play a more important role in their formation ( mao et al . 2008 ) .
in addition to the large - area phenomena discussed above , the dynamical history of galaxy clusters should also be imprinted on the constituent galaxies . in particular , increased secondary star formation either observed in the optical ( burns et al .
1994 ; caldwell & rose 1997 ) or through increased low - power radio emission ( owen et al .
1999 ) is one possible signature of the history of a dynamical cluster .
however , the balance between mergers driving gas towards the centre of the galaxies and hence feeding the central agn to induce a starburst ( bekki 1999 ) and an increase in the ram pressure which would strip gas from galaxies and hence curtail the feeding of the agn ( e.g. balogh et al . 1998 ;
fujita et al .
1999 ) still needs to be addressed observationally . to date
, few studies have been conducted with both sufficient sensitivity in the radio to probe the low - powered radio sources , and good optical observations to confirm cluster membership .
the most notable progress to date has been made on sections of the shapley supercluster with several areas analysed in both the radio and optical , including radio observations of a3556 ( venturi et al .
1997 ) , a3571 ( venturi et al .
2002 ) , the a3528 complex ( venturi et al .
2001 ) and the a3558 complex ( venturi et al . 2001 ; miller 2005 ) and its outskirts ( giacintucci et al . 2004 , miller 2005 ) .
in addition to clusters in superclusters , the host radio population for several individual clusters have also been examined including a2255 ( miller & owen , 2005 ) , a2256 ( miller et al .
2003 ) , a2125 ( owen et al .
2005 ) and a2111 ( miller et al .
however , at present , those observational studies that have been undertaken give incomplete results with both instances in which effects are seen on the radio population , and cases where they are not observed . sometimes neighbouring clusters in the same supercluster show markedly different results in terms of radio population .
this may well be due to the radio galaxy population only being affected in certain phases of dynamical activity ( venturi et al .
2002 ) however , the exact point at which a merger might affect the radio population is , as yet , unclear .
nevertheless , large - area observations of superclusters , such as shapley , present ideal datasets with which to investigate the way in which environment affects the galaxy population . in this paper , which will be the first in a series , we present a study of the radio population of a3158 which lies within the kinematic core of the massive horologium - reticulum supercluster ( hrs ) .
the hrs spans around 180 square degrees on the sky ( fleenor et al .
2006 ) , contains upwards of 20 galaxy clusters ( zucca et al .
1993 ) , is comprised of at least two major filaments ( einasto et al .
2003 ) , and is the second largest mass concentration in the local 300 mpc ( hudson et al .
a3158 lies adjacent to the so - called center of the hrs , defined by the double cluster complex a3125/a3128 ( rose et al .
2002 ) . the paper will be laid out as follows : section 2 will discuss a3158 in detail ; section 3 will give details of the radio observations of the cluster ; section 4 will present the radio source population and in sections 5 and 6 we will discuss the resultant radio source counts and radio luminosity functions ; sections 7 & 8 will present the discussion and conclusions . throughout this paper
we will assume h@xmath4=71 km s@xmath3 mpc@xmath3 which at the average redshift of a3158 ( z=0.0597 ) means that 1 arcsecond @xmath5 1.1 kpc . [ optp ]
.known cluster parameters for a3158 .
column 1 and 2 give the parameter and the most recent value from the literature .
the reference is given in column 3 where 1 means havlen & quintana ( 1978 ) , 2 is ebeling et al .
( 1996 ) , 3 is lokas et al .
( 2006 ) and 4 is fleenor et al .
( 2006 ) . [
cols="<,<,<",options="header " , ] [ tab : rlf ]
several studies have been conducted to examine the way in which the cluster environment affects the radio galaxy population either via composite studies such as that undertaken in ledlow and owen ( 1996 ) , `` lo96 '' , or detailed studies of individual clusters or regions .
unfortunately , the results of the individual studies are often conflicting , possibly as a result of low number statistics , however some clear trends are emerging with respect to the three main indicators of the status of the radio population : source counts , radio luminosity functions and colour analysis of faint excesses . of these
, source counts provide an effective way to search for enhancements over the expected background . to date
there have been several source count studies undertaken for parts of the shapley supercluster ( venturi et al .
1997 , 2000 , 2001 , 2002 ) all of which have returned results which are consistent with the predicted background suggesting that the increased optical density of the supercluster environment is not at all reflected in the radio galaxy population .
however , all of these studies have been relativity shallow , typically only examining sources with log p@xmath6 @xmath7 21.8 and are thus , barely probing the regime of starforming galaxies .
the situation with radio luminosity functions and in particular those for early - type galaxies seems to be complex .
venturi et al . ( 1997 ) examined the rlf for a3556 in shapley and found it to be consistent with that found in other environments , hence concluding that , in this case , the dynamical state of the cluster played no role in the likelihood of radio galaxies forming .
however , in a further study venturi et al .
( 2000 ) presented a detailed study of the a3558 complex in the center of the shapley supercluster finding that neither local nor the cluster environment played a significant role in the power distribution of radio sources in the region .
therefore , venturi et al . ( 2000 ) conclude that the probability of forming low - powered radio sources ( log p@xmath8 22.5 ) in shapley is lower than in both the field and ( composite ) cluster environments .
furthermore , they suggest that cluster merging switches off previously active radio galaxies , i. e. , mergers anti - correlate with radio emission in early - type galaxies .
next they examined the a3528 cluster complex in shapley again finding that the rlf was consistent with that found in other environments ( venturi et al .
finally , they considered the radio emission in a3571 ( venturi et al . 2002 ) again finding that it was suppressed with respect to the expected results from lo96 .
giacintucci et al .
( 2004 ) reobserved the outskirts of the a3558 complex and found the rlfs for both early and late - type galaxies to be consistent with the universal " rlfs of ledlow and owen ( 1996 ) and gavazzi & boselli ( 1999 ) respectively , suggesting that the deficit of radio galaxies in the previous observation was entirely due to the rich cluster a3558 itself .
venturi et al .
( 2001 , 2002 ) attribute the difference in the rlfs to different stages in the merger history of the regions playing a critical role in the determination of radio emission in cluster galaxies . in the case of early mergers
suggested for a3556 and the a3528 complex they suggested that the radio galaxy population has not yet been affected , whereas for the late - stage merger in a3571 and a3558 , in the very core of shapley , the radio galaxy population has been switched off by the merging .
given that cluster mergers are particularly long lived ( @xmath5 gyr ) compared to the lifetimes of radio sources which are typically only 10@xmath9 -10@xmath10 years it seems likely that the radio population could give vital clues as to the exact point in a cluster s dynamical history , thus providing to be an extremely important evolutionary indicator .
this is particularly true if the radio emission is considered together with other signposts of cluster dynamics such as x - ray observations . despite not being evident via an excess in source counts , giacintucci et al .
( 2004 ) found a population of low powered ( log p@xmath11 22 ) galaxies on the outskirts of the a3558 complex , many of which showed a blue excess which they suggested are likely to be starbursts , although they noted that their observations provided only weak evidence that the fraction of radio - emitting galaxies was higher than in non - merging environments .
miller ( 2005 ) conducted a deeper observations of the a3558 complex over 7 deg@xmath12 and , on the basis of colour and radio luminosity , confirmed the presence of an increased starforming population associated with the recent merging of sc 1329 - 313 and a3562 with the rest of the supercluster .
an excess of low - powered blue galaxies was also observed in the merging clusters a2125 ( owen et al .
2005 ) , a2255 ( miller et al .
2003 ) and in a2111 ( miller et al .
2006 ) which despite lacking statistical evidence of substructure at right angles to the plane of the sky , is likely to be dynamically active . in both cases
the authors attribute the excess of dusty starburst galaxies to the merger activity . in all cases
an increase in low - power , blue galaxies either confirmed via spectroscopic analysis as starburst ( owen et al .
2005 ) or assumed through colour analysis ( giacintucci et al .
2004 ) has been associated with recently merging clusters . in the case of a3158
we see both an excess of low - powered radio galaxies ( log p@xmath1 22 ) in the source counts and a suppression of the early - type rlf with respect to universal indicators .
thus , we seem to be seeing both a suppression of agn as observed in the late mergers in a3571 and a3558 ( venturi et al . 2000 and 2002 ) and an increase in the low - powered , blue galaxies , likely to be starbursts , as seen in the mergers for a2125 , a2255 , a2111 ( owen et al .
2005 , miller et al .
2003 & miller et al .
2006 respectively ) and the outskirts of a3558 ( giacintucci et al .
in addition , the axis of the starforming galaxy candidates is elongated and aligned in the direction which connects a3158 to the more massive a3125/a3128 complex which lies at the southern centre of the hrs .
this is similar to the distribution of star - forming galaxies seen in a2125 .
additionally the lack of ht galaxies is similar to both a3571 and a3558 and was used by venturi et al .
( 2001 ) to argue for these clusters being in the later stages of a dynamical encounter .
we present the the first of a series of papers investigating radio emission in the horologium - reticulum supercluster . in this paper
we present radio imaging at 1.4 and 2.5 ghz of the cluster a3158 , from which we detect 109 radio sources of which 64 are found to have optical counterparts . by considering a reduced subsample of 87 sources within a @xmath13 radius , over which the primary beam attenuation of the atca is negligible , we are able to investigate the statistical properties of radio emission in the cluster .
our results can be summarised as follows .
\(1 ) the radio source counts in a3158 are unlikely to be consistent with the background source counts and in particular there appears to be an excess of galaxies with s@xmath14 3.96 mjy ; a ks test of the entire sample against the background only returns a @xmath53% - 8% chance that the distributions are drawn from the same population , while a ks test over the sample with s@xmath14 3.96 mjy gives a much higher chance ( @xmath560% ) that the distributions are drawn from the same population suggesting that the low - powered excess is the significant difference .
\(2 ) the percentage of radio galaxies associated with the cluster as determined by spectroscopic redshifts is highly correlated to radio flux / power which is consistent with the excess source counts being the result of low - powered radio galaxies associated with a3158 ; in addition colours of these galaxies suggest they are blue which is likely to indicate starformation as a driving mechanism .
\(3 ) the alignment of the candidate starforming galaxies associated with the increased source counts follows the bridge of galaxies connecting a3158 to the a3125/a3128 complex suggestive of merger - induced formation .
\(4 ) the radio luminosity function for early - type galaxies is significantly lower than for other composite cluster environments and the field ; this result is similar to the results for the late merger in a3558 obtained by venturi et al .
( 2000 ) .
thus , if we can use the radio galaxy population as an indicator of merger history it is likely that a3158 is observed in a late merger state with a significant excess of low - powered , blue galaxies aligned along the axis connecting this cluster to the a3125/a3128 complex .
these galaxies are likely the result of merger - induced starformation .
we thank the anonymous referee for their careful reading of the manuscript and insightful comments . mj .-
h . acknowledges the support of irgs grant j0015125 & j0016583 administered by the university of tasmania .
m.s . was supported in this work via a university of tasmania vacational scholarship in astrophysics .
j.a.g acknowledges the support of a dave warren scholarship administered by the university of tasmania .
m.b . was supported by a university of tasmania honours scholarship .
this research was supported by the dfg cluster of excellence `` origin and structure of the universe '' .
a portion of this work was supported by nsf grants ast-9900720 and ast-0406443 to the university of north carolina - chapel hill .
the australia telescope compact array is part of the australia telescope which is funded by the commonwealth of australia for operation as a national facility managed by csiro .
this research has made use of 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 .
the digitized sky surveys were produced at the space telescope science institute under u.s .
government grant nag w-2166 .
the images of these surveys are based on photographic data obtained using the oschin schmidt telescope on palomar mountain and the uk schmidt telescope .
the plates were processed into the present compressed digital form with the permission of these institutions .
99 abell , g. o. , corwin , jr . , h. g. , olowin , r. p. , 1989 ,
apjs , 70 , 1 balogh , m. l. , schade , d. , morris , s. l. , yee , h. k. c. , carlberg , r. g. , ellingson , e. , 1998 , apj , 504 , l75 bekki , k. , 1999 , apj , 510 , l15 bardelli , s. , zucca , e. , zamorani , g. , vettolani , g. , scatsmella , r. , 1998 , mnras , 296 , 599 burns , j. o. , roettiger , k. , ledlow , m. , klypin , a. , 1994 , apj , 427 , l87 burns , j. o. , 1998 , science , 280 , 400 cadwell , n. , rose , j. a. , 1997 , aj , 113 , 492 condon , j. j. , cotton , w. d. , broderick , j. j. , 2002 , aj , 124 , 675 dalton , g. b. , efstathiou , g. , maddox , s. j. , sutherland , w. j. , 1992 , apj , 390 , l1 dalton , g. b. , croft . , r. a. c. , efstathiou , g. , sutherland , w. j. , maddox , s. j. , davis , m. , 1994 , mnras , 271 , l47 de vaucouleurs , g. , de vaucouleurs , a. , corwin , jr , h. g. , buta , r. j. , paturel , g. , fouque , p. , 1991 , ` third reference catalogue of bright galaxies ' , vol 13 , xii , 20009 , spinger
verlag berlin .
dupke , r. a. , bregman , j. n , 2005 , apjs , 161 , 224 ebeling , h. , voges , w. , bhringer , h. , edge , a. c. , huchra , j. p. , briel , u. g. , 1996 , mnras , 281 , 799 einasto , m. , jaaniste , j. , einasto , j. , heinmki , p. , mller , v. , tucker , d. l. , 2003 , aap , 405 , 821 fleenor , m. c. , rose , j. a. , christiansen , w. a. , hunstead , r. w. , johnston - hollitt , m. , drinkwater , m. j. , 2004 , in iau colloq .
195 : outskirts of galaxy clusters : intense life in the suburbs , ed .
diaferio , a. , 89 fleenor , m. c. , rose , j. a. , christiansen , w. a. , hunstead , r. w. , johnston - hollitt , m. , drinkwater , m. j. , saunders , w. , 2005 , aj , 130 , 957 fleenor , m. c. , 2006 , phd thesis , university of north carolina , chapel hill .
fleenor , m. c. , rose , j. a. , christiansen , w. a. , johnston - hollitt , m. , hunstead , r. w. , drinkwater , m. j. , saunders , w. , 2006 , aj , 131 , 1280 fujita , y. , tazikawa , m. , nagashima , m. , enoki , m. , 1999 , pasj , 51 , l1 fukugita , m. , shimasaki , k. , ichikawa , t. , 1995 , pasp , 107 , 945 giacintucci , s. , venturi , t. , bardelli , s. , dallacasa , d. , zucca , e. , 2004 , a&a , 419 , 71 havlen , r. j. , quintana , h. , 1978 , 220 , 14 den hartog , r. , katgert , p. , 1996 , 279 , 349 hudson , m. j. , smith , r. j. , lucey , j. r. , schlegel , d. j. , davies , r. l. , 1999 , apjl , 512 , l795 johnston - hollitt , m. , fleenor , m. , rose , j. , christiansen , w. , hunstead , r. w. , 2004 , in iau colloq . 195 : outskirts of galaxy clusters : intense life in the suburbs , ed .
diaferio , a. , 423 katgert , p. , mazure , a. , den hartog , r. , adami , c. , biviano , a. , perea , j. , 1998 , a&as , 129 , 399 kennicutt , r. c. , jr , ajs , 79 , 255 ku , w. h .-
m . , abramopulos , f. , nulsen , p. e. j. , fabian , a. c. , stewart , g. c. , chincarini , g. l. , tarenghi , m , 1983 , mnras , 203 , 253 lambas , d. g. , nicotra , m. , muriel , h. , ruiz , l. , 1990 , aj , 100 , 1006 ledlow , m. j. , owen , f. n. , 1996 , aj , 112 , 9 lokas , e. l. , wojtak , r. , gottlber , s. , mamon , g. a. , prada , f. , 2006 , mnras , 367 , 1463 loveday , j. , peterson , b. a. , maddox , s. j. , efstathiou , g. , 1996 , apjs , 107 , 201 lucey , j. r. , dickens , r. j. , mitchell , r. j. , dawe , j. a. , 1983 , mnras , 203 , 545 mao , m. y. , johnston - hollitt , m. , stevens , j. b. , wotherspoon , s. j. , mnras , 2008 , ( submitted ) melnick , j. , quintana , h. , 1975 , apj , 198 , l97 miller , n. a. , owen , f. n. , 2002 , aj , 124 , 2453 miller , n. a. , 2005 , aj , 130 , 2541 miller , n. a. , owen , f. n. , 2003 , aj , 125 , 2427 miller , n. a. , oegerle , w. r. , hill , j. m. , 2006 , aj , 131 , 2426 mauch , t. , sadler , e. m. , 2007 , mnras , 375 , 931 ohata , y. , kumai , y. , wantanabe , m. , furuzawa , a. , akimoto , f. , tawara , y. , sato , s. , yamashita , k. , arai , k. , shiratori , y. , miyoshi , s. , mazure , a. , 2001 , proc .
series , 251 owen , f. n. , ledlow , m. j. , keel , w. c. , morrison , g. e. , 1999 , aj , 118 , 633 owen , f. n. , ledlow , m. j. , keel , w. c. , wang , q. d. , morrison , g. e. , 2005 , aj , 129 , 31 prandoni , i. , gregorini , l. , parma , p. , de ruiter , h. r. , vettolani , g. , wieringa , m. h. , ekers , r. d. , a&a , 2001 , 365 , 392 quintana , h. , havlen , r. j. , 1979 , a&a , 79 , 70 rose , j. a. , gaba , a. e. , christiansen , w. a. , davis , d. s. , caldwell , n. , hunstead , r. w. , johnston - hollitt , m. , 2002 , aj , 123 , 1216 sault r.j .
, teuben p.j . ,
wright m.c.h . , 1995 , in astronomical data analysis software and systems iv , ed .
r. shaw , h.e .
payne , j.j.e .
hayes , asp conf .
, 77 , 433 .
srsic , j. l. , 1974 , ap&ss , 28 , 365 smith , r. j. , hudson , m. j. , nelan , j. e. , moore , s. a. w. , quinney , s. j. , wegner , g. a. , lucey , j. r. , davies , r. l. , malecki , j. j. , schade , d. , suntzeff , n. b. , 2004 , aj , 128 , 1558 venturi , t. , bardelli , s. , morganti , r. , hunstead , r. w. , 1997 , mnras , 285 , 898 venturi , t. , bardelli , s. , morganti , r. , hunstead , r. w. , 2000 , mnras , 314 , 594 venturi , t. , bardelli , s. , zambelli , g. , morganti , r. , hunstead , r. w. , 2001 , mnras , 324 , 1131 venturi , t. , bardelli , s. , zagaria , m. , prandoni , i. , morganti , r. , 2004 , a&a , 385 , 39 wieringa , m. h. , kesteven , m. j.,1992 , atca technical memo series , 39.2/010 zucca , e. , zamorani , g. , scaramella , r. , vettolani , g. , 1993 , apj , 407 , 470 | we present 1.4 and 2.5 ghz australia telescope compact array ( atca ) observations of the galaxy cluster a3158 ( z=0.0597 ) which is located within the central part of the horologium - reticulum supercluster ( hrs ) .
spectroscopic data for the central part of the hrs suggest that a3158 is in a dynamically important position within the supercluster and that it is moving toward the double cluster system a3125/a3128 which marks the centre of the hrs .
a total of 110 radio galaxies are detected in a 35 arcminute radius about the cluster at 1.4 ghz , of which 30 are also detected at 2.5ghz .
we examine the source counts and compute the radio luminosity function ( rlf ) at 1.4 ghz from the subset of 88 sources found within the full - width half - power area of the atca beam .
comparison of the source counts in the area over the background , as computed by prandoni et al .
( 2001 ) , shows some evidence of an excess of galaxies with l@xmath0 .
this result seems to indicate a star forming population and is a result similar to that found recently by owen et al .
( 2005 ) for the merging cluster a2125 .
in addition we find that the radio luminosity function for early - type galaxies ( e and s0 ) below log p@xmath1 22.5 is lower than that found for a composite cluster environment ( ledlow & owen , 1996 ) but is similar to the early - type rlf for clusters in the centre of the shapley super cluster ( venturi et al .
2000 ) which are believed to be in the latter stages of merging .
this result implies that the cores of superclusters are environments where radio emission , particularly resultant from agn , is suppressed in the later stages of merging .
thus , radio observations of clusters might be sensitive indicators of the precise merger stage of the cluster but more observational evidence is still required to establish this trend .
[ firstpage ] galaxies : clusters : general - galaxies : clusters : individual : a3158 - radio continuum : galaxies . |
the effective field theory ( eft ) approach provides a systematic and model - independent framework to analyse the hadronic parity - violating ( pv ) effects in few - body system @xcite .
we use the pionless effective field theory ( eft(@xmath2 ) ) for the pv process which is at the energies well below the pion production threshold .
recently , the leading - order ( lo ) pv nucleon - nucleon ( @xmath3 ) lagrangian has been expressed in a variety of bases @xcite .
the lo pv lagrangian in the partial - wave basis is written in terms of five independent unknown coupling constants corresponding to five different s - p wave combinations .
this lagrangian has been applied for neutron - neutron , proton - proton and neutron - proton longitudinal asymmetries without dibaryon fields in eft(@xmath2 ) @xcite .
the calculation of @xmath4 in @xmath5 @xmath6 at threshold was introduced with the pv dibaryon - nucleon - nucleon ( @xmath7 ) lagrangian in eft(@xmath2 ) .
the pv @xmath7 lagrangian is constructed in terms of three weak coupling constants .
in fact , parts of the pv lagrangian in dibaryon formalism are given by this lagrangian structure .
result for @xmath4 was obtained in terms of two unknown weak @xmath7 coupling constants for the parity mixing in the spin - singlet and spin - triplet states @xcite .
the photon asymmetry @xmath8 and circular polarization @xmath4 in @xmath5 @xmath6 process were also derived using the pv @xmath3 lagrangian based on the partial - wave basis with and without dibaryon fields .
eft(@xmath2 ) with dibaryon fields has proven to be convenient for few - nucleon systems at low energies @xcite . in the three - body sector , a systematic analysis of the parity - conserving ( pc ) @xmath0 @xmath1 process at threshold
was carried out using the eft(@xmath2 ) @xcite . in the present work ,
we concentrate on the pv @xmath0 @xmath1 amplitude .
we will make predictions for the pv observable in @xmath0 @xmath1 process with the pv lagrangian based on the five independent lecs in the dibaryon eft(@xmath2 ) .
the analysis of the pv three - nucleon interaction ( 3ni ) and calculation of neutron - proton and neutron - deuteron spin rotation have been performed in the eft(@xmath2 ) framework .
the results show that no pv 3ni enters at lo or next - to - leading order ( nlo ) in the nucleon - deuteron ( @xmath9 ) system @xcite . according to these results
we adopt the pv lagrangian for the two - nucleon interaction in the present calculation .
theoretical calculations of the pv photon asymmetry and circular polarization of the outgoing photon in @xmath0 @xmath1 process were previously performed with the potential model .
these analyses were based on desplanques , donoghue , and holstein ( ddh ) meson exchange potential .
the calculated results were @xmath10 and @xmath11 for the two super - soft - core ( ssc ) and reid - soft - core ( rsc ) potentials , respectively .
the ddh calculation of @xmath4 was also given for the two ssc and rsc potentials , @xmath12 and @xmath13 , respectively @xcite .
recently , the pv effects in @xmath0 @xmath1 were calculated for ddh and eft weak interaction potentials @xcite .
we compare these @xmath4 values with our results in sect.[result ] .
finally , the experimental observation for the pv photon asymmetry in @xmath0 @xmath1 process was measured at ill @xcite and the result was presented as @xmath14 . the experimental data for @xmath4 in this process has not been reported to this point at the thermal energy . our paper is organized as follows . in sect .
[ pc sector ] we present briefly the pc formalism for the @xmath9 scattering and neutron - deuteron radiative capture .
the pv @xmath9 scattering and pv @xmath0 @xmath1 processes are explained in sect .
[ pv sector ] .
we obtain the pv amplitude of @xmath0 @xmath1 process in sect .
[ pv nd radiative capture sector ] with the calculation of the appropriate diagrams .
the numerical methods for the evaluation of the pc and pv amplitudes of @xmath0 @xmath1 process are explained in sect .
[ numerical ] . in sect .
[ result ] the results of @xmath4 in @xmath0 @xmath1 are estimated and compared with the ddh framework .
we summarize the paper and discuss extension of the investigation to other observable such as the pv photon asymmetry and other few - body systems in sect .
[ conclusion ] .
in the study of the weak interaction effects , we calculate the observables which depend on the pv interactions . for this purpose the @xmath4 in @xmath0@xmath1 process is chosen . to this aim
we need the pc amplitudes for both @xmath9 scattering and @xmath0@xmath1 processes .
we briefly describe below the procedure to obtain these two pc amplitudes at the leading order .
we proceed by introducing the lagrangian of two- and three - nucleon systems in z - parametrization @xcite @xmath15 where @xmath16d^{a}_{s } \nonumber \\ + d^{i^\dag}_{t}\big[\delta_t - c_{0t}\big(i\partial_0+\frac{\nabla^2}{4m_n}+\frac{\gamma^2_t}{m_n}\big)\big]d^{i}_{t}\;\;\ : \nonumber\\ -y\big(d^{a^\dag}_{s}(n^\dag p^a n)+h.c.\big)\qquad\qquad\quad\;\;\ ; \nonumber \\
-y\big(d^{i^\dag}_{t}(n^\dag p^i n)+h.c.\big)+ ... \,,\qquad\qquad\end{aligned}\ ] ] @xmath17\quad\;\ ; \nonumber \\+n^\dag\big(d^{a}_{s}\tau_a\big)^\dag\big(d^{a}_{s}\tau_a\big)n\bigg)+ ... \,,\quad\;\;\end{aligned}\ ] ] the faddeev equation for the @xmath9 scattering .
single solid line denotes a nucleon .
double line is propagator of the two intermediate auxiliary fields @xmath18 and @xmath19 , denoted by @xmath20 . @xmath21 and @xmath22 are the propagator of exchanged nucleon and three - body interaction at lo , respectively.,width=336 ] where @xmath23 is the nucleon iso - doublet and the auxiliary fields @xmath24 and @xmath25 carry the quantum number of @xmath26 di - nucleon and the deuteron , respectively .
the projectors @xmath27 and @xmath28 are defined by @xmath29 and @xmath30 where @xmath31 and @xmath32 are iso - triplet and vector indices , respectively .
@xmath33(@xmath34 ) are isospin ( spin ) pauli matrices .
@xmath35 is the nucleon mass and the three - nucleon interaction at the leading order is @xmath36 with cut - off @xmath37 .
the parameters @xmath38 are the mass differences between the dibaryons and two nucleons . according to the naive dimensional analysis : the parameters @xmath39 , with mass - dimension 1 , are entered at lo ; dimensionless parameters @xmath40 first appear at nlo ; since they are expressed with two powers of momentum , @xmath41 , where @xmath42 is small momentum and appears in @xmath43 or @xmath44 as an expansion parameter ( @xmath45 the pion mass , @xmath46 a symmetry breaking scale ) in eft(@xmath2 ) .
the parameters of the lagrangians in eqs.([eq:2 ] ) and ( [ eq:3 ] ) are fixed using z - parametrization @xcite .
@xmath47 coupling constant is chosen as @xmath48 and the lo parameters @xmath38 are obtained from the poles of the @xmath3 s - wave scattering amplitude at @xmath49 .
the physical observables are cutoff independent , so in the three - body systems at the leading order all - dependence on the cutoff as @xmath50 can be made to vanish by introducing @xmath22 .
we know that the @xmath36 is of order @xmath51 @xcite .
neutron - deuteron scattering is shown schematically in fig.[fig : nd scattering ] .
we denote the single straight line as a nucleon with the propagator @xmath52 .
the double line represents the dibaryon field with propagators for the singlet and triplet cases .
they are given by @xmath53 where @xmath54 where @xmath55 , @xmath56 is the scattering length in @xmath26 state and @xmath57 is the binding momentum of the deuteron .
we work in the center of mass frame for the @xmath9 scattering with @xmath58 denotes the initial ( on - shell ) relative momentum of the deuteron and the third nucleon , and @xmath59 is the final ( off - shell ) momentum .
we use notation suggested by griesshammer in @xcite . for scattering in the quartet ( @xmath60=@xmath61 ) channel all spins
are aligned and there is no three - body interaction in this channel because of the pauli principle forbids the three nucleons to be at the same point in space . in this channel , the initial and final dibaryon fields have to be in the triplet channel .
so , in this channel we have only the @xmath62 transition and the corresponding amplitude in the cluster - configuration space can be written as @xmath63 for the @xmath9 scattering in the doublet ( @xmath60=@xmath64 ) channel , the parity - conserving amplitudes in cluster - configuration space are given by @xmath65\qquad\qquad\qquad\qquad \;\nonumber\\-\frac{1}{\pi}\int^\lambda_0dqq^2\left[\mathcal{k}^{pc}_{(l)}(e , q , p)\left ( \begin{array}{cc } 1 & -3 \\ -3 & 1 \\ \end{array } \right)+\delta_{l0}\mathcal{h}(e,\lambda)\left ( \begin{array}{cc } 1 & -1 \\
-1 & 1 \\ \end{array } \right)\right ] \qquad\qquad\;\nonumber\\ \times\mathcal{d}^ { lo}(e-\frac{q^2}{2m_n},q)\left ( \begin{array}{cc } t^{pc,(l)}_{d_{d_tn\rightarrow d_tn}}(e;k , q ) & t^{pc,(l)}_{d_{d_sn\rightarrow d_tn}}(e;k , q ) \\ t^{pc,(l)}_{d_{d_tn\rightarrow d_sn}}(e;k , q ) & t^{pc,(l)}_{d_{d_sn\rightarrow d_sn}}(e;k , q ) \\
\end{array } \right),\qquad\qquad\;\:\end{aligned}\ ] ] where @xmath66 denotes the @xmath67 transition amplitude ( @xmath68 or @xmath69 ) in the doublet channel .
the propagator of the exchanged nucleon , @xmath70 , is @xmath71 where @xmath72 denotes the @xmath73-th legendre polynomial of the first kind and @xmath74 indicates the angle between @xmath58 and @xmath59 vectors .
@xmath75 specifies the three - body force that renormalizes the amplitude and it is introduced for the s - wave ( @xmath76 ) .
@xmath77 is the total non - relativistic energy .
the leading - order diagrams which contribute to pc ( m1 ) amplitude of @xmath0 @xmath1 .
the small circles represent magnetic photon - nucleon - nucleon interaction .
wavy line shows a photon .
dashed oval and dashed half - oval indicate insertion of @xmath9 scattering amplitude at lo from fig.[fig : nd scattering ] and the formation of triton ( the normalized triton wave function ) .
remaining notation as fig.[fig : nd scattering].,width=245 ] the dominated m1 amplitude of pc @xmath0 @xmath1 process receives the contribution from the neutron and proton magnetic moment operators coupling to the magnetic field . at the leading order , the magnetic m1 interaction of the photon with the single nucleon
is introduced by the lagrangian @xmath78 where @xmath79 and @xmath80 are the isoscalar and isovector nucleon magnetic moments in the nuclear magnetons , respectively .
@xmath81 and @xmath82 denote the proton and neutron magnetic moments , respectively .
@xmath83 is the electric charge and @xmath84 is magnetic field .
the diagrams of pc @xmath0 @xmath1 at lo are shown in fig.[fig : pc nd capture ] .
the dashed oval is the @xmath9 scattering calculated in sect .
[ pc nd scattering sector ] .
the dashed half - oval is the normalized triton wave function .
we note that for obtaining the normalized triton wave function , one should solve the homogeneous part of eq.([eq:5 ] ) with application of @xmath85 where @xmath86 is the binding energy of the triton .
the wave function is then normalized using the nucleon and dibaryon propagators as briefly explained in [ appendix a ] .
the homogeneous part of eq.([eq:5 ] ) can be written as @xmath87 \nonumber\\ \times\mathcal{d}^ { lo}(-b_t-\frac{q^2}{2m_n},q)\,t^{^3h}(q),\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad\end{aligned}\ ] ] where @xmath88 .
generally , @xmath89 denotes the contribution of the @xmath90 transition ( @xmath68 or @xmath69 ) for making the triton . for pc sector at the lowest order , we have two dominated m1 transitions , to which @xmath91 and @xmath92 belong , @xmath93 and @xmath94 .
the following parametrization of the corresponding contributions to the matrix element are , @xmath95 with @xmath23 , @xmath69 , @xmath96 , @xmath97 and @xmath98 are the 2-component spinors of the initial nucleon field , the final @xmath92 ( or @xmath91 ) field , the 3-vector polarization of the produced photon , the 3-vector polarization of the deuteron and the unit vector along the 3-momentum of the photons , respectively . for both possible magnetic transitions with @xmath99 ( @xmath100 amplitude ) and @xmath101 ( @xmath102 amplitude )
we can write @xmath103\,,\ ! \nonumber\\ \mathcal{w}^{pc}(^4s_{\frac{3}{2}})\big[t^\dag\big(i\vec{\varepsilon}_d\cdot\vec{\varepsilon}^\ast_\gamma\times\vec{\tilde{q}}+\vec{\sigma}\times\vec{\varepsilon}_d\cdot\vec{\varepsilon}^\ast_\gamma\times\vec{\tilde{q}}\big)n\big]\,.\end{aligned}\ ] ] the pc amplitude of @xmath0 @xmath1 at the thermal energy is calculated and then used in the evaluation of the @xmath4 in sect.[result ] .
the pv two - body transitions at the lowest order are @xmath104 @xmath105 , @xmath106 @xmath107 and @xmath108 @xmath109 .
the leading - order pv lagrangian with dibaryon formalism is given by @xmath110\quad \nonumber \\
+ h.c.+ ... \,.\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad\qquad\end{aligned}\ ] ] in the above equation , @xmath111 denotes the weak @xmath7 coupling constant for the pv two - body transition between @xmath112 and @xmath113 partial waves .
@xmath114 represents the isospin change in the pv vertex and @xmath115=diag(1,1,-2 ) is a diagonal matrix in isospin space .
the three - point in the last line represents all transitions that will be neglected at the very low energy . for the three - nucleon systems , at the lowest order , there are four possible transitions which mix s- and p - waves and conserve total angular momentum , @xmath116 a systematic analysis by griesshammer _ et al .
_ @xcite show that no parity - violating three - nucleon interaction ( pv 3ni ) is participated in the nucleon - deuteron system at the leading and next - to - leading orders in eft(@xmath2 ) .
so , in the three - body systems , since all amplitudes converge at lo and nlo , no pv 3ni are needed for the renormalization .
thus , we use the lagrangian in eq.([eq:8 ] ) to obtain the lo pv @xmath9 scattering amplitude for all possible channels in eq.([eq:35 ] ) .
the feynman diagrams that contribute to the pv @xmath9 scattering amplitude are shown in fig.[fig2 ] @xcite . in the cluster - configuration space ,
the contribution of these diagrams in fig.[fig2 ] can be written after doing integration over energy and solid angle as @xmath117+\quad \nonumber \\
+ \frac{1}{\pi^3}\int^\lambda_0dq_1q^2_1\;\int^\lambda_0dq_2q^2_2\;\bigg[t^{lo , pc}(y;e , q_2,p)\mathcal{d}^{lo}(e-\frac{q^2_2}{2m_n},q_2)\qquad\qquad \nonumber \\
\times \:\mathcal{k}^{pv}_{up}(x\rightarrow y;e , q_1,q_2)\;\ , \mathcal{d}^{lo}(e-\frac{q_1 ^ 2}{2m_n},q_1)\;t^{lo , pc}(x;e , k , q_1)\bigg],\end{aligned}\ ] ] where @xmath118 is @xmath119 matrix for lo pc @xmath9 scattering amplitudes , eqs.([eq:05 ] ) , ( [ eq:5 ] ) , depending on the incoming ( outgoing ) @xmath120 ( @xmath121 ) partial wave . @xmath122 and @xmath58 @xmath123 are the center - of - mass energy and the incoming ( outgoing ) momentum , respectively .
the pv kernel @xmath124 for all transitions with total angular momentum @xmath125 is given by @xcite ( see eq.([eq:35 ] ) for all possible @xmath126 ) @xmath127\left ( \begin{array}{cc } \mathcal{s}_1 & -\mathcal{t }
\\ \mathcal{s}_1 & -\mathcal{t } \\ \end{array } \right ) , \nonumber \\
\mathcal{k}^{pv}_{up}(^2p_{\frac{1}{2}}\rightarrow\,^2s_{\frac{1}{2}};e , k , p)=\frac{ym_n}{6\sqrt{2}\pi kp}\,\big[2pq_1(-\varepsilon)+kq_0(-\varepsilon)\big]\left ( \begin{array}{cc } \mathcal{s}_1 & -\mathcal{t }
\\ \mathcal{s}_1 & -\mathcal{t } \\ \end{array } \right ) , \nonumber \\
\mathcal{k}^{pv}_{up}(^2s_{\frac{1}{2}}\rightarrow\,^4p_{\frac{1}{2}};e , k , p)= \frac{ym_n}{3\pi kp}\,\big[2pq_0(-\varepsilon)+kq_1(-\varepsilon)\big]\left ( \begin{array}{cc } \frac{\mathcal{s}_1-\mathcal{s}_2}{3 } & \mathcal{t } \\ 0 & 0 \\ \end{array } \right ) , \nonumber \\ \nonumber \\
\mathcal{k}^{pv}_{up}(^4p_{\frac{1}{2}}\rightarrow\,^2s_{\frac{1}{2}};e , k , p)=\frac{ym_n}{6\pi kp}\,\big[2pq_1(-\varepsilon)+kq_0(-\varepsilon)\big]\left ( \begin{array}{cc } \mathcal{s}_2 & 0 \\ \mathcal{s}_2 & 0 \\ \end{array } \right),\qquad\:\end{aligned}\ ] ] and with @xmath128 by @xmath129\left ( \begin{array}{cc } \mathcal{s}_2 & 0 \\ \mathcal{s}_2 & 0 \\ \end{array } \right)q^r_s , \qquad\;\nonumber \\ \nonumber \\ \mathcal{k}^{pv}_{up}(^2p_{\frac{3}{2}}\rightarrow\,^4s_{\frac{3}{2}};e , k , p)=\frac{ym_n}{3\sqrt{2}\pi kp}\,\big[2pq_1(-\varepsilon)+kq_0(-\varepsilon)\big]\left ( \begin{array}{cc } \frac{\mathcal{s}_1-\mathcal{s}_2}{3 } & \mathcal{t } \\ 0 & 0 \\ \end{array } \right)q^r_s , \;\:\:\nonumber \\ \nonumber \\
\mathcal{k}^{pv}_{up}(^4s_{\frac{3}{2}}\rightarrow\,^4p_{\frac{3}{2}};e , k , p)=-\frac{\sqrt{10}ym_n}{6\pi kp}\,\big[2pq_0(-\varepsilon)+kq_1(-\varepsilon)\big]\left ( \begin{array}{cc } \frac{\mathcal{s}_1-\mathcal{s}_2}{3 } & 0 \\ 0 & 0 \\ \end{array } \right)q^r_s , \nonumber \\ \nonumber \\ \mathcal{k}^{pv}_{up}(^4p_{\frac{3}{2}}\rightarrow\,^4p_{\frac{3}{2}};e , k , p)=-\frac{\sqrt{10}ym_n}{6\pi kp}\,\big[2pq_1(-\varepsilon)+kq_0(-\varepsilon)\big]\left ( \begin{array}{cc } \frac{\mathcal{s}_1-\mathcal{s}_2}{3 } & 0 \\ 0 & 0 \\ \end{array } \right)q^r_s,\end{aligned}\ ] ] where @xmath130 and the @xmath131 is the spin quartet projector .
the index @xmath132 @xmath133 denotes the spin of the outgoing ( incoming ) nucleon and @xmath134 @xmath135 is the spin component of the outgoing ( incoming ) @xmath136 dibaryon .
@xmath137 is the @xmath73-th legendre polynominals of the second kind with the complex argument .
the `` up '' subscript represents the pv vertex in the upper side of the diagrams .
the @xmath138 , @xmath139 and @xmath140 are independent linear combinations of the pv coupling constants , @xmath111 , @xmath141 where the value of the @xmath142 is + 1 and -1 for the scattering on the proton and neutron , respectively .
we note that the eq.([eq:035 ] ) is not a faddeev equation .
however the faddeev equations in the eqs.([eq:05 ] ) , ( [ eq:5 ] ) should be solved numerically for the pc @xmath9 scattering amplitude and then replace the results in eq.([eq:035 ] ) .
we solve eq.([eq:035 ] ) numerically .
the calculation of the time - reversed contributions of the diagrams in fig.[fig2 ] is the same as explained above .
we must only substitute the @xmath143 in eq.([eq:035 ] ) with @xmath144 to obtain the time - reversed contribution of the diagrams in fig.[fig2 ] .
the @xmath145 @xmath146 can be written as @xmath147^\dag\!\!.\end{aligned}\ ] ] finally , after multiplication of the wave function normalization factor the physical amplitude of the neutron - deuteron scattering in the on - shell point ( @xmath148 ) is @xmath149 where @xmath150 is the sum of both @xmath151 and @xmath152 and @xmath153 is the lo deuteron wave function normalization factor in z - parametrization . in this section ,
we concentrate on the pv amplitude of @xmath0 @xmath1 process .
for this purpose , we need to obtain the electromagnetic ( em ) matrix elements that contribute at the thermal energy ( @xmath154 ) near threshold . in the low - energy regime
, we work with the lagrangian introduced in eq.([eq:8 ] ) for the pv interaction . at threshold
, e1 transition contributes a dominate part in the pv neutron - deuteron radiative capture process .
obviously , as we move above the threshold we must also add the parity - violating m1 transition to the leading e1 amplitude . in the present calculations , we introduce the diagrams that contribute to the e1 pv neutron - deuteron radiative capture process .
these diagrams are shown schematically in figs.[fig3 ] and [ fig4 ] .
we identify the diagrams in fig.[fig3 ] as the set ( i ) and those of fig.[fig4 ] as the set ( ii ) .
the dashed rectangular with dashed line around it in fig.[fig3 ] denotes the pv @xmath9 scattering amplitude corresponding to the diagrams @xmath155 , @xmath156 and time - reversed contribution of @xmath155 , @xmath157 shown in fig.[fig2 ] .
the dashed rectangular with solid line around it in fig.[fig4 ] is the pv @xmath9 scattering amplitude depicted in fig.[fig2 ] . in figs.[fig3 ] and [ fig4 ] dashed oval and dashed half - oval are the same as before .
we consider only the one - body current for e1 interaction and we employ the `` convection '' nucleon current for the photon - nucleon - nucleon ( @xmath158 ) vertex that can be written generally by @xcite @xmath159\cdot\:\vec{\varepsilon_\gamma^\ast}\,,\end{aligned}\ ] ] where @xmath160 and @xmath161 are momenta for the incoming and outgoing nucleons . @xmath162 and @xmath163 are the @xmath164-component of the isospin operator and the position of the @xmath165-nucleon , respectively . for the single nucleon , it is reduced to @xmath166 the pv photon - dibaryon - nucleon - nucleon ( @xmath167 ) vertex in fig.[fig3 ] is given by the minimal substitution of @xmath168 in the lagrangian of eq.([eq:8 ] ) with @xmath169 as an external field .
we have both e1 and pv interactions together , so it is obvious that the neutron - deuteron system can change from the initial channel @xmath170 or @xmath171 to the final channel @xmath170 or @xmath171 at the thermal energy . thus , there are four possible transitions , @xmath172 where the first and third transitions can make @xmath173@xmath174 and @xmath173@xmath175 .
so , we have two e1 transitions for the pv neutron radiative capture by deuteron , corresponding to the @xmath120 partial wave of the incoming neutron - deuteron system ( @xmath176@xmath177 and @xmath178@xmath179 )
. the spin structure of the e1 matrix element of pv @xmath0 @xmath1 process can be written as @xmath180 we can obtain for both incoming doublet ( @xmath176@xmath177 ) and quartet ( @xmath178@xmath179 ) channels @xmath181\ , , \nonumber\\ \nonumber\mathcal{w}^{pv}(^4s_{\frac{3}{2}})\big[\;t^\dag\big(\vec{\varepsilon}_d\cdot\vec{\varepsilon}^\ast_\gamma+i\vec{\sigma}\cdot\vec{\varepsilon}_d\times\vec{\varepsilon}^\ast_\gamma\big)n\big]\,.\end{aligned}\ ] ] where @xmath182 is the amplitude of the e1 transition for the incoming @xmath120 partial wave . in this section , we are going to calculate @xmath182 from all possible feynman diagrams introduced in figs.[fig3 ] and [ fig4 ] for the pv @xmath0 @xmath1 process .
if we want to use the pv @xmath9 scattering amplitude for the calculation of some diagrams in the set ( i ) and also the set ( ii ) , we should make clear the possible incoming and outgoing partial waves for pv @xmath9 scattering amplitude . for the @xmath183 , @xmath184 , @xmath185 and @xmath186 diagrams in fig.[fig4 ]
, we can schematically represent the pv and e1 transitions as @xmath187 where above scheme shows the incoming and outgoing partial wave of the pv @xmath9 scattering are @xmath120 and @xmath121 , respectively . at the thermal energy ,
@xmath120 can be @xmath188@xmath177 and @xmath189@xmath179 , as incoming doublet and quartet channels , and @xmath121 also can be @xmath188@xmath190 , @xmath188@xmath191 , @xmath189@xmath190 or @xmath189@xmath191 ( see eq.([eq:35 ] ) ) .
furthermore , because the electric interaction does not change the relative spin of the particles , all e1 transitions which can create the triton are @xmath192 so , from eqs.([eq:35 ] ) , ( [ eq:025 ] ) and ( [ eq:026 ] ) , we finally have only two possible transitions @xmath193 from the initial s - wave state to triton for the @xmath183 , @xmath184 , @xmath185 and @xmath186 diagrams which are shown in fig.[fig4 ] .
hence , @xmath194@xmath195@xmath196 and @xmath197@xmath198@xmath199 amplitudes are needed for calculating the contribution of these diagrams .
a similar arguments can be expressed for the @xmath200 ( @xmath201 ) , @xmath202 and @xmath174 diagrams in the set ( i ) .
consequently , for these diagrams , one can see that there are two possible transitions which are presented schematically as @xmath203 thus , for these diagrams , we require @xmath204@xmath205@xmath206 and @xmath207@xmath205@xmath206 amplitudes to obtain the contribution of @xmath200 , @xmath202 and @xmath174 diagrams in fig.[fig3 ] .
@xmath208 denotes the pv @xmath9 scattering amplitude corresponding to the diagrams @xmath155 , @xmath156 and time - reversed contribution of @xmath155 , @xmath157 shown in fig.[fig2 ] .
now , one can be able to begin the calculation of the sets ( i ) and ( ii ) diagrams .
initially , we describe the set ( i ) ( fig.[fig3 ] ) and then proceed to explain the set ( ii ) ( fig.[fig4 ] ) .
let us consider the pv lecs in eq.([eq:8 ] ) and rename them conveniently from up to down by @xmath209 to @xmath210 , respectively ; for example , @xmath211 is @xmath212 .
the amplitude of all diagrams in fig.[fig3 ] , for the incoming x channel , can be written as @xmath213 where subscript `` @xmath214 '' in @xmath215 represents the contribution of the set ( i ) diagrams and @xmath120 is the partial wave of the incoming channel ( doublet or quartet ) .
we assume @xmath216 and @xmath217 . in the above equation ,
@xmath218 kernel is given by @xmath219 where @xmath220 ( @xmath32 ) is the contribution of all diagrams in the @xmath221-th line of fig.[fig3 ] .
the results of the @xmath222 and @xmath223 kernels for two incoming doublet and quartet channels are presented in [ appendix b ] , however we introduce the @xmath224 kernel in the following .
we note that the groups of the diagrams represented by @xmath225 , @xmath226 and @xmath227 in fig.[fig3 ] have two contributions corresponding to the poles in the nucleon propagators before and after the photon creation .
the contribution of all diagrams in the line 3 of fig.[fig3 ] , for the incoming @xmath120 channel , in the cluster - configuration space is given by @xmath228 where , as previously explained in eq.[eq:027 ] , @xmath121 partial wave should be @xmath229 and @xmath230 for @xmath231@xmath170 and @xmath171 , respectively .
so , we evaluate for the corresponding partial waves the following @xmath232 matrices , @xmath233 \nonumber \\+g(e_i , e_f , q)\left ( \begin{array}{cc } 4g_5 & -\tau_3(g_1+g_2 ) \\
-\tau_3(g_1+g_2 ) & 0 \\ \end{array } \right)\bigg\},\qquad\qquad\qquad\qquad\quad \nonumber \\ \nonumber \\
o(^4s_{\frac{3}{2}};e_i , e_f , q)=\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\qquad\quad\:\:\nonumber\\ -\frac{eym_n}{9\sqrt{6}\pi}\bigg\{3\bigg[\sqrt{\frac{3q^2}{4}-m_ne_f}\left ( \begin{array}{cc } g_5 & 0 \\ \tau_3g_1 & 0 \\ \end{array } \right)+\sqrt{\frac{3q^2}{4}-m_ne_i } \left ( \begin{array}{cc } g_5 & 0 \\ \tau_3g_2 & 0 \\ \end{array } \right)\bigg ] \nonumber \\+g(e_i , e_f , q)\left ( \begin{array}{cc } -2g_5 & 0 \\ -\tau_3(g_1+g_2 ) & 0 \\ \end{array } \right)\bigg\}.\qquad\qquad\qquad\qquad\qquad\quad\end{aligned}\ ] ] the calculated @xmath234 matrix in eq.([eq:0000 ] ) is given for the required partial waves by @xmath235\left ( \begin{array}{cc } 1 & 0 \\ 0 & 1 \\ \end{array } \right ) , \nonumber\\ \tilde{o}(^4s_{\frac{3}{2}}\rightarrow^4p_{\frac{1}{2}};e_i , e_f , q)= \sqrt{\frac{2}{3}}\tilde{o}(^2s_{\frac{1}{2}}\rightarrow^2p_{\frac{1}{2}};e_i , e_f , q)\left ( \begin{array}{cc } 1 & 0 \\ 0 & 0 \\ \end{array } \right),\qquad\qquad\qquad\qquad\qquad\!\!\nonumber\\ \tilde{o}(^2p_{\frac{1}{2}}\rightarrow^2s_{\frac{1}{2}};e_i , e_f , q)= \tilde{o}(^2s_{\frac{1}{2}}\rightarrow^2p_{\frac{1}{2}};e_i , e_f , q),\qquad\quad\;\;\qquad\qquad\qquad\qquad\qquad\qquad \nonumber\\ \tilde{o}(^2p_{\frac{3}{2}}\rightarrow^2s_{\frac{1}{2}};e_i , e_f , q)= \frac{2}{\sqrt{3}}\tilde{o}(^2s_{\frac{1}{2}}\rightarrow^2p_{\frac{1}{2}};e_i , e_f , q).\qquad\quad\;\,\qquad\qquad\qquad\qquad\qquad\quad\!\!\end{aligned}\ ] ] we have also presented the results of @xmath236 and @xmath237 in eq.([eq:00028 ] ) because they are required in the following calculation regarding the diagram @xmath186 in fig.[fig4 ] . in eq.([eq:023 ] ) the @xmath238 function is @xmath239\:\:\,\nonumber \\-\left[(\frac{3q^2}{4}-m_ne_i)^{3/2}-(\frac{3q^2}{4}-m_ne_f)^{3/2}\right]\bigg\}.\quad\ , \nonumber\\\end{aligned}\ ] ] the above results are evaluated by considering the suitable partial - wave projections for the incoming and outgoing channels .
one can see all projection operators in the cluster - configuration space which are used in our calculations in the ref.@xcite .
all other possible diagrams in the calculation of pv @xmath0 @xmath1 amplitude , at the thermal energy , are shown in fig.[fig4 ] as the set ( ii ) .
we note that the @xmath184 diagram in fig.[fig4 ] has no contribution in @xmath240 because the three - body force introduced in sect.[pc nd scattering sector ] is zero for the p - wave ( @xmath241 ) .
the contribution of the @xmath183 and @xmath185 diagrams in fig.[fig4 ] , for two incoming doublet ( @xmath231@xmath188@xmath177 ) and quartet ( @xmath231@xmath189@xmath179 ) channels is given by @xmath242\times\:\ : \nonumber\\ \times \mathcal{d}^{lo}(e_i-\frac{q^2}{2m_n},q)t^{lo , pv}(x\rightarrow y;e_i , k , q),\quad\end{aligned}\ ] ] where @xmath183 and @xmath185 can be written as @xmath243\qquad\nonumber\\ \times\left ( \begin{array}{cc } -3(1+\tau_3 ) & 9(1+\tau_3 ) \\ 3(3-\tau_3 ) & -(3-\tau_3 ) \\
\end{array } \right),\qquad\qquad\qquad\qquad\qquad\quad\!\!\!\nonumber\\\end{aligned}\ ] ] and @xmath244 \left ( \begin{array}{cc } -3(1-\tau_3 ) & 3(3+\tau_3 ) \\
3(3+\tau_3 ) & -(3 + 5\tau_3 ) \\
\end{array } \right ) .\nonumber\\\end{aligned}\ ] ] in the above equations , @xmath245 is @xmath246 for @xmath247@xmath190 and @xmath248 for @xmath247@xmath191 partial wave .
the final diagram in the @xmath4 calculation is the @xmath186 diagram shown in fig.[fig4 ] . after applying the feynman rules and the appropriate incoming and outgoing projections
, one can see that the contribution of the @xmath186 diagram is given by @xmath249 in the eqs.([eq:0045])-([eq:00026 ] ) , @xmath121 partial wave as explained in eq.[eq:027 ] is @xmath188@xmath190 and @xmath188@xmath191 for the incoming doublet and quartet channels , respectively .
the calculated results of the @xmath250@xmath206 and @xmath251@xmath206 are previously introduced in eq.([eq:00028 ] ) .
we emphasize that the diagrams for the interaction of @xmath22 with a photon has no contribution because @xmath22 has no derivatives , so it is not affected by the minimal substitution @xmath252 .
finally , the physical amplitude of all diagrams in the sets ( i ) and ( ii ) is given by @xmath253 where @xmath215 and @xmath254 are the amplitudes of all diagrams in the set ( i ) and ( ii ) , respectively .
we note that the value of @xmath142 is -1 for @xmath0 @xmath1 as previously explained in the sect.[pv nd sector ] .
in the calculation of the photon circular polarization , @xmath4 , in @xmath0 @xmath1 process we need to evaluate the total amplitude of this process . the total amplitude is sum of both pc and pv contributions that introduced in the previous sections .
we generally classify the numerical computation in three steps . in the first step ,
we compute the full - offshell doublet and quartet pc neutron - deuteron scattering amplitudes and the triton wave function at lo .
the pc neutron - deuteron scattering and the triton wave function are used in the calculations of the pc and pv @xmath0 @xmath1 amplitudes .
the pc neutron - deuteron scattering is obtained by solving numerically the faddeev equations in eqs.([eq:05 ] ) and ( [ eq:5 ] ) .
we solve it by the hetherington - schick method @xcite in a mathematica code with an arbitrary cutoff momentum @xmath37 , for detail see ref.@xcite .
we also obtain the triton wave function by solving eq.([eq:0005 ] ) as previously explained . in the second step ,
we use the off - shell pc neutron - deuteron scattering amplitude which is computed numerically in the previous step in order to obtain the pv neutron - deuteron scattering amplitude of the diagrams in fig.[fig2 ] .
we have written a new mathematica code for this step and use the results of the pc neutron - deuteron scattering amplitude as input data .
note that we choose the same cutoff momentum @xmath37 for the pc neutron - deuteron scattering and the pv sector calculations .
coordinate system used in the integration over solid angles.,width=302 ] the final step is the calculation of the pc ( m1 ) and pv ( e1 ) @xmath0 @xmath1 amplitudes .
the pc amplitude of @xmath0 @xmath1 process is obtained by using the pc neutron - deuteron scattering amplitude and the triton wave function results . for the calculation of the pv amplitude
we require to evaluate the set ( i ) diagrams of pv @xmath0 @xmath1 , fig.[fig3 ] .
therefore , we must compute the @xmath255 kernel which is introduced in eq.([eq:00021 ] ) ( the contribution of the diagrams in the lines 1 , 2 and 3 of set ( i ) ) . for the diagrams in the lines 4 , 5 and 6 of the set ( i ) , we add the half - offshell pc neutron - deuteron scattering amplitude from the left side to the @xmath255 kernel ( see eq.([eq:0021 ] ) ) .
finally , the set ( ii ) diagrams in fig.[fig4 ] also computed similar to the set ( i ) diagrams by using the appropriate half - offshell pv neutron - deuteron scattering and the triton wave function data .
we solve the integrations numerically by using the gaussian quadrature weights and the same cutoff @xmath37 as before .
we emphasize that for integration over the solid angle , we choose a coordinate system in which the momentum of outgoing photon is in the z direction and the @xmath256 momentum is in the x - z plane ( see fig.[fig5 ] ) .
in this work , we have concentrated on the evaluation of photon circular polarization , @xmath4 , in @xmath0 @xmath1 process and obtaining a new relation , which is useful for extracting five leading independent lecs .
the pv polarization of photon is given by @xmath257 where @xmath258 and @xmath259 are the cross section for the photons with right and left helicity , respectively .
we know that the cross sections are calculated with spin summation on the square amplitude
. therefore one can conclude that the pv polarization of photon can be written as @xmath260}{\big|\mathcal{w}^{pc}(^2s_{\frac{1}{2}})\big|^2+\big|\mathcal{w}^{pc}(^4s_{\frac{3}{2}})\big|^2},\nonumber\\\end{aligned}\ ] ] where @xmath261 and @xmath182 are the pc and pv amplitudes of @xmath0 @xmath1 process which are introduced and evaluated in sects.[pc nd radiative capture sector ] and [ pv nd radiative capture sector ] , respectively . from eq.([eq:320 ] ) , one can evaluate the @xmath4 in terms of the pv lecs .
so , for the photon circular polarization , we have @xmath262 where @xmath263 is calculated numerically and shown in fig.[fig9 ] .
the @xmath4 does not contain @xmath264 term because the @xmath9 system is an iso - doublet and pv coupling @xmath264 can not contribute . according to the table [ tab:3 ] which is shown below ,
we have calculated the @xmath265 $ ] for the different incoming and outgoing partial waves .
cc @xmath263 & @xmath266 $ ] + @xmath267 & @xmath268 + @xmath269 & @xmath270 + @xmath271 & @xmath272 + @xmath273 & @xmath274 + these results indicate that the cutoff dependence is less than @xmath275 .
this small variation can be removed with the consideration of the higher order corrections .
therefore , the variation of the @xmath263 indicates the ignorable dependence on cutoff at the leading order .
so , it can be concluded that at lo for pv neutron radiative capture by deuteron no new three - nucleon interaction is needed . in the numerical calculation for the results shown in [ fig9 ] , we used for the nucleon mass @xmath276 mev , deuteron binding momentum @xmath277 mev , effective range of deuteron @xmath278 fm , effective range of @xmath3 singlet channel @xmath279 fm , scattering length in singlet channel of @xmath280 fm and a triton binding energy @xmath281 mev .
it is obvious that the pv coupling constants are not predicted by the eft but can be estimated on dimensional grounds .
we expect that the magnitude of the pv coupling to be of the order @xcite @xmath282 of course , we note that this result is dimensional , namely order - of - magnitude estimation and may be off by factors of 10 or more . with this
estimated value for the pv coupling constants and order of magnitude of the results of @xmath263 in fig.[fig9 ] , the value of @xmath4 in @xmath0 @xmath1 is estimated as @xmath283 experimentally , the @xmath4 in @xmath0 @xmath1 is not reported up to now at the thermal energy .
therefore we can compare our estimated result with a theoretical methods such as ddh model .
the calculated values of @xmath4 in @xmath0 @xmath1 based on ddh model are given by @xmath12 and @xmath13 for two super - soft - core ( ssc ) and reid - soft - core ( rsc ) potentials , respectively @xcite .
recently , the value of @xmath4 in @xmath0 @xmath1 is calculated with the ddh - best parameter values and 4-parameter fits for the different potential models in ref.@xcite .
the order of magnitude of the results of @xmath4 in the ref.@xcite is the order of @xmath284 .
we can see that our eft(@xmath2 ) estimation for the @xmath4 in @xmath0 @xmath1 process at the thermal energy , eq.([eq:034 ] ) , agrees with the order of other theoretical methods .
in the present eft(@xmath2 ) calculation , the @xmath4 is calculated in order to minimise the uncertainty in determination of the lecs .
we have used the pv two - body interaction in our calculation .
no pv three - body interaction is taken into account at the leading order .
the value of the coefficient @xmath263 can provide the valuable information in order to use in eq.([eq:032 ] ) .
one can be able to compare eq.([eq:032 ] ) with the future experimental value of @xmath4 .
the variation of the coefficient @xmath263 indicates that the cutoff dependence is ignorable at the leading order .
this small variation can be removed with the consideration of the higher order corrections .
the photon asymmetry @xmath8 for the neutron - deuteron radiative capture process is an open problem which we are going to calculate it in the near future .
the pv proton - deuteron radiative capture is another process which can help us to identify parity - violating few - body coupling constants . for this process
the coulomb interaction should be considered together with the pv and em interactions , simultaneously .
more challenging reactions such as pv four - body scattering will make the use of eft more interesting in the future study .
this shed light on the accurate evaluation of weak @xmath7 coupling constants .
the authors would like to thank harald w. grie@xmath185hammer for the valuable @xmath9 scattering mathematica code .
this work was supported by the research council of the university of tehran .
in order to normalize the triton wave function , one can be able to follow the derivation of the normalization condition for bethe - salpeter equation in @xcite .
we replace the normalization condition of the relativistic two - body bs vertex function @xmath285 with the non - relativistic one which is suitable for the @xmath9 scattering leading to the formation of triton .
one can start from the homogenous equation @xmath286 with the normalization condition introduced by adam _
et al . _ , @xcite as @xmath287 where @xmath288 denotes the interaction kernel and @xmath289 represents the bs propagator . @xmath290 and @xmath291 in eq.([eq:51 ] ) are total four - momentum of the system and the bound state mass , respectively . based on eq.([eq:0005 ] ) , the non - relativistic expression similar to eq.([eq:51 ] ) can be written as @xmath292\bigg|_{e =- b_t } \,t^{^3h}(q')\bigg(\begin{array}{c } 1 \\ 0 \end{array } \bigg).\nonumber \\\end{aligned}\ ] ] where , after taking into account the energy and solid angle integrations , we have @xmath293\bigg|_{e =- b_t}\!\!\!t^{^3h}(q')\bigg(\begin{array}{c } 1 \\ 0 \end{array } \bigg),\nonumber \\\end{aligned}\ ] ] where @xmath294 has been introduced in the text .
we note that @xmath295 vector in eq.([eq:52 ] ) projects @xmath294 into incoming dibaryon in the deuteron case .
similarly , for the incoming dibaryon in the singlet case , @xmath295 and @xmath296 should be replace by @xmath297 and @xmath298 in eq.([eq:52 ] ) , respectively . by comparing eq.([eq:0005 ] ) with ( [ eq:50 ] ) , one can be able to see @xmath299 .
\nonumber \\\end{aligned}\ ] ]
we introduce the contribution of all diagrams in the first and second lines of fig.[fig3 ] as @xmath300 which is the sum of @xmath222 and @xmath223 , @xmath301 in eq.([eq:0059 ] ) , @xmath120 and @xmath256 are the incoming partial wave and momentum , respectively . @xmath302 and @xmath303 are the same as introduced in the text . the @xmath300 can be written in the cluster - configuration space , after applying the energy integration , as @xmath304\qquad\quad \nonumber\\ + \frac{1}{4\pi^4}\int\,p^2\,dp\,\tilde{t}^{lo , pv}(y\rightarrow ^2\!\!s_{\frac{1}{2}};e_f , p , q')\times\qquad \nonumber\\ \times\mathcal{d}^ { lo}(e_f-\frac{p^2}{2m_n},p)\big[\bar{\mathcal{s}}^{e}+\bar{\mathcal{s}}^{f}\big]\;\;\big\ } \nonumber\\-\mathcal{s}^{c^{up}}-\mathcal{s}^{c^{down}}\bigg\},\qquad\qquad\qquad\qquad\qquad\qquad\quad\;\;\nonumber\\ \end{aligned}\ ] ] where @xmath305 ( @xmath306 , @xmath226 ) is a matrix function of @xmath307 and @xmath308 ( @xmath309 , @xmath202 ) is a matrix function of @xmath310 . in this section of appendix b ,
we introduce the relations of all @xmath311 and @xmath308 matrix functions for the incoming doublet channel . in the following ,
we have replaced @xmath312 , @xmath212 , @xmath313 , @xmath314 and @xmath315 by @xmath209 , @xmath211 , @xmath316 , @xmath317 and @xmath210 , respectively , as in the text . after applying the adequate projections for @xmath176@xmath177 channel , in the cluster - configuration space , the @xmath318 and @xmath319 are given by @xmath320 \nonumber\\ \times\left ( \begin{array}{cc } ( 1-\tau_3)(3g_1 - 2g_5 ) & -(3+\tau_3)g_2 + 2(1+\tau_3)g_3 ) \\ ( 3+\tau_3)(g_1 + 2g_5 ) & -(3 + 5\tau_3)g_2 - 2(1+\tau_3)g_3 ) \\
\end{array } \right ) , \nonumber\\ \end{aligned}\ ] ] @xmath321 we can also obtain the @xmath322 and @xmath323 for the doublet channel by @xmath324^\dagger .
\end{aligned}\ ] ] for all other @xmath311 and @xmath308 matrices , in the incoming doublet channel , we have @xmath325\quad \nonumber\\ \times\left ( \begin{array}{cc } ( 1+\tau_3)(3g_1 + 2g_5 ) & -(1+\tau_3)(3g_2 + 2g_3 ) \\
( 3-\tau_3)(g_1 + 2g_5 ) & -(3-\tau_3)g_2 - 2(1+\tau_3)g_3 \\
\end{array } \right ) , \end{aligned}\ ] ] @xmath326 \nonumber\\ \times\left ( \begin{array}{cc } ( 1\!+\!\tau_3)(3g_1\!+\!2g_5 ) & ( 1\!+\!\tau_3)(3g_1\!+\!2g_5 ) \\
-(3\!-\!\tau_3)g_2\!+\!2(1\!-\!\tau_3)g_3 & -(3\!-\!\tau_3)g_2\!+\!2(1\!-\!\tau_3)g_3 \\ \end{array } \right ) , \nonumber\\ \end{aligned}\ ] ] and @xmath329 \left ( \begin{array}{cc } -3(1-\tau_3 ) & 3(3+\tau_3 ) \\
3(3+\tau_3 ) & -(3 + 5\tau_3 ) \\ \end{array } \right ) .
\end{aligned}\ ] ] in the eqs.([eq:65 ] ) and ( [ eq:66 ] ) the @xmath330 , @xmath331 and @xmath332 matrix functions are given by @xmath333 , \nonumber\\ \nonumber\\
n_2(e,\vec{q},\vec{l},\vec{q'})=\frac{(\vec{q'}+2\vec{l})\cdot\vec{l}}{m_n\,e\!-\!l^2\!-\!q'^2\!-\!\vec{l}\cdot\vec{q'}}\mathcal{d}^ { lo}(e-\frac{q'^2}{2m_n},q')\qquad\qquad\quad \nonumber\\ \times\bigg[\frac{1}{m_n\,e\!-\!q^2\!-\!l^2\!-\!\vec{q}\cdot\vec{l}}\left ( \begin{array}{cc } -3(1+\tau_3 ) & 9(1+\tau_3 ) \\ 3(3-\tau_3 ) & -(3-\tau_3 ) \\ \end{array } \right)\;\;\:\ ; \nonumber\\ + \frac{2h_0(\lambda)}{\lambda^2}\left ( \begin{array}{cc } 3(1+\tau_3 ) & -3(1+\tau_3 ) \\
-(3-\tau_3 ) & ( 3-\tau_3 ) \\ \end{array } \right)\bigg ] , \nonumber\\ \nonumber\\ n_3(e,\vec{q},\vec{l},\vec{p})=\frac{\vec{p}\cdot\vec{l}}{m_n\,e\!-\!l^2\!-\!p^2\!-\!\vec{l}\cdot\vec{p}}\mathcal{d}^ { lo}(e-\frac{p^2}{2m_n},p)\qquad\qquad\quad\:\ : \nonumber\\ \times\bigg[\frac{1}{m_n\,e\!-\!q^2\!-\!l^2\!-\!\vec{q}\cdot\vec{l}}\left ( \begin{array}{cc } -3(1+\tau_3 ) & 9(1+\tau_3 ) \\ 3(3-\tau_3 ) & -(3-\tau_3 ) \\ \end{array } \right)\;\;\,\ , \nonumber\\ + \frac{2h_0(\lambda)}{\lambda^2}\left ( \begin{array}{cc } 3(1+\tau_3 ) & -3(1+\tau_3 ) \\
-(3-\tau_3 ) & ( 3-\tau_3 ) \\ \end{array } \right)\bigg ] , \nonumber\\ \end{aligned}\ ] ] where
@xmath138 and @xmath140 are introduced previously in sect.[pv nd sector ] .
similar to the doublet channel , we can write the following equations for the quartet channel .
these results are obtained in the cluster - configuration space after applying the adequate incoming and outgoing partial - wave projections . all calculated @xmath311 and @xmath308 matrix functions in the quartet channel are @xmath336 \nonumber\\ \times\left ( \begin{array}{cc } ( 1-\tau_3)(3g_1+g_5 ) & 0 \\ ( 3+\tau_3)(g_1-g_5 ) & 0 \\ \end{array } \right),\qquad\qquad\:\:\qquad\qquad\quad\ ! \nonumber\\ \end{aligned}\ ] ] @xmath337 \nonumber\\ \times\left ( \begin{array}{cc } -(1-\tau_3)g_5 & 0 \\ 2(3+\tau_3)g_2 + 4(1+\tau_3)g_3 & 0 \\ \end{array } \right ) , \qquad\qquad\quad\qquad\!\!\!\nonumber\\ \end{aligned}\ ] ] @xmath338 @xmath339 we introduce two contributions in eqs.([eq:62]-[eq:67 ] ) and ( [ eq:73]-[eq:77 ] ) for the poles in the nucleon propagators before and after the photon creation , as mentioned in the text .
we note that the groups of the diagrams represented by @xmath226 and @xmath227 in fig.[fig3 ] have no contribution in the incoming quartet channel because in these diagrams the nucleon which interacts with the photon is a neutron ( for @xmath178@xmath179 channel ) and a photon has no e1 interaction with a neutron . | we study the parity - violating effects in @xmath0 @xmath1 process using pionless effective field theory ( eft(@xmath2 ) ) at thermal energy . for the weak @xmath3 interaction the parity - violating lagrangian contains five independent low - energy coupling constants ( lecs ) .
one can fix the coupling constants by comparison of the calculated observables with a sufficient number of experimental data .
so , we need five observables to obtain lecs .
one of the five parity - violating observables is the photon circular polarization ( @xmath4 ) in @xmath0 @xmath1 process .
we calculate @xmath4 at the leading order in terms of lecs .
we compare our results with the previous @xmath4 calculation based on the model of desplanques , donoghue , and holstein .
o # 1#2_#1 -2.5pt # 1#2#3_#1 |
the galactic center ( gc ) is a unique laboratory for a detailed study of star formation and its impact on the nuclear environment of galaxies . because of its proximity
, the gc can be imaged at resolutions more than 100 times better than the nearest galaxies like our own ( e.g. , the andromeda galaxy ) .
this unmatched high - resolution capability is particularly important to the study of the formation and destruction processes of stellar clusters as well as the overall population of massive stars in such an environment .
indeed , the gc is known to host three very massive clusters ( the arches and quintuplet clusters , as well as the central parsec cluster surrounding sgr a@xmath4 ; ages @xmath5 yrs , masses @xmath6 each ; e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) , which are responsible for about 5% of the total lyman continuum flux for the galaxy , in less than 0.01% of its volume .
massive stars are known to exist outside the clusters @xcite , however , locating these important stars has been sporadic and incomplete ( e.g. * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these isolated massive stars may have formed individually or in small groups , although the extreme conditions in the gc make this mode of star formation unusually difficult .
they may also have been spun out of the clusters due to their internal dynamics .
this process is expected to be accelerated by the strong external tidal force in the gc @xcite .
the star formation process depends critically on the interplay of massive stars with the interstellar medium ( ism ) .
massive stars release large amounts of ultraviolet radiation .
they also produce strong stellar winds throughout their short lives and die in supernova explosions .
such energetic stellar output dramatically shapes the morphology of the ism , as is revealed in existing radio , mid - infrared , and x - ray maps ( fig .
[ f : fig1 ] ; e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
perhaps the best - known example of this is the radio arc region which surrounds the arches and quintuplet clusters : a collection of arc - like thermal filaments , synchrotron - bright non - thermal linear filaments , and diffuse , hot gas .
some of the structures within this region , the sickle hii region ionization front and photoevaporated pillars in particular , suggest that gas is being collected and compressed ( * ? ? ?
* cotera et al . in preparation ) , and is morphologically resembles known regions outside the gc where such processes are believed to have led to the formation of a new generation of stars ( e.g. m16 ; @xcite ) . the energy injection from massive stars and possibly from the central supermassive black hole ( smbh ) , that is coincident with the radio source sgr a * , may also strongly affects the thermal and/or dynamical properties of the ism .
molecular gas in the gc , for example , has unusually high turbulent velocities and temperatures compared to clouds in other parts of the galaxy ( e.g. , * ? ? ?
this may bias the stellar initial mass function toward heavier stars and favor the production of massive star clusters ( * ? ? ?
* ; * ? ? ?
* and references therein ) .
to constrain the dynamical process of the known clusters , the overall population of massive stars , and their formation modes , we sought to obtain a uniform survey of massive stars and their interplay with the ism across the central 90 parsecs .
ionized warm gas provides the ideal tracer of massive stars and their interplay with their surroundings .
although previous radio continuum observations are free of extinction and trace both thermal and non - thermal emission , the scale - dependence of radio interferometers makes it difficult to accurately measure flux densities on arcsecond scales . in addition , fine features ( e.g. , ionization fronts ) are often found in the midst of diffuse non - thermal emission ( e.g. , the radio arc ) , making decomposition of thermal and non - thermal components challenging .
radio recombination line observations typically have a substantially lower signal - to - noise and spatial resolution than radio continuum data . in the near - ir ,
the 2.16 br@xmath7 line is accessible from the ground ( no suitable filter is available for existing and planned space - based platforms ) , but has also not proved effective for large scale mapping of warm ionized gas . to our knowledge , no large - scale br@xmath7 line survey of the gc has ever been published .
the difficulty lies in the lack of the combination of high spatial resolution and stable point spread function ( psf ) in a ground - based survey .
observations with adaptive optics ( ao ) can have superb resolution , but only over small fields of view ( typically @xmath8 ) .
the ao psf also changes strongly , both spatially and with time , and has extended psf wings , making the psf subtraction extremely difficult .
in addition , photometric conditions rarely , if ever , persist for the time required for such a survey , hampering the study of large - scale features .
we have carried out the first large - scale , high - resolution , near - ir survey of the gc , using _ hst _
the primary objective of the survey is to map out the hydrogen paschen-@xmath0 ( p@xmath0 for short ) line ( wavelength 1.876 @xmath9 ) in a field of @xmath10 around sgr a@xmath4 ( corresponding to 90 pc @xmath11 35 pc at the gc distance of 8 kpc ; * ? ? ?
[ f : fig1 ] ) .
this field is known to be rich in clustered massive star formation ( i.e. , including the three known clusters ) .
nicmos observations of the p@xmath0 emission were taken previously only for a few discrete regions , centered on the three known massive clusters ( see [ ss : comparion ] for further discussion ) .
our survey field also includes much unexplored territory . in particular , the region between sgr a and sgr c along the galactic plane ( to the right in fig .
[ f : fig1 ] ) shows a number of compact _ spitzer _ irac sources ( several with radio counterparts ) likely related to some of the earliest stages of massive star activities .
this field selection also facilitates an unbiased study of the apparent asymmetry between the positive and negative galactic longitude sides relative to sgr a@xmath4 .
our nicmos observations overcome all of the potential problems of a ground - based survey .
the wide nic3 camera on this instrument provides a resolution of @xmath12 with a stable psf .
this capability is important for detecting and cleanly removing point - like sources as well as resolving fine structures of extended features .
both the stable psf and the absence of atmosphere for our survey means that the entire survey is photometrically accurate and consistent .
the extra magnitude of extinction along the sight - line to the gc at 1.87 ( as compared to 2.16 ) is more than compensated for by an intrinsic p@xmath0 line intensity 12 times greater than the br@xmath7 line @xcite .
in addition , the background at p@xmath0 with nicmos is a factor of @xmath13 lower than the sky background at br@xmath7 as observed from the ground .
this low background of nicmos also allows us to detect faint point - like sources and extended features , particularly over our large mosaicked field .
our survey was completed in 144 hst orbits between february 22 and june 5 , 2008 .
all of these orbits had an identical design , each consisting of @xmath14 nic3 pointing positions for both f187n ( on - line ) and f190n ( off - line ) filters .
these filters have a 1% bandpass centered at the wavelengths 1.875@xmath9 and 1.900 @xmath9 .
nic3 is a 256@xmath11 256 hgcdte array detector with a pixel size of @xmath15 in the x and y directions .
adjacent positions and orbits were designed to have 3@xmath16 overlaps in the vertical direction of the nic3 array coordinates , partly to allow for the bottom 15 pixels that do not provide useful data due to afocal vignetting of that portion of the array .
this is in addition to the overlaps resulting from the 4-point dithers at each of the four positions of an orbit .
each dither exposure was taken in the multiaccum readout mode ( samp - seq=16 ; nsamp=9 ) for photometry over a large flux range and for effective identification and removal of cosmic - rays and saturated pixels .
the 4-point dithers were in an inclined square wave pattern ( i.e. , the shifts were in the directions of [email protected] relative to the x - axis of the detector coordinates ) ; each dither step produced a shift of 233 and 563 pixels in the x and y directions .
this dither pattern enabled both a sub - pixel sampling of the undersampled nic3 psf and a consistent overlap between orbits .
each orbit effectively covered an area of roughly 102@xmath18 ( 2.8 arcmin@xmath19 ) . with this strategy and a single telescope orientation ,
the survey mapped out a field of 416 arcmin@xmath19 ( as outlined in fig .
[ f : fig1 ] ) , and with a uniform 192 sec exposure per filter .
the entire science data set includes 144 ( orbits ) @xmath11 4 ( pointing positions ) @xmath11 4 ( dithers ) @xmath11 2 ( filters ) @xmath20 images .
in addition , between target visibility periods ( during earth occultations ) after each of the 144 orbits , 16 dark frames were obtained with the detector clocked in a manner identical to the science exposures .
these inter - orbit raw dark exposures were assembled to form a `` superdark '' frame used in our image calibration .
-emitting complex ( g359.866 + 0.002 ) , about 5@xmath21 northwest of sgr a * : ( a ) f187n image , ( b ) continuum - subtracted f187n image , and ( c ) point source removed ( diffuse ) image .
the intensity is logarithmically scaled over the range of 0.1 - 30 @xmath22 in ( a ) and 0.1 - 1 @xmath22 in ( b ) and ( c).,scaledwidth=48.5% ] the data reduction and analysis , based on a combination of several newly developed tools and the standard routines implemented in the software iraf / stsdas , will be detailed in @xcite .
briefly , in addition to the standard dc offset correction , we use a global fitting method to determine the relative instrumental background offsets among all nic3 array quadrants , pointing positions , and orbits , based on multiple overlapping fields . a similar method is also used to optimize the correction for the relative astrometry offsets among orbits .
the absolute astrometry is corrected in respect to radio counterparts of the near - ir sources in the field , reaching an accuracy better than @xmath23 .
the source detection is based primarily on the _ starfinder _ software @xcite .
the procedure and product of the data reduction are illustrated in fig .
[ f : fig2 ] for global views and fig .
[ f : fig3 ] for full - resolution close - ups . to estimate the continuum contribution in the f187n band ( figs . [
f : fig2]a and [ f : fig3]a ) , we adaptively construct a mean f187n to f190n intensity ratio map , with the value of each pixel given as the median ratio of the nearest 100 sources , most of which are assumed to be located at the gc
. however , this assumption does not hold for regions with exceptionally high extinction , as judged from a low surface number density of detected sources ( i.e. , @xmath24 sources arcsec@xmath25 ) . for such regions ,
we calculate the ratio values from the extinction map obtained from _ spitzer
_ observations ( * ? ? ?
* ; * ? ? ?
* for details ) .
the continuum - subtracted map ( figs . [
f : fig2]b and [ f : fig3]b ) gives a crude look of the potential -emitting sources as well as the diffuse emission .
however , one can not judge whether or not a peak ( e.g. , in fig .
[ f : fig2]b or [ f : fig3]b ) is a emission source based only on its apparent excess above the background ; such a peak can easily be caused by the statistical flux fluctuation of a subtracted source , especially a bright one .
in fact , all peaks in [ f : fig3]b , except for the brightest one , are below our current source detection threshold .
we identify source candidates based on their integrated continuum - subtracted fluxes at a statistical confidence @xmath26 , where @xmath27 accounts for all known measurement errors @xcite . to produce a _ diffuse
_ map ( figs .
[ f : fig2]c and [ f : fig3]c ) , we use the actual f187n to f190n flux ratio of a detected source for its affected pixels ( defined to have its psf - predicted intensity contribution @xmath26 above the local background in the f190n map ) .
we further remove any uniformly - distributed foreground and residual instrument background in a map by shifting its intensity to zero , averaged over a few darkest regions ( see [ s : results ] ) .
[ f : fig2 ] shows a panoramic view of our survey data . clearly , the distribution of both the stellar light and the diffuse emission is highly inhomogeneous , partly due to the variations in foreground extinction , which varies strongly from one region to another and on different spatial scales .
the most outstanding extinction features in fig .
[ f : fig2 ] are the distinct dark filaments ( tendrils ) in the field just above sgr a , which are also noticeable in published k - band images ( e.g. , * ? ? ?
the darkness of these filaments indicates that they are foreground thick dusty clouds silhouetted against the ir - luminous nuclear region of the galaxy .
the filaments appear to be structured down to arcsecond scales [ or @xmath28 cm , where @xmath29 is the distance to such a filament ; see the full resolution close - ups in fig .
[ f : fig3 ] ] . with our data
, we can distinguish between extinction and lack of ionized emission by comparing the and f190n images if a region is dark in both images , we can say with confidence that the dark features indicate regions with extinctions av @xmath30 . fig . [
f : fig2]c shows extended -emitting complexes across much of the field .
strong and distinctive emission features are seen on angular scales from sub - arcsecond to about 10@xmath21 , with surface brightness ranging from as high as @xmath31 down to a 3@xmath27 level of @xmath32 for the image with the 04 @xmath11 04 pixel .
the rms noise of a map depends on its pixel size and is @xmath33 for the full resolution image with a pixel size of 01 @xmath11 01 , accounting for both statistical and systematic fluctuations ( photon counting , artifacts from stellar continuum subtraction , etc . ) .
this noise is estimated from selected regions in the map ( e.g. , at the upper right corner of the map ; see * ? ? ? * for details ) that show little emission but are bright in the f190n continuum image ( hence the rms contribution from clumpy structures in the emission or extinction should be minimal ) .
the overall distribution of the emission is highly lopsided toward the positive galactic longitude side of the gc , similar to what is seen in radio and mid - ir ( fig .
[ f : fig1 ] ) .
the diffuse emission is greatly enhanced around the three known clusters responsible for the ionization of the inner rims of surrounding gas .
the emission around the gc cluster is largely confined within a small spiral - like nuclear structure , the emission of which was first detected by @xcite .
our new survey now enables us to trace the faint emission features to large distances , which extend vertically below and above this structure ( fig .
[ f : fig4]a ) .
these features may represent outflows from the central few parsecs of the galaxy .
the bright emission itself indicates the ionized inner portion of a dusty circumnuclear disk around sgr a@xmath4 .
this disk appears more extended , as traced by the 8 emission ( fig .
[ f : fig4]a ) .
detailed analysis / modeling of these structures is needed to further investigate the physical state of the disk and the energetics of the gc cluster and/or the smbh , which is currently in a radiatively inefficient accretion state .
[ f : fig4]b presents a close - up of the sickle hii region , revealing fingers of ionized gas resembling the pillars of creation " in m16 @xcite .
these fingers are apparently illuminated by ionizing photons from hot stars in the quintuplet cluster ( * ? ? ?
* ; * ? ? ?
* cotera et al . in preparation ) .
the -emitting `` arches '' around the arches cluster shows an amazing array of fine linear structures ( fig .
[ f : fig4]c ) , which are preferentially oriented in the same directions as magnetic field vectors measured in infrared polarization @xcite .
similar organized narrow linear -emitting features are also abundant in some of the more compact hii regions seen in fig .
[ f : fig2]c .
these structures may indicate a critical role that locally strong magnetic fields play in shaping the morphology and dynamics of the ionized gas ( e.g. , * ? ? ?
* ) fig .
[ f : fig5 ] shows a selection of compact -emitting nebulae
. some of these enigmatic nebulae may represent individual massive stars in formation ( e.g. , first row in fig .
[ f : fig5 ] ) , whereas others may be due to stellar ejecta and/or their interaction with dense ambient medium ( second and third rows ) .
the survey also enables us to perform an unprecedentedly uniform , high resolution , photometrically sensitive census of the stellar population .
our preliminary detection based on the f190n continuum image gives a total of about 0.6 million point - like sources , which accounts for about 86% of the total observed f190n intensity in the survey field .
this is the highest ratio of resolved point sources to continuum emission achieved to date in the very crowded gc field ( e.g. , figs .
[ f : fig2]a and [ f : fig3]a ) .
the median of the unresolved intensity is @xmath34 , averaged over the surveyed field .
our 50% source detection limit is about 17th mag in the f190n band .
a preliminary analysis shows that we should detect most of the massive main - sequence stars ( earlier than b3 , corresponding to @xmath35 ) in the region as well as the bulk of red giants , including red clump stars .
a detailed analysis of the spatial distribution of the detected sources and the still unresolved intensity will allow us to infer the overall intrinsic structure of the stellar light distribution in the gc and to provide quantitative measurements of the foreground extinction .
our preliminary analysis has resulted in the detection of about 200 -emitting star candidates . before the publication of such candidates ,
however , we still need additional careful examination , especially for relatively weak sources , to minimize potential systematic contamination ( e.g. , due to local enhancements of diffuse complexes ) .
we easily detect all previously known emission line stars throughout the survey region ( @xmath36 total ; mostly wolf - rayet stars ) .
we have already verified an additional 30 emission line stars via spectroscopic follow - up observations ( * ? ? ?
* cotera et al . in preparation ) .
we anticipate that further follow up observations will at least double that number again .
more importantly , over half of the newly identified stars are located outside the three known clusters .
therefore we have already more than doubled the number of known intercluster emission line stars .
the newly discovered stars will enable us to address in greater detail the formation of massive stars outside clusters in our upcoming papers .
in particular , the star candidates outside the clusters are not distributed in a completely random fashion .
in addition to the p@xmath0 emission line , the f187n filter contains a few helium lines which are often strong emission lines in wolf - rayet stars .
therefore the most easily detected stars in the f187n filter are typically wolf - rayet stars , although future refinements to the data analysis will enable us to detect o and possible b supergiants as well , which have smaller equivalent widths , in emission or absorption .
such refinements may be achieved by modeling the broad - band spectral energy distribution together with the narrow band measurements of the p@xmath0 emission and absorption .
while the data analysis , multi - wavelength comparison , follow - up observations , and modeling are still ongoing and will be presented in later papers , we discuss here the implications of the survey for addressing several key issues related to star formation and its impact on the gc environment : two large questions remain unanswered about massive star formation in the gc : ( 1 ) do massive stars form exclusively in the clusters , or in looser associations throughout the field ? and ( 2 ) is star formation continuous , or does it happen in discrete bursts ? until recently ,
massive ( e.g. , emission - line ) stars have mostly be found within the 3 gc clusters .
a number of searches for young massive stars at infrared and x - ray wavelengths have been carried out ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
these studies , each with diverse search criteria , have yielded a rather incomplete picture of the young massive star population in this gc .
the p@xmath0 emission is a sensitive tracer of massive stars and their environs .
the emission traces warm ionized gas , produced and maintained primarily by lyman continuum radiation from massive stars .
such gas can be associated with individual stars ( e.g. , massive stellar winds ) , their immediate vicinities ( circumstellar material ) , and surrounding ism structures .
the fact that the majority of our new sources are actually located outside the three known clusters has strong implications for understanding the dynamics of massive stars with ages of only a few @xmath37 yr . the exact time scales for clusters to be disrupted and the place where the stars end up provide information about the internal dynamics of the clusters as well as the large - scale tidal force of the gc @xcite .
the spatial distribution of the massive stars detected in our survey can be used to better constrain these processes throughout the field .
the distributions of the stars and their types as well as follow - up work on both radial velocities and proper motions @xcite will allow us to constrain how quickly the massive stellar clusters are currently dissolving , how many star clusters have been disrupted , and whether or not the stars are formed in massive clusters which have been subsequently dispersed , or formed in relative isolation . in the high pressure and high density environment of the gc , the diffuse p@xmath0-emitting gas tends to be found at ionization fronts bordering molecular clouds .
the gas density and pressure are expected to change drastically at such fronts ( e.g. , * ? ? ?
the high resolution p@xmath0 map resolves various ionization fronts , which helps to determine the local neutral gas density and the thermal pressure of diffuse warm ionized gas .
the comparison between figs .
[ f : fig1 ] and [ f : fig2 ] demonstrates the complicated relationship among various stellar and ism components . within the sickle region ( fig .
[ f : fig4]b ) , for instance , we see two distinctive ionization front morphologies .
we see columns of photo - evaporating gas similar to the elephant trunks or pillars seen in regions such as m16 in one part , but also a remarkably smooth ionization front located on another side of the large hii region .
although very different , both are morphologically perpendicular to the brightest non - thermal radio filaments , which are uniquely found in the gc field and have been a long - standing mystery .
they are known to trace enhanced magnetic field and/or cosmic rays .
but the underlying physical processes remain unclear .
the observed relative morphological configuration then suggests that the large - scale , inter - cloud , vertical ( poloidal ) magnetic field is combing through the dense gas and gets enhanced in the process .
the expansion of the magnetized inter - cloud material , probably accompanied with cosmic ray acceleration , is likely driven by the strong mechanical energy output from the quintuplet and arches clusters , consistent with the barrel - shaped morphology of the nonthermal radio filaments ( fig .
[ f : fig1 ] ) .
therefore , the feedback from the clusters likely plays an important role in shaping the ism and the eco - system of the gc in general .
a few p@xmath0 observations were made previously in the gc region .
the sgr a west region , covering the inner @xmath38 4 pc of the galaxy , was observed using both nic3 and nic2 ( e.g. , * ? ? ?
several other programs concentrated only on the central parsec with nic1 .
a nic2 p@xmath0 observation was also made for the core of the arches cluster @xcite .
the p@xmath0 emission from the pistol nebula ( fig .
[ f : fig4]b ) was covered with a @xmath39 mosaic of nic2 observations , which had a smaller field of view ( 192 on a side ; * ? ? ?
our map shows considerably more structure in the diffuse emission of the nebula than the image shown in @xcite , apparently due to the higher sensitivity of our nic3 observations .
in addition , our new data provide additional epochs for variability analysis in these regions with the existing observations ( e.g. , * ? ? ?
our study of the gc region complements other efforts to understand galactic star forming regions , such as orion ( e.g. , * ? ? ?
* ) , the carina nebula ( e.g. , * ? ? ?
* ) , ngc 3603 @xcite , and westerlund 1 ( e.g. , * ? ? ?
it is crucial to compare the time scales and efficiency with which stars form in the gc to those in other regions of the galactic disk in order to gain a detailed understanding of the interplay between massive stars and the ism .
ultimately , understanding the gas removal , via the formation of massive stars and their impact on the ism or other processes ( e.g. , type ia supernova - driven galactic bulge wind ; * ? ? ?
* ; * ? ? ?
* ) , will help to determine why some host smbhs accrete at rates near their eddington limits , while others , including sgr a * , are at rates orders of magnitude lower .
we gratefully acknowledge the support of the staff at stsci for implementing the survey and for helping in the data reduction and analysis . support for program hst - go-11120 was provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas 5 - 26555 .
[ lastpage ] 99 arendt r. g. et al . , 2008 ,
apj 682 , 384 blum r. d. et al . , 2001 , aj , 122 , 1875 chuss d. t. et al . , 2003 ,
apj , 599 , 1116 cotera a. s. et al . , 1996 , apj , 461 , 750 cotera a. s. et al . , 1999 ,
apj , 510 , 747 cotera a. et al . , 2006 ,
jphcs , 54 , 183 cotera a. et al . , 2009 , in preparation diolaiti e. et al . , 2000 , , in adaptive optical systems technology , ed .
p. l. wizinowich , proc .
spie , 4007 , 879 dong h. , et al .
, 2009 , in preparation figer d. f. et al . , 1999 ,
apj , 525 , 759 figer d. f. et al . , 2002 ,
apj , 581 , 258 feigelson e. d. et al . , 2005 , apjs , 160 , 379 genzel r. et al . , 2003 ,
apj , 594 , 812 ghez a. m. et al . , 2008 , apj , 689 , 1044 hummer d. g. , & storey p. j. , 1987 , mnras 224,801 kim s. s. , morris m. & lee h. m. , 1999 , apj 525 , 228 kim s. s. ,
figer d. f. , lee h. m. , & morris m. , 2000 , apj 545 , 301 kim s. s. , & morris m. , 2003 , apj 597 , 312 kim s. s. , figer d. f. , & morris m. , 2004 , apj 607 , l123 lang , c. c. , goss w. m. , & wood d. o. s. , 1997 , apj , 474 , 275 lang c. c. , goss w. m. , & morris m. , 2001 , aj , 121 , 2681 li z. y. , wang q. d. , & wakker b. , 2009 , mnras , 397 , 148 mauerhan j. , muno m. , & morris m. r. , 2007 , apj , 662 , 574 mauerhan j. et al . , 2009 , apj , submitted morris m. , 1993 , apj , 408 , 496 morris m. , & serabyn e. , 1996 , ara&a , 34 , 645 muno m. et al . , 2006a , apj , 638 , 183 muno m. p. et al . , 2006b , apj , 650 , 203 muno m. p. et al . , 2009 , apjs , 181 , 110 nishiyama s. et al . , 2009 ,
apj . , 690 , 1648 philipp s. et al . , 1999 ,
a&a , 348 , 768 portegies zwart s. f. et al . , 2002 , apj , 565 , 265 portegies zwart s. f. , mcmillan s. l. w. , & gerhard o. , 2003 , apj , 593 , 352 schultheis m. et al . , 2009 , a&a , 495 , 157 scoville n. , et al . , 2003 , apj , 594 , 294 scowen et al . , 1998 , aj , 116 , 163 serabyn e. , shupe d. , & figer d. f. , 1998 , nature , 394 , 448 simpson , j. p. , et al . , 1997 , apj , 487 , 689 smith n. stassun k. g. , & bally j. , 2005 , apj , 129 , 888 stolovy s. et al . , 2006 , jphcs , 54 , 176 stolte a. et al . , 2004 , aj , 128 , 765 stolte a. et al . , 2006 ,
aj , 132 , 253 stolte a. et al . , 2008 , apj , 675 , 1278 tang s. k. et al . , 2009 , mnras , 392 , 77 tapia m. et al . , 2006 , mnras , 367 , 513 wang q. d. , gotthelf e. , & lang c. c. , 2002 , nature , 415 , 148 wang q. d. , dong h , & lang c. c. , 2006 , mnras , 371 , 38 yusef - zadeh f. , morris m. , & chance d. , 1984 , nature , 310 , 557 yusef - zadeh f. et al .
, 2002 , apj , 570 , 665 | we have recently carried out the first wide - field hydrogen paschen-@xmath0 line imaging survey of the galactic center ( gc ) , using the nicmos instrument aboard the hubble space telescope .
the survey maps out a region of @xmath1 ( @xmath2 ) around the central supermassive black hole ( sgr a * ) in the 1.87 and 1.90 narrow bands with a spatial resolution of @xmath3 pc ( 02 fwhm ) at a distance of 8 kpc . here
we present an overview of the observations , data reduction , preliminary results , and potential scientific implications , as well as a description of the rationale and design of the survey .
we have produced mosaic maps of the paschen-@xmath0 line and continuum emission , giving an unprecedentedly high resolution and high sensitivity panoramic view of stars and photo - ionized gas in the nuclear environment of the galaxy .
we detect a significant number of previously undetected stars with paschen-@xmath0 in emission .
they are most likely massive stars with strong winds , as confirmed by our initial follow - up spectroscopic observations .
about half of the newly detected massive stars are found outside the known clusters ( arches , quintuplet , and central ) .
many previously known diffuse thermal features are now resolved into arrays of intriguingly fine linear filaments indicating a profound role of magnetic fields in sculpting the gas . the bright spiral - like paschen-@xmath0 emission around sgr a * is seen to be well confined within the known dusty torus . in the directions roughly perpendicular to it ,
we further detect faint , diffuse paschen-@xmath0 emission features , which , like earlier radio images , suggest an outflow from the structure .
in addition , we detect various compact paschen-@xmath0 nebulae , probably tracing the accretion and/or ejection of stars at various evolutionary stages .
multi - wavelength comparisons together with follow - up observations are helping us to address such questions as where and how massive stars form , how stellar clusters are disrupted , how massive stars shape and heat the surrounding medium , how various phases of this medium are interspersed , and how the supermassive black hole interacts with its environment .
[ firstpage ] galaxy : center ; interstellar medium ( ism ) ; infrared : stars ; circumstellar matter ; stars : formation |
in the following , we give the series expansions in the different phases for the ground - state energy per plaquette @xmath72 and the quasiparticle gaps @xmath91 , for positive and negative signs of the dimensionless parameter @xmath92 respectively . for the sake of clarity , instead of fractions
, we give below the numerical values of the coefficients with 16 digits . for @xmath15 ,
the ground state is a nontrivial topological state without any flux in the plaquettes and elementary excitations are obtained by adding fluxes in plaquettes .
these elementary excitations can be either @xmath34-flux or @xmath35-flux although , on a torus , only single @xmath35-flux are allowed because of the fusion rules .
thus , at order 0 , the ground - state energy per plaquette is @xmath93 and the elementary gap is @xmath94 . in the limit @xmath95 , we computed @xmath72 up to order @xmath96 using operator perturbation theory @xcite , whereas @xmath97 were obtained up to order @xmath98 using perturbative continuous unitary transformations @xcite .
@xmath99\\ \nonumber\\ \delta^+_\sigma
/ j_{\rm{p}}= & \ , 1 - 1.5\ , t-1.875\ , t^2 - 2.8125\ , t^3 - 7.09375\ , t^4 -16.52335611979167\ , t^5 - 48.05825297037760\ , t^6 \nonumber\\ & -133.1833968692356\ , t^7 - 409.2101262765166\ , t^8 - 1231.461932247098\ , t^9 - 3900.735208706145\ , t^{10},\nonumber \displaybreak[0]\\
\nonumber\\ \delta^+_\psi / j_{\rm{p}}= & \ , 1 - 1.5\ , t-1.875\ , t^2 - 2.8125\ , t^3 - 7.09375\ , t^4 - 16.523681640625\ , t^5 - 48.06215243869358\ , t^6 \nonumber\\ & -133.1707469092475\ , t^7 - 409.2326163570086\ , t^8 - 1231.410148056503\ , t^9 - 3900.790315182440\ , t^{10},\nonumber \displaybreak[0]\\ \nonumber\\ \delta^-_\sigma / j_{\rm{p}}= & \ , 1 + 0.75\ , t+0.09375\ , t^2 + 0.421875\ , t^3 + 0.36962890625\ , t^4 + 0.6806233723958333\ ,
t^5 + 1.173678080240885\ , t^6\nonumber \nonumber\\ & + 2.567433410220676\ , t^7 + 5.445718276795046\ , t^8 + 12.50392429651884\ , t^9 + 29.51492451663127\ , t^{10},\nonumber \displaybreak[0]\\ \nonumber\\ \delta^-_\psi / j_{\rm{p}}=&\ , 1 + 0.75\ , t+0.09375\ , t^2 + 0.421875\ , t^3 + 0.36962890625\ , t^4 + 0.6807861328125\ , t^5 + 1.169758266872830 \ , t^6\nonumber \nonumber\\ & + 2.556954012976752\ , t^7 + 5.404811399512821\ , t^8 + 12.36748536516701\ , t^9 + 29.04858138563500\ , t^{10}.\nonumber\end{aligned}\ ] ] for @xmath25 , the ground state is the unique polarized state made of @xmath0-links and the elementary excitation consists in loops of six @xmath52-links or six @xmath2-links around one plaquette .
thus , at order 0 , the ground - state energy per plaquette is @xmath100 and the gap @xmath101 . in the limit @xmath102
, we computed @xmath72 up to order @xmath103 using a partitioning technique provided by lwdin @xcite and the low - energy gap @xmath104 were obtained up to order @xmath105 using operator perturbation theory @xcite .
@xmath106\\ \nonumber\\ \delta^+ / j_{\rm{e}}= & \ , 6 -0.5 \,t^{-1}-0.05 \ , t^{-2}-1.510416666666667\cdot 10^{-2 } \ , t^{-3}-2.042258522727273\cdot 10^{-3 } \ , t^{-4 } \nonumber\\ & -2.409132234394131\cdot 10^{-4 } \ , t^{-5}-7.207724010454154\cdot 10^{-5 } \
, t^{-6 } -1.073283118010885\cdot 10^{-5 } \
, t^{-7 } \nonumber\\ & -2.421548499055980 \cdot 10^{-6 } \ , t^{-8},\nonumber \displaybreak[0]\\ \nonumber\\ \delta^- / j_{\rm{e}}= & \ , 6 + 0.25\ , t^{-1}-0.04375 \ , t^{-2}-1.223958333333333\cdot 10^{-2 } \ , t^{-3}-1.817215119949495\cdot 10^{-3 } \ , t^{-4}\nonumber\\ & -3.438872179417279\cdot 10^{-4 } \ , t^{-5}-8.285969963376602\cdot 10^{-5 } \ , t^{-6}-1.826013298799764\cdot 10^{-5 } \ , t^{-7 } \nonumber\\ & -3.800686519319271\cdot 10^{-6 } \ , t^{-8}.\nonumber\end{aligned}\ ] ]
this supplementary material describes the numerical results obtained for the effective quantum dimer model ( qdm ) describing the string - net model [ see eq .
( 6 ) in the manuscript ] in the vicinity of @xmath27 .
this effective model is a hard - core quantum dimer model on the hexagonal lattice similar to that discussed in ref . @xcite but with a different potential term .
we performed some lanczos diagonalizations on finite - size clusters with periodic boundary conditions up to . in this largest system ,
the total hilbert space dimension is 1.648.213.392 .
the size and shape of the clusters we studied are given in tab .
[ tab : clusters ] .
all these clusters are compatible with the @xmath68 unit cell of the star crystal ( sc ) and plaquette crystal ( pc ) depicted in fig .
[ fig : sc_pc ] that turn out to be relevant for the original string - net model . .size
( @xmath17 ) and shape of the different clusters used in this study .
each cluster is defined by the two vectors @xmath107 and @xmath108 whose coordinates @xmath109 $ ] and @xmath110 $ ] are written in the basis @xmath111 .
when the coordinates @xmath110 $ ] of @xmath108 are not specified , they are defined by a @xmath39 rotation of @xmath107 ( @xmath112 and @xmath113 ) ensuring the @xmath114 rotation symmetry of the cluster .
periodic boundary conditions are imposed .
the cluster corresponding to @xmath115 plaquettes is displayed for illustration . [ cols="^,^,^",options="header " , ] we diagonalized the hamiltonian in each topological sector .
for dimer coverings on bipartite lattices , each topological sector is labeled by two winding numbers @xmath116 and @xmath117 . here , @xmath116 ( @xmath117 ) is defined as the `` electric flux '' crossing a loop winding around the torus in the @xmath107 ( @xmath108 ) direction .
the `` electric '' field @xmath118 is defined on each bond @xmath119 of the hexagonal lattice by the following rules : a ) we assume that @xmath120 belongs to the even sublattice if not , use @xmath121 ; b ) @xmath122 if @xmath119 is occupied by a dimer ; c ) @xmath123 if @xmath119 is empty ( no dimer ) .
these rules ensure that @xmath124 is divergence - free ( each site touches one dimer and two empty bonds ) and the flux crossing a closed line is therefore invariant under local moves of this line .
these winding numbers are also conserved under any local dimer move . with the definition above , the sc and the pc found in ref .
@xcite belong to the @xmath125 sector .
note that the staggered crystal also discussed in ref .
@xcite has a macroscopic electric flux , _ i.e. _ , winding numbers proportional to the linear dimensions of the system .
the qdm hamiltonian considered here is defined by @xmath126{./figures / flip_plaq1.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq2.pdf}\end{array } \hspace{-1mm}\right| + \left|\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq2.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq1.pdf}\end{array } \hspace{-1mm}\right| \big ) + v\ : \left|\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / diag_plaq.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / diag_plaq.pdf}\end{array } \hspace{-1mm}\right| \label{eq : hameff}.\ ] ] the string - net model in the vicinity of @xmath127 ( @xmath64 ) is described by @xmath60 which is related to @xmath128 with @xmath129 and @xmath130 while , for @xmath79 ( @xmath131 ) , one must consider the parameters @xmath58 and @xmath59 .
when the cluster has an even number of hexagons , one can redefine the signs of dimer configurations ( unitary transformation ) to change @xmath62 into @xmath132 ( @xmath63 unchanged ) .
however , for systems with an odd number of hexagons , the symmetry @xmath133 is broken . in order to compare various sizes reliably , we set @xmath58 in the following except when focussing on the special case @xmath130 . the evolution of the low - energy spectrum is shown in fig . [ fig : sp ] for three different system sizes ( @xmath135 and 57 ) . for @xmath134
, we observe an almost perfect degeneracy of the two lowest eigenvalues , corresponding to a three - fold degenerate ground state in the thermodynamic limit ( since one state belongs to a two - dimensional irreducible representation ) .
in fact , the spectrum at @xmath130 appears to be smoothly connected to the spectrum at @xmath136 where ground states are scs .
one can check in particular that the energy splitting @xmath137 of this multiplet vanishes as @xmath138 ( @xmath139 is a constant ) at @xmath130 when increasing the system size ( right panel of fig .
[ fig : scaling_v-1 ] ) .
by contrast , the excitation gap @xmath140 above the ground - state multiplet ( middle panel of fig .
[ fig : scaling_v-1 ] ) converges to a finite value ( @xmath141 ) in the thermodynamic limit .
we note that its value is close to the perturbative result , @xmath142 , obtained at the order 2 in a @xmath69 expansion [ see eqs .
( 7 - 8 ) in the manuscript ] .
all these data unambiguously show that the system is indeed in a sc phase at @xmath130 , with a large excitation gap and a short correlation length .
( left pannel ) and the two lowest excitation energies @xmath140 ( middle panel ) and @xmath137 ( log - normal plot in the right panel ) , for @xmath129 and @xmath130 .
@xmath140 is finite in the thermodynamic limit while @xmath137 corresponds to a vanishingly small splitting of the ground - state multiplet ( spontaneous symmetry breaking sc phase ) .
dotted green lines indicate the order 2 perturbative results in a @xmath69 expansion as given in eqs .
( 7 - 8 ) in the manuscript .
the red line is a fit showing the exponential decrease of @xmath137 with the system size .
, width=604 ] and @xmath140 at @xmath81 ( left ) and @xmath59 ( right ) .
superscripts @xmath12 or @xmath145 stands for odd or even plaquette numbers .
monte - carlo simulations @xcite have concluded that the system is in a pc phase at @xmath81 .
in such a situation the first gap @xmath137 must go to zero exponentially in the thermodynamic limit ( finite ground - state degeneracy due to the spontaneous symmetry breaking ) while the second gap @xmath140 remains finite ( local excitation of the crystal ) .
when the linear size of the system is smaller than the correlation length we have a apparent @xmath146 scaling ( see linear fits ) .
finite - size corrections are expected to become exponentially small in @xmath17 for larger systems .
although the available sizes do not allow to definitely conclude that the situation is the same at @xmath59 , we note a rather similar behavior for @xmath81 and @xmath59.,width=529 ] at @xmath81 , @xmath128 coincides with the hamiltonian studied in ref .
@xcite . using monte carlo simulations
, it was shown that the ground state of this hamiltonian with only the kinetic - energy term realizes a pc that has the same quantum numbers as those of the sc discussed above ( same broken lattice symmetries ) @xcite .
the pc is however `` softer '' , with a smaller excitation gap and a longer correlation length than in the sc phase at @xmath130 .
this can be seen from the finite - size scaling of the ground - state energy splitting ( @xmath137 ) and from the scaling of the excitation gap ( @xmath140 ) at @xmath81 .
the right panel of fig .
[ fig : scaling_v0_v1 ] shows the finite - size behavior of two lowest excitation energies with an apparent @xmath146 scaling indicating that the available system sizes are only comparable or smaller than the correlation length of the crystal .
we also note that the transition between the sc phase to the pc phase ( presumably occuring in the interval @xmath148 ) is accompanied by a sudden increase in the finite - size splitting @xmath137 of the ground - state multiplet ( see fig .
[ fig : sp ] ) .
this is also naturally explained by larger quantum fluctuations and a larger correlation length in the pc phase .
as far as the string - net model is concerned , we are interested in the point @xmath59 .
the spectra show a smooth evolution from @xmath81 ( where the plaquette long - range order was found on large systems @xcite ) to @xmath59 .
the spectra in the zero - winding sector are quantitatively very similar at @xmath81 and @xmath59 .
_ from these observations we argue that the model is also likely in the same pc phase at @xmath59_. however , we note that with the system sizes available in exact diagonalization , the scaling of the gaps does not allow to safely conclude that the excitation gap is finite at this point ( see left panel of fig .
[ fig : scaling_v0_v1 ] ) .
in addition , the apparent @xmath146 scaling should likely cross over to an exponential behavior on larger system once the correlation length is attained .
an alternative scenario would be a critical gapless state .
this is very unlikely since all qdm studied so far in two dimensions are gapped for generic values of the parameters and no mechanism producing an extended gapless _ phase _ in these models ( as opposed to critical _ points _ such as the rohsar - kivelson point ) is known .
increasing @xmath63 beyond @xmath59 leads to a ( first - order ) transition towards a phase where the ground state is no longer in the zero - winding sector . the position @xmath150 of this level crossing is shown in fig .
[ fig : vc ] , where the largest system sizes indicate that the transition takes place around @xmath151 .
this phase is not of direct interest for the original string - net model and will be analyzed in a future work . | we consider the string - net model on the honeycomb lattice for ising anyons in the presence of a string tension .
this competing term induces a nontrivial dynamics of the non - abelian anyonic quasiparticles and may lead to a breakdown of the topological phase . using high - order series expansions and exact diagonalizations
, we determine the robustness of this doubled ising phase which is found to be separated from two gapped phases .
an effective quantum dimer model emerges in the large tension limit giving rise to two different translation symmetry - broken phases .
consequently , we obtain four transition points , two of which are associated with first - order transitions whereas the two others are found to be continuous and provide examples of recently proposed bose condensation for anyons .
more than 20 years after its discovery @xcite , topological quantum order remains one of the most fascinating fields of condensed matter physics .
topologically ordered systems are characterized by several features such as , e.g. , the topological degeneracy , exotic braiding statistics , or long - range entanglement ( see ref .
@xcite for a recent review ) .
contrary to conventional phases , topological phases can not be described by a local order parameter so that the landau - ginzburg theory can not be used to investigate transitions between them . in this context
, new tools have been developed to understand transition mechanisms . among them
, an appealing approach relying on condensation of bosonic quasiparticles @xcite has been proposed to determine some possible connections between different phases .
we refer the interested readers to refs .
@xcite for concrete examples in lattice models and to refs .
@xcite for more mathematical considerations .
however , a complete description of topological phase transitions is still missing and , in particular , a classification of universality classes for the critical properties is still to be established .
from that respect , it seems essential to study microscopic models in order to explore possible scenarios . in this rapid communication
, we analyze the zero - temperature phase diagram of the string - net model @xcite defined on the honeycomb lattice with ising anyons in the presence of a string tension .
first , we give some properties of the unperturbed ising string - net model and we discuss several limiting cases allowing for a qualitative understanding of the phase diagram . to go beyond , we compute high - order series expansions of the low - energy spectrum in two limiting cases that we compare with exact diagonalization ( ed ) results .
apart from a trivial ( polarized ) phase and the doubled ising ( dising ) topological phase , we find two different translation symmetry - broken phases emerging from an effective quantum dimer model whose analysis is given in ref .
@xcite .
furthermore , our results also suggest the possibility of universality classes associated with the condensation of ising quasiparticles . _
hilbert space_. microscopic degrees of freedom of the string - net model are defined on the edges of a trivalent graph @xcite . for the ising theory considered thereafter , they can be in three different states @xmath0 , @xmath1 , and @xmath2 .
the hilbert space @xmath3 is then defined by the set of states that satisfy the so - called branching rules ( at each vertex ) stemming from the su@xmath4 fusion rules @xmath5 for any trivalent graph with @xmath6 vertices , the dimension of the hilbert space is then given by @xcite @xmath7 _ model_. let us consider the following hamiltonian @xmath8 first introduced in refs .
@xcite .
the first term is the string - net hamiltonian introduced by levin and wen @xcite .
it involves the projector @xmath9 onto states with no flux @xmath10 through .
the second term is diagonal in the canonical basis introduced above since @xmath11 is the projector onto state @xmath0 on edge @xmath12 .
this latter term is a string tension since it breaks the topological properties of the ground state described in ref .
@xcite . without loss of generality , we set @xmath13 and @xmath14 . for @xmath15 , the system is , by construction , in a doubled ( achiral ) ising topological phase , dubbed dising in the following @xcite . consequently , the degeneracy of the eigenstates depends on the graph topology .
for instance , the degeneracy of the @xmath16 energy level ( ) on a torus with @xmath17 plaquettes is @xcite @xmath18,\ ] ] where the binomial coefficient results from the different ways to choose @xmath19 plaquettes carrying the flux excitations among @xmath20 . in particular
, one finds @xmath21 ground states that are labeled by the ( trivalued ) flux contained in each of the two non - contractible loops of the torus @xcite .
note that , using the euler - poincar relation for a trivalent graph on this genus - one surface ( @xmath22 ) , it is easy to check that @xmath23 matches with eq .
( [ eq : dimh ] ) . one goal of the present work is to analyze the stability of this topological phase when @xmath24 is switched on as well as to characterize the transition between various phases .
indeed , for @xmath25 , the ( unique ) ground state is the state where all edges are in the state @xmath0 .
thus , there must be at least one phase transition in the range @xmath26 $ ] . for @xmath27 ,
the ground state is infinitely many degenerate in the thermodynamical limit so that ( at least ) one phase transition is expected in the range @xmath28 $ ] .
finally , for @xmath29 , the ground - state degeneracy ( on a torus ) is given by @xmath30 so that transitions must also occur in the range @xmath31 $ ] . in the following , we consider the simplest two - dimensional trivalent graph , namely , the honeycomb lattice .
+ _ the simple " case : @xmath32$]_. to determine the boundaries of the dising topological phase ( ) , we computed the ground - state energy as well as the quasiparticle gaps by means of high - order series expansions in powers of @xmath33 ( lengthy expressions are given in ref .
@xcite ) using various techniques @xcite . in the vicinity of @xmath15
, one must make the distinction between two different low - energy gaps corresponding to quasi-@xmath34 and quasi-@xmath35 excitations .
these excitations are usually referred to as @xmath36 and @xmath37 in the literature ( see for instance ref .
@xcite ) but , for simplicity , we will adopt here a quasiparticle language keeping in mind that these are achiral objects .
contrary to quasi-@xmath35 excitations , a single quasi-@xmath34 excitation can not exist on a compact surface such as the torus because of the branching rules @xcite .
using standard extrapolation methods , we determined the points where these gaps vanish and thus established the stability range of the dising phase .
however , if level crossings due to higher - energy levels are present , a first - order transition may also arise and can not be captured by our perturbative approach that only deals with low - energy states . to check the validity of the conclusions drawn from the series expansions , we performed ed of @xmath38 using periodic boundary conditions and systems with unit vectors of equal norms forming an angle of @xmath39 . on the torus
, the hamiltonian @xmath38 can be split into topological different sectors that must not be confused with the nine flux sectors discussed previously for @xmath15 . indeed , for any @xmath40
, branching rules impose that a @xmath1 link is always connected to a single @xmath1 link . in particular , there exist non - contractible loops of @xmath1 links enclosing the torus .
fusion rules impose that @xmath38 only conserves the parity of the number of such loops and , since there are two independent non - contractible loops on the torus , one has @xmath41 different sectors for any @xmath40 .
we display in fig .
[ fig : gse ] a comparison between the ed results and the series expansions for the ground - state energy performed around @xmath15 ( red ) and @xmath25 ( blue ) . as can be seen , at each order ,
the series intersect at two different points .
similarly , we show in fig .
[ fig : gap1 ] the results for the low - energy gap that intersect in a unique point .
after extrapolations , using the same analysis as in ref .
@xcite , we found that all these crossing points converge towards a unique value defining a second - order transition point at @xmath42 .
this point also matches with the position of the infinite - size extrapolation of the gap minimum as well as the minimum of @xmath43 computed from ed .
this critical point separates the dising phase originating from @xmath15 from the polarized ( non - topological ) phase near @xmath25 .
we stress that the relevant low - energy gap in the ed must be interpreted as a single quasiparticle gap associated with quasi-@xmath35 excitations .
the corresponding energy level is indeed adiabatically connected to @xmath44 at @xmath15 . within the topological symmetry - breaking formalism proposed in ref .
@xcite , if quasi-@xmath35 excitations condense while quasi-@xmath34 excitations remain gapped , one should switch towards another topological phase , which is not the case here . in this framework ,
the only possibility to enter a non - topological phase is that the quasi-@xmath34 gap also vanishes at @xmath45 .
as explained above , this gap can not be observed in ed on a torus but it can be computed perturbatively since this approach is independent of the surface topology in the thermodynamical limit .
series expansions of both quasiparticle gaps ( quasi-@xmath35 and quasi-@xmath34 ) are given in ref .
@xcite .
as can be checked explicitly , they are strictly identical up to order @xmath46 and differ beyond .
however , the sign of the ( tiny ) difference between both gaps changes at each order , which is compatible with a simultaneous vanishing of these gaps at @xmath45 .
note that we performed similar calculations for the ladder geometry and we found that both gaps are identical at all orders we .
for @xmath47 , we did not find any indication of a transition but , as for the fibonacci theory @xcite , the first derivative of the ground - state energy per plaquette @xmath48 displays a jump at @xmath29 indicating a first - order transition .
thus , the trivial phase originating from @xmath25 extends from @xmath45 ( second - order transition point ) to @xmath49 ( first - order transition point ) .
+ _ the original " case : @xmath50$]_. for @xmath27 , the ground - state manifold is spanned by all states minimizing the number of edges in state @xmath0 . interestingly ,
for the ising theory , fusion ( branching ) rules allow some states without any @xmath0 bond provided each vertex touches exactly one @xmath51 and two @xmath52 .
these constraints are nothing but those of hard - core dimer coverings of the hexagonal lattice if the state @xmath51 is viewed as a bond occupied by a dimer . "
the exponential ground - state degeneracy at @xmath27 prevents a simple series expansion around this point . in addition , the alternating signs in the series around @xmath53 prevents from an analysis similar to the one used in the range @xmath26 $ ] .
however , a close inspection of the gap series expansion around @xmath15 indicates a transition point near @xmath54 . as can be seen in fig .
[ fig : gap2 ] , ed results are consistent with a unique phase transition in the range @xmath28 $ ] , but accessible sizes are definitely too small to characterize properly the phase for @xmath55 as well as the nature of the transition . to gain a deeper understanding of this region , we derived the low - energy effective theory near @xmath27 , at leading order , by considering the effect of the string - net hamiltonian on the infinitely many degenerate ground - state manifold of the unperturbed ( @xmath56 ) problem .
the effective hamiltonian can be written in the following form @xmath57{./figures / flip_plaq1.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm
} \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq2.pdf}\end{array } \hspace{-1mm}\right| + \left|\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq2.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / flip_plaq1.pdf}\end{array } \hspace{-1mm}\right| \big ) + \nonumber \\ & v&\:\ : \left|\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / diag_plaq.pdf}\end{array } \hspace{-1mm}\right\rangle \left\langle\hspace{-1 mm } \begin{array}{c}\includegraphics[width=0.5cm]{./figures / diag_plaq.pdf}\end{array } \hspace{-1mm}\right| + \mathds{1 } \big ] , \label{eq : hameff}\end{aligned}\ ] ] with @xmath58 and @xmath59 . in this representation , a red ( blue )
link corresponds to a @xmath52 ( @xmath51 ) state .
we emphasize that @xmath60 is only valid to compute the correction of order 1 ( @xmath61 ) to the spectrum of @xmath38 near @xmath27 .
this effective hamiltonian consists of two terms : a kinetic term @xmath62 acting on flippable " plaquettes and a potential term @xmath63 proportional to the number of empty plaquettes ( without dimers ) .
this model looks very similar to the famous quantum dimer model introduced by rokhsar and kivelson on the square lattice @xcite and later studied on the honeycomb lattice @xcite .
the only difference between both models comes from the potential term that , in the rokhsar - kivelson model , is proportional to the number of flippable plaquettes . to our knowledge
, the hamiltonian ( [ eq : hameff ] ) has yet to be studied and can not be solved exactly for arbitrary couplings .
however , for @xmath64 , it is possible to infer its low - energy properties by considering the limit @xmath65 while keeping in mind that @xmath66 in our problem . for @xmath67 ( and @xmath64 ) ,
the energy is minimized by maximizing the number of empty plaquettes ( without dimers ) . in the thermodynamical limit , there are three possible ground states satisfying this constraint so that the system is in the so - called @xmath68 star crystal ( sc ) phase @xcite ( see ref .
@xcite for an illustration ) .
first excited states are obtained by flipping one plaquette in one of these ground states .
the kinetic term @xmath62 induces quantum fluctuations that can be captured using perturbation theory . at order 2 in @xmath69 , the ground - state energy per plaquette
is given by @xmath70 whereas the low - energy gap reads @xmath71 of course , one could reach higher orders in this @xmath69 expansions but it is not of crucial importance for the present study .
indeed , these expressions already provide a very good approximation of @xmath72 and @xmath73 for @xmath66 since , as shown in ref .
@xcite , they only differ from the infinite - size values extrapolated from ed results by less than @xmath74 and @xmath75 , respectively . in the range @xmath76 , we thus find a gapped translation symmetry - broken phase with a three - fold degenerate ground state .
furthermore , the three momenta ( center and the two corners of the hexagonal brillouin zone ) of the ground states in the sc coincide with the locations of the minima of the excitation gap in the dising phase .
this suggests that the transition from the topological phase to the crystal also corresponds to a simultaneous condensation of the anyonic quasiparticles .
furthermore , momenta of these scs also minimize the dispersion in the range @xmath77 which is compatible with a second - order transition .
this is an example of a continuous phase transition between a non - abelian topological phase and a non topological ( translation ) symmetry - broken phase . +
_ the tricky " case : @xmath78$]_. obviously , the effective model ( [ eq : hameff ] ) is also valid for @xmath79 but , contrary to the case @xmath64 where a sc is favored ( see discussion above ) , the ground state is infinitely - many degenerate at @xmath67 so that it is difficult to consider a perturbative @xmath69 expansion .
consequently , we performed ed of @xmath60 up to relatively large system sizes ( @xmath80 ) and we found that the point @xmath58 , @xmath59 lies in the same phase as the point @xmath58 , @xmath81 ( see ref .
@xcite ) . as discussed in ref .
@xcite for @xmath81 , the ground state displays the so - called plaquette order that breaks the translational symmetry .
thus , for @xmath79 , we expect a non topological ordered ( gapped ) plaquette crystal ( pc ) phase with a three - fold degenerate ground state .
however , in the absence of perturbative analysis near @xmath29 and because of important variations with the system size of the ed results ( at least up to @xmath82 which is our current limit ) , we did not succeed in characterizing the whole interval @xmath83 $ ] .
however , we observe that @xmath48 displays a jump for @xmath29 and @xmath27 so that first - order transitions occur at these points . _ summary and outlook_. a sketch of the phase diagram gathering all informations discussed throughout this rapid communication is given in fig .
[ fig : diagram ] .
one of the main results is the possibility to condense simultaneously quasi-@xmath35 and quasi-@xmath34 excitations at the critical point @xmath84 , unveiling a likely new universality class .
unfortunately , in the absence of an alternative description , it is difficult to predict the associated critical exponents . setting @xmath85
, a standard ed data collapse analysis gives results that obey the hyperscaling relation for @xmath86 and @xmath87 , and the resulting specific - heat exponent is @xmath88 .
one can also compute directly the exponent @xmath89 using different _
d_log pad approximants of the gap series and we found values in the range @xmath90 $ ] .
this rather broad range clearly indicates the lack of precision of such an approach for order 10 series , but suggests that this phase transition may belong to a new universality class .
we hope that the present work will stimulate further studies to deepen our understanding of the topological phase transitions .
+ m.d.s .
acknowledges financial support from the dfh - ufa . k.p.s . acknowledges esf and eurohorcs for funding through his euryi .
g.m .
is supported by a jcjc grant of the agence nationale pour la recherche ( project no . anr-12-js04 - 0010 - 01 ) .
28ifxundefined [ 1 ] ifx#1 ifnum [ 1 ] # 1firstoftwo secondoftwo ifx [ 1 ] # 1firstoftwo secondoftwo `` `` # 1''''@noop [ 0]secondoftwosanitize@url [ 0 ]
+ 12$12 & 12#1212_12%12@startlink[1]@endlink[0]@bib@innerbibempty link:\doibase 10.1103/physrevb.39.11413 [ * * , ( ) ] link:\doibase 10.1103/physrevb.40.7387 [ * * , ( ) ] link:\doibase 10.1142/s0217979290000139 [ * * , ( ) ] link:\doibase 10.1155/2013/198710 [ * * , ( ) ] link:\doibase 10.1103/physrevb.79.045316 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.102.220403 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.105.216804 [ * * , ( ) ] link:\doibase 10.1103/physrevb.84.125434 [ * * , ( ) ] link:\doibase 10.1088/1367 - 2630/14/1/015004 [ * * , ( ) ] link:\doibase 10.1103/physrevb.88.235120 [ * * , ( ) ] @noop link:\doibase 10.1007/s00220 - 013 - 1723 - 0 [ * * , ( ) ] @noop @noop
link:\doibase 10.1103/physrevb.71.045110 [ * * , ( ) ] @noop @noop link:\doibase 10.1038/nphys1396 [ * * , ( ) ] @noop @noop link:\doibase 10.1063/1.1724312 [ * * , ( ) ] \doibase doi:10.1088/0022 - 3719/10/8/031 [ * * , ( ) ] link:\doibase 10.1007/s100510050026 [ * * , ( ) ] link:\doibase 10.1103/physrevlett.110.147203 [ * * , ( ) ] @noop link:\doibase 10.1103/physrevlett.61.2376 [ * * , ( ) ] link:\doibase 10.1103/physrevb.64.144416 [ * * , ( ) ] @noop |
there are two main reasons for revisiting soft @xmath5 high energy interactions at this time .
* a. * to predict soft processes at the lhc we need a reliable model .
a detailed analysis performed in the late sixties @xcite showed that there could be different scenarios for the high - energy behaviour of the interaction amplitude .
in the ` weak coupling ' regime the total cross section @xmath6 , and in order not to violate unitarity , and to satisfy the inequality @xmath7 the triple - pomeron vertex must vanish when @xmath8 , that is @xmath9 . in this case , the large logarithm coming from the integration over the mass of the system produced in diffractive dissociation ( @xmath10 ) is compensated by the small value of the mean momentum transferred through the pomeron @xmath11 .
on the other hand , for ` strong coupling ' , with @xmath12 with @xmath13 , the inequality ( [ eq:1 ] ) is provided by a small value of the rapidity gap survival factor @xmath14 which decreases with energy .
the present diffractive data are better described within the ` strong coupling ' approach , and here we show predictions for the lhc for this scenario .
however , the possibility of the ` weak coupling ' scenario is not completely excluded yet .
therefore , it is quite important to study the different channels of diffractive dissociation at the lhc in order to reach a final conclusion and to fix the parameters of the model for high - energy soft interactions . *
b. * the second reason for studying soft interactions arises because it may not be an easy task to distinguish the production of a new object at the lhc when it is accompanied by hundreds other particles emitted in the same event . for the detailed study of the new object , @xmath15 , it may be better to select the few , very clean , events with the large rapidity gaps ( lrg ) on either side of the new object @xcite .
that is to observe the exclusive process @xmath16 .
in such a central exclusive process ( cep ) the mass of @xmath15 can be measured with very good accuracy ( @xmath17 gev ) by the missing - mass method by detecting the outgoing forward protons . moreover
, a specific @xmath18 selection rule @xcite reduces the background and also greatly simplifies the spin - parity analysis of @xmath15 . however , the cep cross section is strongly suppressed by the small survival factor , @xmath19 , of the rapidity gaps .
thus we need a reliable model of soft interactions to evaluate the corresponding value of @xmath14 .
the total and elastic proton - proton cross sections are usually described in terms of an eikonal model , which automatically satisfies @xmath20-channel elastic unitarity . to account for the possibility of excitation of the initial proton , that is for two - particle intermediate states with the proton replaced by @xmath21
, we use the good - walker formalism @xcite . already at tevatron energies the absorptive correction to the elastic amplitude , due to elastic eikonal rescattering , is not negligible ; it is about 20% in comparison with the simple one pomeron exchange .
after accounting for low - mass proton excitations ( that is @xmath21 s in the intermediate states ) the correction becomes twice larger ( that is , up to 40% ) .
next , in order to describe high - mass diffractive dissociation , @xmath22 , we have to include an extra factor of 2 from the agk cutting rules @xcite .
thus , the absorptive effects in the triple - regge domain are expected to be quite large .
the previous triple - regge analyses ( see , for example , @xcite ) did not allow for absorptive corrections and the resulting triple - regge couplings must be regarded , not as bare vertices , but as effective couplings embodying the absorptive effects . since the inelastic cross section ( and , therefore , the absorptive corrections ) expected at the lhc are more than twice as large as that observed at fixed - target and cern - isr energies , the old triple - regge vertices can not be used to predict the diffractive cross sections at the lhc .
thus , it is necessary to perform a new triple - regge analysis that includes the absorptive effects explicitly .
cross section @xmath23 data obtained in the strong triple - pomeron coupling fit , where @xmath24 .
the individual triple - regge contributions are also shown.,height=566 ] [ fig1 ] together with e.g.d.s.luna , we fitted @xcite the fixed - target fnal , cern - isr and tevatron data that are available in the triple - regge region .
we included ` @xmath25 ' , ` @xmath26 ' , ` @xmath27 ' , ` @xmath28 ' and @xmath29 contributions , assuming either the ` strong ' or ` weak ' coupling scenarios for the behaviour of the triple - pomeron vertex . to account for the absorptive corrections we used a two - channel ( good - walker ) eikonal model , which describes well the total , @xmath30 , and elastic , @xmath31 , @xmath5 and @xmath32 cross sections . in the ` strong ' coupling case , a good @xmath33dof=167/(210 - 8)=0.83
was obtained .
the quality of the description to a sample of the data can be seen in fig.1 , where the various triple - regge contributions to the cross section are also shown . in comparison with the old triple - regge analysis @xcite
, we now have about a twice larger relative contribution of the ` @xmath26 ' term .
this is mainly due to the inclusion in our analysis of the higher - energy tevatron data .
since the absorptive effects are included explicitly , the extracted values of the triple - reggeon vertices are now much closer to the _ bare _ triple - regge couplings .
in particular , the value @xmath34 is consistent with a reasonable extrapolation of the perturbative bfkl pomeron vertex to the low scale region @xcite ; here @xmath35 is the pomeron - proton coupling .
note also that these values of ` @xmath25 ' and ` @xmath26 ' vertices allow us to describe the hera data @xcite on inelastic @xmath36 photoproduction , @xmath37 , where the screening corrections are rather small .
the ` weak ' coupling scenario leads to a larger @xmath33dof=1.4 and to a worse description of the @xmath38 process at the lowest values of @xmath39 . at the lhc energy
the ` weak ' coupling fit predicts about 3 times smaller inclusive cross section @xmath40 at @xmath41 in comparison with that predicted in the ` strong ' coupling case .
since the triple - pomeron vertex ( [ eq:2 ] ) turns out to be rather large , the contribution of the so - called ` enhanced ' diagrams , with a few vertices , is not negligible .
moreover , we can not expect that more complicated multi - pomeron interactions , driven by the @xmath42 vertices , which describe the transition of @xmath43 to @xmath44 pomerons , will not affect the final result .
it looks more reasonable to assume that @xmath45 than to assume that @xmath46 for any @xmath47 .
thus we need a model which accounts for the possibility of multi - pomeron interactions ( with arbitrary @xmath43 and @xmath44 ) . while the eikonal formalism describes the rescattering of the incoming fast particles , the enhanced multi - pomeron diagrams represent the rescattering of the intermediate partons in the ladder ( feynman diagram ) which describes the pomeron - exchange amplitude .
the ladder - type pomeron amplitude may be generated by the evolution equation ( in rapidity , @xmath48 , space ) @xmath49 where @xmath50 is the two - dimensional vector in impact parameter space and @xmath51 is the slope of the pomeron trajectory .
@xmath52 is the probability to emit new intermediate partons within unit rapidity interval ; the pomeron intercept @xmath53 .
the solution of ( [ eq:3 ] ) , @xmath54 , is the opacity ( at point @xmath55 ) , corresponding to the incoming particle placed at @xmath56 and @xmath57 .
it looks natural to account for absorptive effects by including on the r.h.s . of ( [ eq:3 ] )
the factor @xmath58 , where the subscripts @xmath59 and @xmath60 denote the opacities of the beam and the target incoming hadrons .
we thus have the evolution equations @xmath61 @xmath62 where here @xmath63 .
the coefficient @xmath64 accounts for the fact that the absorptive cross section for the intermediate parton , @xmath65 , may be different from that for the incoming particle ( proton ) . since the equations ( [ eq:3]-[eq:5 ] ) describe the interaction amplitude and not the cross section we have a factor @xmath66 in the exponent . in terms of multi - pomeron vertices , the absorption factors , @xmath67 , in ( [ eq:4],[eq:5 ] )
correspond to @xmath68 note that , since the intermediate parton may be absorbed by the interaction with the particles ( partons ) from the wave function of both beam or target hadron , we now need to solve the two equations , ( [ eq:4 ] ) and ( [ eq:5 ] ) .
this is done iteratively .
the resulting solution solution can then be used in the eikonal formulae to determine all soft @xmath5 interactions . in comparison with our previous model @xcite
, we now include a non - zero slope ( @xmath69 ) of the pomeron trajectory .
the incoming proton wave function is described by three good - walker eigenstates , that is we use a 3-channel eikonal for the rescattering of fast particles .
the transverse size squared of each eigenstate is proportional to the corresponding absorptive cross section ; @xmath70 .
that is we assume that the parton density at the origin is the same for each eigenstate .
the shape of the pomeron - nucleon vertex is parametrised by the form factor @xmath71 , whose fourier transform , @xmath72 , plays the role of the initial conditions for @xmath73 .
next , we allow for four different @xmath39-channel states , which we label @xmath74 : one for the secondary reggeon ( @xmath75 ) trajectory and three pomeron states ( @xmath76 ) to mimic the bfkl diffusion in the logarithm of parton transverse momentum , @xmath77 @xcite . to be precise , since the bfkl pomeron @xcite is not a pole in the complex @xmath78-plane , but a branch cut , we approximate the cut by three @xmath39-channel states of a different size
. the typical values of @xmath79 in each of the three states is about @xmath80 gev , @xmath81 gev and @xmath82 gev .
thus the system ( [ eq:4],[eq:5 ] ) is replaced by @xmath83 @xmath84 the transition factors @xmath85 were fixed by properties of the bfkl equation . in the exponents , the opacities @xmath86 ( @xmath87 )
are actually the sum of the opacities @xmath88 ( @xmath89 ) with corresponding coefficients ( see @xcite for more details ) .
clearly the number of parameters in such a model is large .
therefore , instead of a straightforward fit of the data , we adjust the majority of parameters in reasonable intervals and demonstrate that such a model allows us to reproduce all the available data on diffractive cross sections , @xmath90 .
the quality of the description is demonstrated in fig.2 , where we also present the prediction for elastic cross section at the lhc energy @xmath91 tev .
dependence of the elastic @xmath5 cross section data .
the dashed and dotted lines are the contributions at the lhc energy from the elastic scattering of the largest size ( @xmath92 ) and the smallest size ( @xmath93 ) components .
interestingly , we see that the complete set of eikonal contributions gives a smooth prediction for the elastic cross section in this @xmath39 interval at the lhc energy.,height=680 ] [ fig2 ] the values of the parameters that we use are : @xmath94 gev@xmath95 , @xmath96 gev@xmath97 , @xmath98 , @xmath99 gev@xmath95 , @xmath100 and , for each of 3 components of the pomeron , @xmath101
. the pomeron intercepts are consistent with the expectations of resummed nll bfkl which gives @xmath102 practically independent of the scale @xmath79 @xcite .
the slopes of the pomeron trajectories are : @xmath103 gev@xmath95 for the large size pomeron component , @xmath104 gev@xmath95 for the second component ( @xmath105 ) , and @xmath106 for the smallest size pomeron component .
ths model , which we have tuned to describe the available data , allows us to predict , in principle , all features of soft @xmath5 high energy interactions .
let us consider the cross section for beam particle diffractive dissociation ( with a rapidity gap up to @xmath48 ) .
first recall the usual eikonal relations @xmath107 for each impact parameter point @xmath50 , the desired cross section for single dissociation is proportional ( i ) to the elastic parton @xmath108 cross section @xmath109 ; ( ii ) to the probability to find the intermediate parton @xmath65 in the interval @xmath110 , that is @xmath111 ; ( iii ) to the amplitude @xmath86 of the parton @xmath112 beam interaction ; ( iv ) to the gap survival factor @xmath113 ( where @xmath114 ) .
the resulting cross section reads @xmath115 @xmath116 where @xmath117 ( @xmath118 ) are the coordinates in the impact parameter plane with respect to the target ( beam ) hadron .
expression ( [ eq : sd ] ) must be averaged over the components ( that is the good - walker eigenstates ) of the beam and target hadrons and summed over the different @xmath39-channel states ( @xmath119 ) with appropriate normalisation factors @xmath120 .
the energy behaviour of the diffractive cross sections , and of the multiplicities of the secondaries produced by the @xmath39-channel pomeron components of different sizes , are shown in fig.3 .
cross sections and ( c ) the cross sections of dissociation to a fixed @xmath121 state , where @xmath24 . plot ( d ) shows the parton multiplicity ( solid lines ) and the number of ` colour tubes ' ( dashed ) produced by the pomeron components of different size.,height=680 ] [ fig3 ] note that , starting with the same ` bare ' intercepts ( @xmath122 ) , after the absorptive correction , the contribution of the large - size component becomes practically flat , while the small - size contribution , which is much less affected by the absorption , continues to grow with energy .
such a behaviour is consistent with experiment , where the density of low @xmath79 secondaries is practically saturated , while the probability to produce a hadron with a large transverse momentum ( say , more than 5 gev ) grows with the initial energy .
after accounting for absorptive effects the new triple - regge analysis leads to a rather large triple - pomeron vertex @xmath123 with @xmath124 .
this indicates that in order to describe diffractive processes at the lhc we must use a model which includes multi - pomeron interactions .
we construct such a model in which the absorption of the intermediate partons is described by conventional - like factors @xmath125 . in terms of the multi - pomeron vertices
, this corresponds to @xmath126 .
the model reasonably well reproduces available diffractive data on @xmath127 using the parameters : @xmath122 , @xmath128 gev@xmath95 , @xmath98 .
it thus leads to the hope that there is a smooth matching between the perturbative qcd treatment of the pomeron and the multi - pomeron description of soft interactions .
since the screening corrections caused by the ` enhanced ' multi - pomeron diagrams , that is by the high - mass dissociation , slow down the growth of the cross section with energy , the model predicts a relatively low total cross section at the lhc energy , namely @xmath129 mb @xcite .
predictions for lhc soft interactions are shown in figs . 2 and 3
. 0 v.n .
gribov and a.a .
migdal , sov .
* 8 * 583 ( 1969 ) .
gribov and a.a .
migdal , sov .
jetp * 28 * 784 ( 1969 ) .
khoze , a.d .
martin and m.g .
ryskin , eur .
j. * c23 * 311 ( 2002 ) .
khoze , a.d .
martin and m.g .
ryskin , eur .
j. * c19 * 477 ( 2001 ) , erratum * c20 * 599 ( 2001 ) ; + v.a .
khoze , a.d .
martin and m.g .
ryskin , arxiv : hep - ph/0006005 .
good and w.d .
walker , phys .
rev . * 120 * 1857 ( 1960 ) .
abramovsky , v.n .
gribov and o.v .
kancheli , sov .
. phys . * 18 * 308 ( 1973 ) .
r.d . field and g.c .
fox , nucl .
* b80 * 367 ( 1974 ) ; + a.b .
kaidalov , v.a.khoze , yu.f .
pirogov and n.l .
ter - isaakyan , phys . lett . *
b45 * 471 ( 1974 ) ; + for a review see a.b .
kaidalov , phys . rep .
* 50 * 157 ( 1979 ) .
luna , v.a .
khoze , a.d .
martin and m.g .
ryskin , arxiv:0807.4115 .
j. bartels , m.g .
ryskin and g.p .
vacca , eur .
j. * c27 * 101 ( 2003 ) .
zeus collaboration : abstract 549 , int .
europhysics conf . on hep , aachen ,
july 2003 .
ryskin , a.d .
martin and v.a .
khoze , eur .
j. * c54 * 199 ( 2008 ) .
lipatov , sov .
jetp * 63 * 904 ( 1986 )
. v.s .
fadin , e.a .
kuraev , and l.n .
lipatov , phys .
b * 60 * ( 1975 ) 50 ; + e.a .
kuraev , l.n .
lipatov , and v.s .
fadin , zh .
* 71 * ( 1976 ) 840 [ sov
jetp * 44 * 443 ( 1976 ) ] ; _ ibid . _ * 72 * ( 1977 ) 377 [ * 45 * ( 1977 ) 199 ] ; + i.i .
balitsky and l.n .
lipatov , yad .
* 28 * ( 1978 ) 1597 [ sov
* 28 * ( 1978 ) 822 ] .
ryskin , a.d .
martin and v.a .
khoze , to be published .
v.s . fadin and l.n .
lipatov , phys . lett . *
b429 * ( 1998 ) 127 ; + g. camici and m. ciafaloni , phys . lett . * b430 * ( 1998 ) 349 ; + g.p .
salam , jhep * 9807 * 019 ( 1998 ) , act .
b30 * 3679 ( 1999 ) ; + m. ciafaloni , d. colferai and g.p .
salam , phys .
* b452 * 372 ( 1999 ) , phys . rev . *
d60 * 114036 ( 1999 ) . | we present a triple - regge analysis of the available @xmath0 high - energy data accounting for absorptive corrections .
we describe a model for high - energy soft interactions which includes the whole set of multi - pomeron ( @xmath1 ) vertices , and give predictions for the lhc . plus 2 mm minus 2 mm 23.0 cm 17.0 cm -1.0 in -42pt ippp/08/77 + dcpt/08/154 + + * soft diffraction at the lhc * v.a .
khoze@xmath2 , a.d . martin@xmath3 and m.g .
ryskin@xmath2 + @xmath3 institute for particle physics phenomenology , university of durham , durham , dh1 3le + @xmath4 petersburg nuclear physics institute , gatchina , st .
petersburg , 188300 , russia |
consider the multigraph @xmath1 with two vertices @xmath2 and three edges @xmath3 none of which are loops .
theta - curve _ is a locally flat embedding of @xmath1 in @xmath4 or in @xmath5 .
each theta - curve @xmath6 has three _ constituent knots _ : @xmath7 , @xmath8 , and @xmath9 .
given a constituent knot @xmath0 , there is exactly one arc @xmath10 of @xmath6 not contained in @xmath0 . +
a theta - curve is _ unknotted _ provided it lies on an embedded @xmath11 and is _ knotted _ otherwise .
we will use two operations on knots and theta - curves : the order-@xmath12 connected sum @xmath13 and order-@xmath14 connected sum @xmath15 .
an order-2 connected sum @xmath16 of a theta - curve @xmath17 and a knot @xmath18 is the result of deleting unknotted ball - arc pairs from each of @xmath19 and @xmath20 , and then identifying the resulting boundary spheres .
the ball - arc pair in @xmath19 must be disjoint from the vertices @xmath21 and @xmath22 . an order-3 connected sum @xmath23 of two theta - curves is the result of deleting an unknotted ball - prong neighborhood of a vertex from each theta - curve , and then identifying the resulting boundary spheres .
each of these operations yields a theta - curve in @xmath4 .
+ wolcott has shown that the order-@xmath14 connected sum is independent of the glueing homeomorphism provided one specifies the vertices at which to sum and the pairing of the arcs @xcite . the operations @xmath16 and @xmath23 are ambiguous as presented .
the former could mean up to six different theta - curves and the latter could mean up to @xmath24 different theta - curves .
in each instance below , this ambiguity is either irrelevant or is sufficiently eliminated by context .
a theta - curve @xmath6 is _ prime _ provided the following three conditions are satisfied : ( i ) @xmath6 is knotted , ( ii ) @xmath6 is not an order-@xmath12 connected sum of a nontrivial knot and a ( possibly unknotted ) theta - curve , and ( iii ) @xmath6 is not an order-@xmath14 connected sum of two knotted theta - curves .
we adopt the convention that the unknot is not prime .
+ let @xmath25 be the unit sphere .
let @xmath26 denote the orientation preserving involution of @xmath4 whose matrix is diagonal with entries @xmath27}$ ] .
note that @xmath28 is a great circle in @xmath4 and is therefore unknotted .
let @xmath29 be the group of order two . throughout this paper ,
equivariance is with respect to @xmath30 .
recall that a subset @xmath31 is _ equivariant _ provided @xmath32 ( setwise ) or @xmath33 . + suppose the theta - curve @xmath6 has an unknotted constituent knot @xmath0 and @xmath34 . by an ambient isotopy
, we can assume @xmath35 .
let @xmath36 be the double branched cover with branch set @xmath28 and such that @xmath37 .
lifting the arc @xmath10 of @xmath6 yields the knot @xmath38 in @xmath4 .
we call @xmath39 the _ lifted knot of @xmath6 with respect to @xmath0_. by stereographic projection from @xmath40 , the map @xmath41 descends to the rotation of @xmath5 about the @xmath42-axis by half a revolution .
+ moriuchi attributes the following theorem to w. thurston without proof ( * ? ? ?
4.1 ) , and moriuchi references an unpublished letter of litherland for thurston s statement of this theorem ( * ? ? ?
suppose @xmath6 has an unknotted constituent knot @xmath0 , and let @xmath39 be the lifted knot of @xmath6 with respect to @xmath0 .
the theta - curve @xmath6 is prime if and only if the lifted knot @xmath39 is prime .
our proof of the main theorem uses the equivariant dehn lemma in two places . in the proof of lemma [ lem : unknotted ] , kim and tollefson s version ( * ? ? ?
* lemma 3 ) suffices as noted by the referee . in the proof of lemma [ lem : torus ] , we use edmond s version @xcite .
it s possible that with some tinkering kim and tollefson s version suffices .
+ in the final section of this paper , we use the main theorem to prove that kinoshita s theta - curve is prime .
we also explain how primality of kinoshita s theta - curve yields an alternate proof of the irreducibility of certain tangles .
+ * acknowledgment .
* we thank the referee for simplifying the proof of lemma [ lem : unknotted ] and for several other helpful comments .
we will prove the contrapositive in both directions .
first , we observe two useful lemmas . [
lem : sum ] suppose @xmath6 , @xmath43 , and @xmath44 are theta - curves , and @xmath18 is a nontrivial knot . 1 .
[ first_part]if @xmath45 has an unknotted constituent knot @xmath0 , then @xmath46 ( that is , @xmath0 is the union of two edges in @xmath6 ) .
let @xmath39 be the lifted knot of @xmath6 with respect to @xmath0 .
then , the lifted knot of @xmath45 with respect to @xmath0 is @xmath47 .
[ second_part]if @xmath23 has an unknotted constituent knot @xmath0 , then there are unknotted constituent knots @xmath48 and @xmath49 such that the order-3 connected sum @xmath23 induces the connected sum @xmath50 .
let @xmath51 be the lifted knot of @xmath52 with respect to @xmath53 for @xmath54 .
then , the lifted knot of @xmath23 with respect to @xmath0 is @xmath55 .
for the first claim in [ first_part ] , label the edges of @xmath6 so that the sum @xmath45 is performed along @xmath56 and let @xmath10 be the resulting edge in @xmath45 .
the knot @xmath57 is the connected sum of the knots @xmath8 and @xmath58 .
as @xmath58 is nontrivial and nontrivial knots do not have inverses under connected sum , we see that the knot @xmath57 is nontrivial .
similarly , the knot @xmath59 is nontrivial .
hence , @xmath9 must be an unknot as desired .
the first claim in [ second_part ] follows similarly .
the remaining claims follow from the definitions of order-2 and order-3 connected sum and from the definition of the double branched cover @xmath36 . for the remainder of this section , we assume @xmath6 is a theta - curve with unknotted constituent knot @xmath35 and @xmath34 .
let @xmath39 be the lifted knot of @xmath6 with respect to @xmath0 .
[ lem : sphere ] if @xmath60 is an equivariant 2-sphere in @xmath4 that meets @xmath0 in exactly two points , then @xmath61 is an embedded 2-sphere in @xmath4 transverse to @xmath0 and meeting @xmath0 in exactly two points .
as @xmath60 meets @xmath35 , equivariance implies @xmath62 .
equivariance also implies that @xmath60 is transverse to @xmath0 , @xmath61 is a closed connected surface , and @xmath61 is transverse to @xmath0 .
let @xmath63 be the equivariant annulus obtained from @xmath60 by deleting the interiors of disjoint equivariant @xmath12-disk neighborhoods of the two points @xmath64 .
the restriction of @xmath65 to @xmath66 is a double cover and @xmath61 is obtained from @xmath67 by glueing in two @xmath12-disks .
it follows that the euler characteristic of @xmath67 is 0 and the euler characteristic of @xmath61 is 2 .
hence , @xmath61 is a @xmath12-sphere .
[ lem : unknotted ] the theta - curve @xmath6 is unknotted if and only if @xmath39 is unknotted .
clearly , if @xmath6 is unknotted , then @xmath39 is unknotted .
suppose @xmath39 is unknotted .
let @xmath68 be a closed regular equivariant neighborhood of @xmath39 in @xmath4 .
as @xmath69 is a solid torus , it has compressible boundary . by the equivariant dehn lemma ( * ? ? ? * lemma 3 ) , there exists a properly embedded equivariant compressing disk @xmath70 .
if @xmath32 , then a slight pushoff @xmath71 of @xmath72 is an equivariant compressing disk with @xmath73 and we may redefine @xmath72 to be @xmath71 instead .
hence , we may assume @xmath74 .
as @xmath75 is a longitude of the solid torus @xmath68 , there exists an embedded annulus @xmath76 such that @xmath77 and @xmath78 .
the @xmath12-sphere @xmath79 contains @xmath39 , is transverse to @xmath35 , and meets @xmath0 in exactly two points .
therefore , @xmath60 divides @xmath80 into two equivariant unknotted ball - arc pairs .
each of these balls contains an equivariant @xmath12-disk with boundary @xmath39 and containing that balls arc of @xmath0 .
the union of these two disks is a new @xmath12-sphere @xmath63 such that @xmath67 is a sphere containing @xmath6 as desired . to prove the reverse implication of the main theorem ,
suppose @xmath6 is not prime .
if @xmath6 is unknotted , then @xmath39 is the unknot which is not prime . if @xmath81 and @xmath58 is nontrivial , then , by lemma [ lem : sum ] , @xmath82 which is not prime
otherwise , @xmath83 , where @xmath43 and @xmath44 are both knotted .
then , by lemma [ lem : sum ] , @xmath84 is a sum of knots which are nontrivial by lemma [ lem : unknotted ] .
this proves the reverse implication of the main theorem .
+ to prove the forward implication of the main theorem , suppose @xmath39 is not prime .
if @xmath39 is unknotted , then so is @xmath6 by lemma [ lem : unknotted ] . otherwise , there is a sphere @xmath60 that splits @xmath85 into two knotted ball - arc pairs .
[ lem : connect_sum ] if @xmath86 , then @xmath6 is a nontrivial order-2 connected sum . if @xmath87 and @xmath60 meets @xmath0 at exactly two points distinct from @xmath21 and @xmath22 , then @xmath6 is a nontrivial order-3 connected sum .
if @xmath86 , then @xmath60 bounds a ball @xmath88 disjoint from @xmath89 .
it follows that @xmath88 is disjoint from @xmath90 . as @xmath91 is connected and disjoint from @xmath60 , @xmath0 must also be disjoint from @xmath88 .
thus @xmath92 maps homeomorphically by @xmath65 to a nontrivial ball - arc pair in @xmath19 .
thus , @xmath6 is a nontrivial order-2 connected sum .
suppose @xmath87 meets @xmath0 at exactly two points distinct from @xmath21 and @xmath22 .
as @xmath93 and @xmath41 interchanges the two lifts @xmath94 and @xmath95 of @xmath10 , @xmath60 meets each of @xmath94 and @xmath95 once .
therefore , the vertices @xmath21 and @xmath22 must lie in different components of @xmath96 , and so @xmath60 meets each arc of @xmath0 once . in particular , the @xmath30-action does not interchange the balls in @xmath4 bounded by @xmath60 . by lemma [ lem : sphere ] , @xmath61 is a sphere and it splits @xmath6 as an order-3 connected sum .
each of these summands must be nontrivial since @xmath60 splits @xmath85 into two knotted ball - arc pairs .
let @xmath60 be a sphere such that : 1 . [ 2.1 ] @xmath60 separates @xmath85 into two knotted ball - arc pairs .
[ transverse ] @xmath60 and @xmath89 are in general position with each other .
[ 2.3 ] @xmath97 .
condition [ transverse ] is achieved by the proof of lemma 1 from @xcite . by lemma [ lem : connect_sum ]
, it suffices to show that we can either improve @xmath60 ( while maintaining [ 2.1][2.3 ] ) and make it disjoint from @xmath89 , or we can produce a sphere @xmath63 which bounds two knotted ball - arc pairs in @xmath85 such that @xmath98 and @xmath63 meets @xmath0 at exactly two points distinct from @xmath21 and @xmath22 . [
lem : inessential ] any curves of @xmath99 which are inessential in @xmath100 can be removed without introducing new intersections .
note that a curve in @xmath99 is essential in @xmath100 if and only if it is essential in @xmath101 .
consider a component @xmath102 of @xmath99 that is inessential in @xmath100 and is innermost in @xmath101 .
then @xmath102 bounds closed disks @xmath103 and @xmath104 and @xmath105 is an embedded 2-sphere . as @xmath106 and @xmath107 are both disjoint from @xmath39 , @xmath105 bounds a ball @xmath88 disjoint from @xmath39 .
case 1 : @xmath108 .
then , there is a neighborhood @xmath68 of @xmath106 such that @xmath109 and @xmath110 .
improve @xmath60 by pushing @xmath107 past @xmath106 into @xmath68 using @xmath88 .
since the only part of @xmath60 that changed now lies in @xmath68 and @xmath68 is now disjoint from @xmath99 , there are no new intersections .
case 2 : @xmath111 .
since @xmath106 is innermost , this means that @xmath112 and @xmath113 . using @xmath88 ,
push @xmath107 past @xmath106 to a parallel copy of @xmath106 .
this removes @xmath102 without adding new intersections . by lemma [ lem : inessential ]
, we may assume all components of @xmath99 are essential in @xmath100 .
let @xmath114 be the components of @xmath99 and let @xmath115 be the annulus with @xmath116 for @xmath117 .
we assume @xmath118 and @xmath119 are adjacent in @xmath100 for @xmath120 .
let @xmath121 be the permutation such that @xmath122 . as @xmath41 is an involution , either @xmath123 is the identity or @xmath123 has order two .
[ lem : fixed_scc ] if @xmath124 and @xmath118 bounds a disk @xmath125 such that @xmath126 , then either @xmath118 can be removed without introducing intersections or @xmath6 is a nontrivial order-3 connected sum .
suppose @xmath72 and @xmath127 are as indicated .
then , @xmath128 is a sphere invariant under @xmath41 .
since @xmath41 is orientation preserving , @xmath128 must have at least one fixed point under @xmath41 ; in fact , it must have two because @xmath0 meets @xmath128 transversely . as @xmath126
, all such fixed points must lie in @xmath118 . as @xmath60 is transverse to @xmath0
, @xmath118 contains a finite number of fixed points under @xmath41 .
let @xmath129 be an arc intersecting @xmath0 exactly at its endpoints , so @xmath130 .
then @xmath131 is a simple closed curve , so @xmath132 and thus @xmath118 can not contain more than two fixed points .
thus , if @xmath128 bounds two knotted ball - arc pairs , then we are done by lemma [ lem : connect_sum ] . otherwise , @xmath128 bounds an unknotted ball - arc pair @xmath92 .
using @xmath88 , we can push @xmath72 ( and any other components of @xmath133 ) past @xmath134 to remove @xmath118 .
thus , we must show that as long as @xmath135 , there is a curve @xmath118 as in lemma [ lem : fixed_scc ] .
[ lem : torus ] suppose @xmath136 is a torus such that @xmath137 and @xmath138 .
let @xmath139 be the half bounded by @xmath140 containing @xmath0 .
if @xmath141 is essential in @xmath140 , null - homotopic in @xmath142 , and equivariant , then it is invariant ( setwise ) .
suppose @xmath140 , @xmath142 , and @xmath102 are as indicated . since @xmath143 , it is clear that @xmath144 .
as @xmath102 is equivariant , null - homotopic in @xmath142 , and disjoint from @xmath0 , @xmath102 bounds an equivariant disk @xmath145 by the equivariant dehn lemma @xcite .
suppose , by way of contradiction , that @xmath146 .
then , as @xmath102 and @xmath72 are equivariant , we have @xmath147 and @xmath74 .
the set @xmath148 has two components , the closures in @xmath142 of these are @xmath149 and @xmath150 .
each of @xmath149 and @xmath150 is bounded by the disks @xmath72 and @xmath134 as well as an annulus in @xmath140 , so both @xmath149 and @xmath150 are 3-balls .
now , @xmath0 is contained in one of these balls . without loss of generality , @xmath151 .
so , both @xmath149 and @xmath150 must be fixed setwise by @xmath41 .
as @xmath149 is fixed by @xmath41 and @xmath41 is orientation preserving , @xmath41 must have a fixed point on @xmath152 .
but , the set of fixed points of @xmath41 is exactly @xmath0 , and @xmath140 , @xmath72 , and @xmath134 are all disjoint from @xmath0 . this is a contradiction , so @xmath102 must be invariant . in the following lemma , we take @xmath153 .
[ lem : there_is_fixed_scc ] if @xmath154 , then there is some @xmath155 such that @xmath124 .
in particular , there is some @xmath127 such that @xmath124 .
suppose , by way of contradiction , that @xmath154 and there is no such @xmath127 .
choose @xmath156 such that : 1 .
[ 1 ] @xmath157 .
2 . @xmath158 is minimal such that [ 1 ] is satisfied .
if @xmath159 , then we can not have @xmath160 .
so , @xmath161 and @xmath162 is a torus .
furthermore , @xmath163 and @xmath140 is disjoint from @xmath0 .
as @xmath0 is connected , one component of @xmath164 contains @xmath0 .
let @xmath142 be its closure in @xmath4 .
since @xmath140 is also disjoint from @xmath39 and @xmath165 is connected , we have @xmath166 .
let @xmath167 be minimal such that @xmath168 . then
, @xmath169 bounds a disk in @xmath60 with interior disjoint from @xmath140 which intersects @xmath39 .
so , @xmath169 and all other curves in @xmath140 of the same homotopy type are null - homotopic in @xmath142 . in particular
, @xmath170 is essential in @xmath140 , null - homotopic in @xmath142 , and equivariant .
hence , @xmath170 is invariant by lemma [ lem : torus ] .
this implies that @xmath171 , a contradiction .
the second part is immediate as either @xmath172 , or @xmath173 and there is @xmath174 such that @xmath124 .
let @xmath175 be minimal such that @xmath124 ( this @xmath127 exists by lemma [ lem : there_is_fixed_scc ] ) .
let @xmath125 be a disk with @xmath176 such that : ( i ) @xmath72 contains @xmath177 in case @xmath178 , ( ii ) @xmath72 does not contain @xmath119 in case @xmath179 , and ( iii ) @xmath72 is either disk in @xmath60 bounded by @xmath180 in case @xmath181 .
we claim that @xmath182 .
suppose , by way of contradiction , that @xmath183 .
then , @xmath184 and @xmath185 .
if @xmath171 , then @xmath156 contradicts minimality of @xmath127 . if @xmath186 , then lemma [ lem : there_is_fixed_scc ] yields @xmath187 such that @xmath188 , which contradicts minimality of @xmath127 .
if @xmath189 , then let @xmath190 .
so , @xmath191 since @xmath192 is the identity .
again , lemma [ lem : there_is_fixed_scc ] yields @xmath193 such that @xmath188 , which contradicts minimality of @xmath127 .
hence , the claim @xmath182 holds . by lemma [ lem : fixed_scc ] , either @xmath118 can be removed or @xmath6 is a nontrivial order-3 connected sum .
thus , if @xmath6 is not a nontrivial order-3 connected sum , then @xmath60 can be made disjoint from @xmath89 and @xmath6 is a nontrivial order-2 connected sum by lemma [ lem : connect_sum ] .
this completes the proof of the main theorem .
to employ the main theorem , we must produce the lifted knot for a given theta - curve .
fortunately , this is not difficult . given a theta - curve @xmath194 with unknotted constituent knot @xmath0 , draw the lifted knot using the following algorithm : 1 .
draw @xmath6 in @xmath195 so that @xmath0 comprises the @xmath42-axis and the point at infinity .
consider the diagram given by projecting @xmath6 onto the @xmath196-plane .
2 . by an ambient isotopy fixing @xmath0 ,
arrange for all self - crossings of @xmath10 to have positive @xmath197-coordinate . by a further isotopy
, we may assume the diagram appears as shown in figure [ fig : ltc2d_1 ] .
3 . now , lifting @xmath10 yields the knot @xmath39 given as the union of the following : ( i ) the diagam @xmath58 , ( ii ) a copy of @xmath58 rotated one half of a revolution about the @xmath42-axis , and ( iii ) arcs between ( i ) and ( ii ) as shown in figure [ fig : ltc2d_2 ] . applying the algorithm to kinoshita s theta - curve , we obtain the lifted knot @xmath39 in figure [ fig : kino_2 ] . with @xmath39
exhibited as a positive @xmath14-braid , it is a simple exercise to isotop @xmath39 to the standard @xmath198-torus knot ( this fact was also observed by wolcott @xcite ) . as torus knots are prime @xcite , @xmath39 is prime and the main theorem implies that kinoshita s theta - curve is prime as well .
previously , the authors @xcite produced uncountably many isotopically distinct unions of three rays in @xmath5 with the brunnian property ( namely , all three rays are knotted , but any two of them are unknotted ) . to achieve this , we used sequences of three - component tangles lying in thickened spheres .
our main tangle @xmath199 is shown in figure [ fig : block_a ] .
we discovered this tangle independently as described in @xcite .
a key property of the tangle @xmath199 proved in @xcite was that @xmath199 is irreducible ( namely , no sphere separates @xmath199 into two nontrivial tangles ) . by taking the thickened sphere containing @xmath199 and crushing each of the boundary spheres to a point , one obtains kinoshita s theta - curve .
as we have just observed , kinoshita s theta - curve is prime .
this immediately implies that the tangle @xmath199 is irreducible and provides an alternate proof of ( * ? ? ?
* theorem 6.1 ) . | we prove a folklore theorem of w. thurston which provides necessary and sufficient conditions for primality of a certain class of theta - curves .
namely , a theta - curve in the 3-sphere with an unknotted constituent knot @xmath0 is prime if and only if lifting the third arc of the theta - curve to the double branched cover over @xmath0 produces a prime knot .
we apply this result to kinoshita s theta - curve . |
it is agreed that type ia supernovae ( sn ia ) are thermonuclear explosions of white dwarfs ( wds ) accompanied by a complete destruction of the wd , or at least one of the two interacting wds ( e.g. @xcite ) . on the other hand ,
there is no consensus yet on the evolutionary route of the exploding wd .
the different scenarios discussed in the literature in recent years can be classified according to different criteria , one of which is summarized in table [ tab : table1 ] .
we list in table [ tab : table1 ] the relevant observational properties of each scenario that might be compatible with the presence of ears , or on the contrary be in conflict with the presence or ears .
we also list what we consider as main predictions , strong characteristics , and main difficulties of each scenario , as these can be relevant to the properties related to ears .
for example , the presence of massive , approximately @xmath10 of csm in the kepler snr , poses a problem to several scenarios , such as the massive csm inferred in the sn ia ptf11kx .
all scenarios suffer from difficulties , but we find that three scenarios suffer from severe difficulties , making them highly unlikely to be the main evolutionary route of sn ia . as all scenarios have difficulties , we list them and discuss them in alphabetical order .
scenario@xmath11}$ ] & core degenerate & double degenerate & double detonation & single degenerate & wd - wd collision + presence of 2 opposite ears in some snr ia . &
explained by the snip mechanism ( studied in this paper ) . &
low mass ears if jets during merger @xcite . &
no ears are expected for he wd companion . & ears by jets from accreting wd @xcite . & no ears are expected + @xmath12 csm in keplers snr & the massive csm shell might be a pn . & no csm shell & any csm is of a much lower mass & can be explained by heavy mass loss from an agb donor .
@xmath13}$ ] & no csm shell + main scenario predictions & 1 .
single wd explodes 2 .
massive csm in some cases ( snip ) & 1 .
sufficient wd - wd close binaries 2 .
dtd @xmath14 & 1 .
asymmetrical explosion 2 .
@xmath15 & 1 .
companion survives 2 .
@xmath16 & asymmetrical explosion + general strong characteristics & & & & & + general difficulties & & & & & + severe difficulties & & & & & + fraction of sn ia & @xmath17 ( this work ) & @xmath18 ( this work ) & @xmath19 @xcite@xmath20}$ ] & @xmath21 & @xmath22 @xcite@xmath23}$ ] + [ tab : table1 ] + notes : + + + + + ( a)_the core - degenerate ( cd ) scenario _ ( e.g. , @xcite ) . within this scenario
the wd merges with the hot core of a massive asymptotic giant branch ( agb ) star . in this case
the explosion might occur shortly or a long time after merger .
by shortly we refer to a time scale within which the sn ejecta will interact with gas originated from the earlier phases of the progenitors , e.g. , the ejected common envelope ( ce ) . by a long time
we refer to a time when any mass lost from the progenitor has dispersed in the ism .
the limit is not sharp , but it is typically @xmath24 .
( b)_the double degenerate ( dd ) scenario _
( e.g. , @xcite ) .
this scenario is based on the merger of two wds .
however , this scenario does not specify the subsequent evolution , namely , how long after merger does the explosion of the remnant take place ( e.g. , @xcite ) .
recent papers , for example , discuss violent mergers ( e.g. , @xcite ) as possible channels of the dd scenario .
this seems very problematic as it leads to a highly non - spherical supernova remnant ( snr ; @xcite ) , while @xcite find that snrs ia generally possess a spherical structure .
others consider very long delays from merger to explosion , e.g. , because rapid rotation keeps the structure overstable @xcite .
the upper limit on the size of the progenitor of sn 2011fe ( e.g. , @xcite ) implies that the exploding object was very compact , ruling out the presence of close gas .
since the merger process ejects gas , the dd scenario is limited to have explosion at a time of @xmath25 after merger @xcite .
as well , the production of manganese @xcite and nickel that requires exploding wds close to the chandrasekhar mass can be explained in the dd scenario if there is enough time for the merging wds to coalesce to a massive remnant .
the dd and the cd scenarios can overlap in the sense that if the merger occurs after the termination of the ce phase , but while the core is not yet on the cooling track of a wd , both scenarios comprise the system .
basically , the scenarios can overlap if the merger occurs during the planetary nebula phase of the system . however , there are two significant differences between the two scenarios . in the dd scenario
the merger is of two cool wds , while in the cd scenario it is a wd with a hot core . in the dd scenario
the delay to explosion is mainly due to gravitational waves emitted by the two wds ( e.g. , review by @xcite ) , while in the cd scenario it is due to angular momentum loss by the merger remnant @xcite .
one significant difference between the two scenarios is that due to the long time scale of gravitational waves , no massive circumstellar matter ( csm ) is expected around the exploding star in the dd scenario . on the other hand , in the cd scenario
a large fraction of sn ia will possess a csm that will influence the spectrum of the sn and/or the morphology of the snr .
since in the cd scenario this csm was ionized at some time by the central wd remnant of the merger , it was a planetary nebula ( pn ) .
such sne ia inside pne , or remnants of pne , are termed snips @xcite .
( c)_the double - detonation ( ddet ) mechanism _
( e.g. , @xcite ) . here
a sub - chandrasekhar mass wd accumulates a layer of helium - rich material coming from a donor .
the helium layer is compressed as more material is accreted and detonates , leading to a second detonation near the center of the co wd ( @xcite and references therein ) .
there are two types of he wd companions : a wd of mass @xmath26 residing at @xmath27 from the exploding co wd @xcite , or a lighter , @xmath28 , wd at an orbital separation of @xmath29 @xcite .
@xcite found that the ddet scenario can account for only a small fraction of all sn ia , but @xcite found a much larger fraction that can be attributed to the sub - chandrasekhar ddet scenario .
@xcite found that the explosion leads to a non - spherical snr , and in a case of a close helium wd , the latter will be ignited and will eject helium .
as elaborated in @xcite , the expected snr has a dipole asymmetry that no well resolved snr ia has .
for example , although sn1006 has a dipole asymmetry , @xcite find this dipole to be different from the one expected due to the effect of the he wd companion on the ejecta in the ddet scenario .
another problem with this scenario is that because the exploding co wd almost does not grow before it explodes @xcite , it predicts that most exploding wds will have a mass of @xmath30 .
this is in odd with recent findings that most sn ia masses are peaked around @xmath31 @xcite .
@xcite also claim that at least @xmath32 of all sn ia come from near chandrasekhar mass ( @xmath33 ) wds .
we conclude that the ddet scenario can lead to explosions similar to sn ia , but not to the common sn ia .
( d)_the single degenerate ( sd ) scenario _
( e.g. , @xcite ) . in this scenario a white dwarf ( wd ) accretes mass from a non - degenerate stellar companion and explodes when its mass reaches the chandrasekhar mass limit .
the search for companions to sne ia came with null results ( e.g. @xcite ) and made this scenario very unlikely .
another grave difficulty is that population synthesis studies find that the sd scenario can account for only a small fraction of sn ia ( e.g. , @xcite ) .
we note that @xcite who included delay detonation due to wd spin - up , on the other hand , argue that the sd scenario can produce the required number of sne ia , and the observed delay time distribution .
a very interesting set of observations is the detection of csm from hydrogen lines ( e.g. , @xcite ) and sodium absorption lines @xcite in some sn ia .
these have been used to argue for the sd scenario ( e.g. , @xcite and @xcite from hydrogen , and @xcite and @xcite from sodium ) .
however , it seems that the csm mass inferred both from hydrogen , as in ptf 11kx @xcite , and from sodium absorption lines @xcite , is too large to be accounted for by the sd scenario . instead
, @xcite and @xcite strongly argue that the cd scenario best accounts for these observations .
\(e ) _ the wd - wd collision scenario _
( e.g. , @xcite ) . in this scenario two
wds collide and immediately ignite .
the collision is most likely induced by a tertiary star @xcite by the lidov - kozai mechanism @xcite . despite some attractive features of this scenario , it can account for at most few per cent of all sne ia @xcite .
another problem is that the explosion in most cases is highly non - spherical , contradicting the observed morphology of most young close - by snrs .
a third problem might be the delay time distribution ( dtd ) that behaves as @xmath14 ( e.g. , @xcite ) .
although the wd - wd collision scenario can lead to the observed @xmath14 dtd , it does so only at late times after star formation , @xmath34gyr according to @xcite .
the fraction of systems that might be explained by the wd - wd collision is @xmath35 of all sn ia @xcite .
based on the discussion above we consider the cd and dd scenarios as the most likely scenarios to account for most or all sn ia . in this paper
we try to set a lower limit on the fraction of sn ia that come from the cd scenario by considering the influence of the csm on the sn spectra ( section [ sec : sncsm ] ) and on the morphology of snrs ( section [ sec : snr ] ) .
our short summary is in section [ sec : summary ] .
@xcite find that @xmath36 of sn ia in star - forming galaxies exhibit time - variable sodium features associated with circumstellar material .
@xcite obtain a similar result , showing that @xmath0 of sne ia possess blueshifted narrow na i d absorption features compared to non - blueshifted na i d features . some papers ( e.g. , @xcite ) attributed the na i d absorption lines to a wind from a giant star in the sd scenario .
however , it seems that the csm that supplies the sodium is too massive to be accounted for in the sd scenario @xcite .
instead , @xcite suggested that the sodium comes from dust in a @xmath37 shell .
this na - from - dust absorption ( nada ) model better fits the cd scenario where the shell is the mass that was ejected in the common envelope ( ce ) process ( assuming it is not ism ) . for absorption lines to appear , the csm should reside along the line of sight . since in many cases the csm might not cover the entire sphere , the @xmath36 is a lower limit . on the other hand , in some cases the absorbing gas might be of ism origin .
overall , the study of @xcite , within the frame of the nada model , suggests that @xmath38 of sn ia have massive csm that qualifies them as snip candidates .
we note that @xcite , in the frame of the ddet scenario , and @xcite , within the frame of the dd scenario , have also attempted to explain the presence of csm around some sne ia .
the csm mass in these models is too low for the nada model @xcite , although can be sufficient for other explanations for the na absorption lines , e.g. @xcite and @xcite .
interestingly , all the sn ia that show time - variable sodium features reside in star - forming galaxies @xcite .
this suggests that either the absorbing gas is of ism origin , as dense clouds are abundant in star forming galaxies , or that the progenitors of these systems are relatively massive stars , @xmath39 that evolve rapidly .
we assume that the time - variable sodium features are due to csm as in that case the csm must reside relatively close to the exploding wd . in that case
the positive correlation of variable absorption with star forming galaxies is compatible with the cd scenario @xcite .
we note though that presently the studies of @xcite and @xcite can not be used to distinguish between ism and csm features for an individual sn @xcite . in the last decade
more and more sne ia have been claimed to interact with a dense csm ( e.g. , @xcite and @xcite ) .
we mention here two such sn ia , sn 2002ic and ptf 11kx .
@xcite argued that the hydrogen lines in sn 2002ic can be explained by a few solar masses that have been lost by the progenitor shortly before the explosion , and attribute it to an agb companion in the frame of the sd scenario .
@xcite attributed the massive csm to the ejection of a ce , where a wd companion merged with the hot core of a massive agb star ; a process that was later termed the cd scenario .
the massive csm was very close , some mass as close as @xmath40 , to the sn origin when explosion occurred .
sn ia ptf 11kx @xcite had narrow hydrogen lines and indications of interaction with a massive csm which started 59 days after the explosion .
@xcite argued that the csm was formed by a giant companion to the exploding wd , as is the case in some channels of the sd scenario .
they dismiss the merger scenario as suggested by @xcite on several grounds , e.g. , it can not account for several csm shells .
@xcite argued to the contrary , and explained ptf 11kx in the frame of the cd scenario .
in particular , @xcite estimated the csm mass within @xmath41 to be @xmath42 , much above what the sd scenario can supply .
@xcite crudely estimate that the fraction of sne ia that exhibit prominent circumstellar interaction near maximum light , e.g. , sn 2002ic and ptf 11kx , is @xmath43 .
as @xmath44@xmath45 sn ia occur per @xmath46 stars formed @xcite , the estimate of @xcite stands at @xmath47 sn ia with massive csm per @xmath48 stars formed .
the monte carlo simulations of @xcite show that the frequency of these systems is consistent with the cd scenario .
@xcite take the sn ia rate at short delay times of @xmath49myr ( what they term prompt ) to be constant .
based on that , we take the probability for explosion after the termination of the ce to be constant with time up to few @xmath50 . in our modelling this rate is higher than that found by @xcite .
namely , we find in this paper that the sn ia rate within about one million years of the ce phase is very high .
for a constant probability per unit time to explode within hundreds of thousands of years after merger , the probability to explode is linear with the radius of the csm up to several , assuming the csm expands with a constant velocity of tens of @xmath51 .
namely , the probability for the system to explode by the time the csm reached a radius of @xmath52 ( as is the case for the na absorbing gas , ) is @xmath53 times that of a system where the csm only reached a radius of @xmath54 , as is the case of the sn ia - csm systems .
therefore , the @xmath55 estimate by @xcite for cases where the csm is close to the exploding site , translates to few @xmath56 of snips ( where the csm is at @xmath52 ) .
this is consistent with the estimate from the na absorption lines discussed in section [ subsec : sodium ] .
out of a dozen of resolved sn ia remnants , at least a few may be attributed to sn ia that exploded inside a csm shell . for kepler
s snr , the interaction of the sn ia with the csm is established @xcite .
our previous work on this snr @xcite suggests that kepler s sn exploded inside a csm shell that originated as a pn .
we term this scenario sn ia inside pn , or snip . in a recent work
, we examined snr g1.9 + 0.3 @xcite and suggested that its observed x - ray morphology also hints at a snip origin .
examining the x - ray and radio snr images in the chandra snr catalogue @xcite , we try to identify other snrs that might have evidence for snip origin . as an indication for a possible snip origin
we take here the morphological feature of two opposite ears in the snr ( nomenclature by @xcite ) , a feature both kepler and g1.9 + 0.3 have . on the one hand
some ` ears ' might originate in ism interaction rather than csm . on the other hand
some snips might contain no ` ears ' at all .
as well , we do nt consider the possibility of forming the ears by jets blown by the merging wd and the core ( see @xcite ) , although it exists .
for the preliminary estimates of the present study we consider our approach to be adequate .
we focus only on close - by snrs that have a significant apparent angular size , and are well - resolved to an extent that allows to identify or disprove the existence of two opposite ears in the snr morphology .
all of the sne in our sample are dynamically young , with estimated ages of less than @xmath57 @xcite .
in other words , we only examine snrs ia that have clearly resolved sharp images that can show relatively detailed filaments structure in the snr shell .
very old snrs that have been significantly influenced by ism interaction are not considered .
though these criteria limit us to snrs in our immediate neighborhood ( the galaxy and the magellanic clouds ; mcs ) , we assume that this is a fair representative of other star - forming galaxies .
we have identified a total of 13 snrs that comply with our criteria and list them in table [ table : snrs ] . in kepler s snr and snr g1.9
+ 0.3 , and possibly snrs dem l71 , g299.2 - 2.9 , 0534 - 69.9 and 0519 - 69.0 , we identify opposite ears structures .
we define the opposite ears structure as two protrusions from the overall spherical shape of an snr , directly opposite one to another .
the protrusions have width of the order of roughly one tenth of the radius of the snr so that they can be easily identified by eye .
no other similar protrusions beside these two should be present for a positive identification .
these snrs are marked as having ears ( `` yes '' ) in table [ table : snrs ] .
some snrs have such protrusions , but they are less proclaimed or the image is not sharp enough to determine if there are indeed opposite ears .
these snrs are marked as possibly having ears ( `` maybe '' ) in table [ table : snrs ] .
we already performed hydrodynamical simulations to explain the first two snrs as snips .
the study of snr g299.2 - 2.9 is a goal of a future work . from the listed snrs ia in the table we derive an initial estimate of @xmath58 of well - resolved snrs ia possessing the opposite - ears morphology .
in estimating the snip fraction other uncertainties should be considered . first , other scenarios can also create opposite ears , lowering the number of snip out of the snrs ia having opposite ears .
such scenarios include dynamical interaction with the ism , in which small randomly positioned protrusions can be created by instabilities in the ejecta @xcite , and jets blown during the merger process @xcite .
another speculative possibility that we raise here is an asymmetric explosion with jets along a symmetry axis that can give rise to opposite ears when the jets interact with the ism .
we also note that detecting ears in snrs can be a function of viewing angle . in @xcite
we estimated that the probability to observe the ears features in kepler s snr is roughly @xmath59 .
therefore , some of the snrs that we identified as not having ears in our sample might in fact be snips , but with the ears projected in front and behind the snr itself .
this introduces an additional uncertainty to our crude estimate , lowering the number of sn ia identified as snip candidates .
another source of uncertainties is the usage of an x - ray catalogue , that might be biased toward sne ia exploding in dense ism and/or within a csm .
if a sn ia occurs in a very low density ism and there is no csm , then the x - ray emission that results from the ejecta interaction with the surrounding gas will be very weak .
we might miss non - interaction sn ia .
this is likely to be the case with sn ia in the halo of the galaxy .
overall the uncertainties are large , and our estimate of the fraction of snip candidates in well resolved snrs ia being approximately @xmath60 is only a crude upper limit .
the resolved snrs are all located in the star - forming milky way galaxy and lmc .
@xcite showed that star - forming galaxies have a mass - normalized sn ia rate of @xmath61 snum @xmath62 , while passive galaxies ( with no recent star formation ) have a lower rate of @xmath63 snum .
our proposed snip scenario is expected to occur in systems that explode through the cd channel ( see section [ sec : intro ] ) .
this will predominantly occur in star - forming galaxies .
@xcite show that @xmath64 of the stellar mass in galaxies having a stellar mass of @xmath65 in our local universe ( @xmath66 ) resides within elliptical ( passive ) galaxies .
assuming that no snips occur in passive galaxies , and taking into account the mass - normalized sn ia rates above @xcite , and the mass fraction in elliptical galaxies of @xmath67 , we find that @xmath68 of sn ia occur in star forming galaxies . with our estimate from table [ table : snrs ] of @xmath60 being snips in star forming galaxies , we arrive to estimate that approximately @xmath69 of all sne ia may be snips by the opposite - ears criterion .
.known well - resolved snrs with ages less than @xmath70 .
we identify the morphological feature of two opposite ears in between 2 and 6 of the 13 snrs .
all images are taken from the chandra snr catalogue ( references therein ) . [
cols="<,^,^,^,^,^",options="header " , ] in the cd scenario the two wds merge at the end of , or very shortly after , the ce phase . in some cases
the two wds might not merge because the agb envelope is not massive enough .
the envelope is then ejected before merger . in such systems , according to the scenario discussed here , the two wds mass is close or over the chandrasekhar mass , and the pn contains `` ears '' , i.e. , small lobes .
we are aware of two pne that have been claimed to host such a wd binary system , and these indeed have `` ears '' .
nova v458 has a binary system at the center , with a very short orbital period of @xmath71 , consisting of a relatively massive wd ( @xmath72 ) and a secondary post - agb star ( @xmath73 ; @xcite ) .
it is an elongated pn that has two opposite lobes @xcite .
the gravitational wave inspiral time for this binary is @xmath74 , making it too long for a snip explosion to take place , but not by far .
pn g135.9 + 55.9 ( ts01 ) hosts a binary system with an orbital period of @xmath75 and a total mass very close to the chandrasekhar limit @xcite .
it is a pn that has two opposite ears @xcite .
these two systems , we argue , support our snip scenario .
they belong to systems just outside the boundary of the parameter space of the cd scenario where merger has been barely prevented , most likely due to a too low agb envelope mass . many others wd - core systems with a total mass close to the chandrasekhar mass merged at the end of the ce phase and formed snips .
in this work we estimated the fraction of sne ia exploding inside pne ( snips ) , using three independent calculations .
@xmath1 an analysis of sodium absorption lines in snrs ia , based on @xcite and @xcite and described in section [ subsec : sodium ] .
the results suggest that @xmath76 of sn ia possess massive csm and thus can be potentially qualified as snips . @xmath3 an estimation of the probability for sn ia explosion that will have a massive hydrogen csm , based on @xcite and described in section [ subsec : hcsm ] .
this yields an estimate of @xmath77 of all sne ia having snip origin . @xmath7
an examination of the x - ray morphologies of well - resolved close - by snrs ia , described in section [ sec : snr ] .
approximately @xmath78 of the well - resolved snrs ia possess opposite ear - like features , which we take as evidence of snip origin .
considering the star - forming nature of the galaxy and the lmc , and taking into account the mass - normalized sn ia rates for star - forming and passive galaxies , this estimate translates to a fraction of @xmath79 of all sne ia having snip origin , as a crude upper limit .
it is not clear yet whether the core - degenerate ( cd ) scenario can account for all sn ia and whether it can reproduce the dtd ( for two opposite views see @xcite and @xcite ) .
however , it seems quite secure that the cd scenario can explain the large fraction of sn ia occurring shortly after star formation , i.e. , within @xmath80 , i.e. , shortly after the two massive stars of a binary system end their evolution on the agb @xcite .
some studies argued for a very small contribution of the cd scenario .
we attribute their results to two inaccurately treated processes .
first is the mass transfer from the more massive star during its rgb and/or agb phases to the still main - sequence companion .
@xcite pointed out , based on the code used by @xcite , that the transferred mass should be higher than usually assumed in population synthesis studies .
the second inaccurately treated process is the ce ejection , which is too `` optimistic '' ( in the sense that the common envelope is assumed to be easily ejected ) in some population synthesis studies , e.g. , @xcite .
we note that @xcite argue that the cd scenario can not account for the csm they find in their study .
we strongly disagree with that conclusion . despite the difficulties and open questions listed in table [ tab : table1 ] , we suggest that the cd scenario is the most likely scenario out of the 5 listed there to account for all the properties of the csm ( e.g. , @xcite ) . our results , as well as other papers cited here , e.g. , that the sd scenario contribution to sn ia is at best very small , suggest that the focus in the open questions of sn ia progenitors should be shifted from the sd scenario to other scenarios .
our view , as summarized in the last row of table [ tab : table1 ] , is that the two main sn ia channels are the dd and cd scenarios . though we could only propose crude estimates that still demand a deeper and more thorough analysis , we are confident in the conclusion that a significant fraction ( at least @xmath81 , and possibly more ) of sne ia are snips , most likely formed through the cd scenario .
we thank an anonymous referee for many detailed comments that substantially improved both the scientific content of the paper and the presentation of our results .
this research was supported by the asher fund for space research at the technion , and the e. and j. bishop research fund at the technion .
raskin , c. , scannapieco , e. , fryer , c. , rockefeller , g. , & timmes , f. x. 2012 , , 746 , 62 raskin , c. , & kasen , d. 2013 , , 772 , 1 raskin , c. , timmes , f. x. , scannapieco , e. , diehl , s. , & fryer , c. 2009 , , 399 , l156 toonen , s. , nelemans , g. , bours , m. , portegies zwart , s. , claeys , j. , mennekens , n. , & ruiter , a. 2013 , asp conference proceedings , vol .
san francisco : astronomical society of the pacific .
, edited by j. krzesi;ski , g. stachowski , p. moskalik , and k. bajan , p.329 | using three independent directions we estimate that the fraction of type ia supernovae ( sne ia ) exploding inside planetary nebulae ( pne ) , termed snips , is at least @xmath0 .
our three directions are as follows .
@xmath1 taking the variable sodium absorption lines in some sn ia to originate in a massive circumstellar matter ( csm ) , as has been claimed recently , we use the results of sternberg et al .
( 2014 ) to imply that @xmath2 of sn ia occur inside a pn ( or a pn descendant ) , hence classify them as snips .
@xmath3 we next use results that show that whenever there are hydrogen lines in sn ia the hydrogen mass in the csm is large @xmath4 , hence the explosion is a snip .
we make the simplest assumption that the probability for explosion is constant in time for up to about @xmath5 after the merger of the core with the white dwarf ( wd ) in the frame of the core - degenerate scenario .
this results with at least few @xmath6 of sne ia that may have a snip origin .
@xmath7 we examine the x - ray morphologies of 13 well - resolved close - by sn remnants ( snrs ) ia and derive a crude upper limit , according to which @xmath8 of all snrs ia possess opposite ear - like features , which we take as evidence of snip origin .
our results , together with several other recent results , lead us to conclude that the two scenarios most contributing to sne ia are the core degenerate and the double degenerate scenarios . together
these two account for @xmath9 of all sne ia . |
when considering a physical system which is far from thermodynamic equilibrium and is open , i.e. which exchanges either mass , momentum or energy with the external world , it is customary to write the total time derivative @xmath2 of boltzmann - gibbs entropy @xmath3 of the system in the form @xcite : @xmath4 here and in the following , the word entropy refers to boltzmann - gibbs entropy only .
we have adopted the notation of @xcite , @xcite and @xcite ; @xmath5 and @xmath6 are boltzmann s constant and the distribution function respectively ; @xmath7 is the probability of finding the system at the time @xmath8 inside the volume element @xmath9 of a @xmath10-dimensional state space which is centered at @xmath11 and satisfies the normalization condition @xmath12 at all times .
for instance , @xmath13 may be coordinates in @xmath14 , concentrations of chemical species , etc .
moreover , @xmath0 and @xmath15 are the contributions of physical processes occurring inside the bulk of the system and at the system boundary respectively ( some authors denote @xmath0 and @xmath15 with @xmath16 and @xmath17 respectively ; we drop also the dependence on @xmath18 and @xmath8 unless otherwise specified below , for simplicity ) .
the quantity @xmath0 may include e.g. joule heating , viscous heating etc .
the quantity @xmath15 may be related to the exchange of mass , energy etc . with the external world ; it is the amount of entropy which escapes away from the system across the system boundaries per unit time and may be @xmath19 , depending on the particular problem .
it is customary to write @xmath0 and @xmath15 as a volume integral of a suitably defined entropy production density @xmath20 ( referred to as entropy source strength in @xcite ) on the system volume and as a surface integral of some suitably defined entropy flow across the system boundary respectively . here
@xmath21 at all times everywhere throughout the system because of the second principle of thermodynamics ( as for the counterexample provided by @xcite , see sec .
[ sec6 ] ) .
it follows that @xmath22 at all times -see e.g. sec .
49 of @xcite for a detailed discussion of this issue in the particular case of a viscous , thermally conducting fluid .
in contrast with closed ( i.e. , not open ) systems where @xmath23 at thermodynamic equilibrium , the relative role of @xmath0 and @xmath15 in open systems far from thermodynamic equilibrium is far from being assessed even in the neighbourhood of steady states - but for a very restricted class of problems @xcite @xcite @xcite @xcite . in particular , a crucial issue is the spontaneous evolution ( relaxation ) towards a final ( relaxed ) state . as for closed systems ( where @xmath24 at all times ) , relaxation is related to a growth of @xmath25 towards its maximum value .
thermodynamic equilibrium is the relaxed state ; its extremum property @xmath23 , which implies also @xmath26 , is related to stability of this state against perturbations . of course , if we decide to apply a transformation of coordinates : @xmath27 the validity of the extremum property @xmath23 remains unaffected , just like the equations of motion involved in liouville theorem underlying it .
this conclusion holds even if the actual value of @xmath25 may be not invariant under - see appendix [ proofs2 ] .
invariance concerns the extremum property , rather than the actual value of the extremised quantity . as for open systems ( where @xmath15 may @xmath28 )
, boundary conditions may keep the relaxed state ( if any exists ) far from thermodynamic equilibrium , so that @xmath29 .
the question if a description of stable relaxed states of an open system involves extremum properties while requiring no detailed knowledge of the microscopic dynamics of the system - in analogy with the familiar equilibrium thermodynamics - is still open .
usually , once the extremum property is proven relaxation is described as the evolution of a system which starts from some initial condition inside a suitable attraction basin and ends at the configuration corresponding to the extremum .
some researchers have invoked the so - called local thermodynamic equilibrium approximation ( lte ) @xcite . in a nutshell
, lte means that - although the total system is not at thermodynamic equilibrium - the relationships among thermodynamic quantities within a small mass element are the same as in real equilibrium .
lte leads to the so - called general evolution criterion , an exact result concerning the total time derivatives of thermodynamic quantities @xcite , which rules relaxation in many physical systems - see ref .
@xcite for a review .
moreover , it has been proven @xcite that relaxed states in open systems satisfy extremum properties if some very restrictive assumptions are satisfied , like onsager symmetry relationships @xmath30 , where @xmath31 is onsager matrix : these properties are glansdorff and prigogine s minimum entropy production principle of suitably constrained minimization of @xmath2 , henceforth referred to as minep , and its equivalent formulation , onsager and machlup s least dissipation principle. the relevant formulation of minep depends on the actual boundary conditions .
the proof of minep relies on the positive sign of scalar quantities like the quadratic form @xmath32 , with @xmath33 @xmath34-th and @xmath35-th thermodynamic force respectively .
scalars are invariant under , hence minep too is invariant under .
if such restrictions are dropped , minep does not apply @xcite @xcite . among the many attempts to go beyond lte and onsager ,
many authors have postulated a maximum entropy production principle ( mepp ) in different forms @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite @xcite
@xcite @xcite @xcite @xcite - for a review , see @xcite .
according to mepp , a stable relaxed state corresponds to a suitably constrained maximum of @xmath2 . to date , and in spite of its present popularity , attempts to derive mepp rigorously meet no general consensus @xcite @xcite @xcite @xcite , as they _are so far unconvincing since they often require introduction of additional hypotheses , which by themselves are less evident than the proven statement_ @xcite - see also the critical review of @xcite .
a way out is to drop altogether , and to postulate - in the framework of the so - called extended irreversbiel thermodynamics - that entropy depends locally not only on temperature , pressure etc . but also on the heat flux and the viscous stress tensor @xcite .
however , the results reviewed in @xcite cast doubt on the very existence of a mepp - like , general - purpose extremum property of @xmath2 in far - from - equilibrium , stable , relaxed states ; for selected classes of problems , however , validity of variational principles concerning particular contributions to the entropy balance remains possible .
for instance , many alleged applications of mepp seem indeed to be statements concerning @xmath15 -see appendix [ meppapp ] for further discussion .
a different , quite successful approach to non - equilibrium , namely stochastic thermodynamics ( st ) , deals precisely with selected classes of physical systems , namely those systems where a non - equilibrium process occurs which is coupled to one ( or several ) heat bath(s ) of constant temperature .
st provides a framework for extending the notions of familiar thermodynamics like work , heat and entropy production to the level of individual trajectories across the space state of the system - see the reviews of @xcite and @xcite .
typically , st describes systems where a few observable degrees of freedom - like the positions of colloidal particles or the gross conformations of biomolecules - are in non - equilibrium due to the action of possibly time - dependent external forces , fields , flow or unbalanced chemical reactants .
in contrast with the approaches described above , which deal with large ( macroscopic ) systems where fluctuations around mean values may be negligible , st is uniquely suitable for describing smaller ( mesoscopic ) systems where the role of ( usually quantum - mechanical ) fluctuations is not negligible . in the following , and with a slightly misleading wording , we take the words macroscopic , classical , large and non - degenerate as interchangeable ; in the spirit of our thermodynamic treatment , a rigorous definition of the words macroscopic and mesoscopic is provided in sec . [ sec6ter ] which invokes no information on the detailed mechanics underlying fluctuations . the time - scale separation between the observable , typically slow , degrees of freedom of the system and the unobservable fast ones ( linked either with thermal baths or with the internal dynamics of the system , like e.g. in the case of biopolymers ) , together with the fact that temperatures remain well - defined at all times , allows for a consistent thermodynamic description in st .
internal energy , entropy and free energy are well - defined and , in principle , computable for fixed values of the slow variables .
usually , the time - scale separation implies that the dynamics becomes markovian , i.e. , the future state of the system depends only on the present one with no memory of the past .
if the states are made up by continuous variables ( like position ) , the dynamics follows a langevin equation for an individual system and a fokker - planck equation for the whole ensemble .
this is another difference from the above described approaches , which rely on no detailed description of dynamics . rather than to conditions for stability of relaxed states ,
st leads to a large number of exact results - usually referred to as fluctuation theorems ( fts ) .
basically , fts are mathematical identities derived from the invariance of the microscopic dynamics under time - reversal - see e.g. assumption ( 2 ) in sec .
2.1 of @xcite .
they acquire physical meaning by associating their mathematical ingredients with the thermodynamic quantities identified within st .
the success of invariance - based fts in the description of some mesoscopic systems is confirmed by experiments .
it suggests that even a rigorous non - equilibrium thermodynamics of some selected classes of macroscopic systems should take advantage of some property of invariance - just like in some systems onsager symmetry follows from invariance of microscopic dynamics under time reversal .
remarkably , a derivation of mepp from the second principle of thermodynamics actually exists , but only provided that _the maximum flow is taken as a zero flow _ [ ... ] _ in practice , this can be realised , e.g. by time / space scaling_ @xcite . generally speaking , therefore
, a different choice of _time / space scaling_ seems to lead to violation of mepp , i.e. mepp may not be invariant under .
in contrast , some authors @xcite @xcite @xcite @xcite @xcite suggest that invariance under is precisely a requirement for meaningful extremum properties for stable relaxed states of selected , open macroscopic systems .
( in the following , by invariance we mean invariance under ) .
the issue is far from being purely academic . generally speaking ,
when it comes to systems which remain confined at all times with a certain region @xmath36 of the phase space , does _ not _ leave the measure of @xmath36 unaffected . as a matter of principle , therefore , should an invariant extremum property for stable relaxed states actually exist , it would provide information concerning stability regardless of the size of the system ; in other words , a common thermodynamic description of stability would be available for both macroscopic and mesoscopic systems .
a simple , far - reaching corollary follows straightforwardly . to date
, the relaxed states are usually assumed to be steady states in the non - equilibrium thermodynamics of macroscopic systems - for an exception concerning minep , see e.g. chap .
xv of ref .
in contrast , cyclic behaviour plays a central role in st .
in fact , despite the fundamental difference between isothermal engines operating at one temperature and genuine heat engines like thermoelectric devices involving two baths of different temperature , st provides a common framework based on the representation of entropy production in terms of cycles of the underlying network of states @xcite .
accordingly , should an invariant extremum property for stable relaxed states actually exist , it would provide information on the stability of both steady and oscillating relaxed states .
a bridge could be established between different problems of stability far from thermodynamic equilibrium . in this paper
we focus our attention on a wide - but well - delimited - set of physical systems and assume that relaxed states exist .
our goal is to find an invariant , necessary condition of stability of these relaxed states .
we examine the invariant quantities in the entropy balance of the system in sec .
we discuss the connection between invariance and stability of steady relaxed states and hint at a possible criterion for stability in sec .
[ sec3 ] ; when discussing invariance , we shall make use of dimensionless quantities only , in order to simplify the maths .
since we are looking for necessary condition of stability , we are free to select the perturbations stability is to be checked against ; this allows us to focus our attention on stability against slowly evolving perturbations , and leads therefore to dramatic simplification , as shown in sec .
[ sec4 ] . in sec .
[ sec5 ] we take advantage of the results of the previous sections and show that the criterion hinted at in sec .
[ sec3 ] is actually the looked - for necessary condition for stability of steady relaxed states , starting from the detailed analysis of a relevant , pivotal problem of plasma physics
. no onsager symmetry is assumed .
generalization to oscillating relaxed states is provided in sec .
[ sec6bis ] .
we show that invariance of physics allows easy reformulation of the result of sec .
[ sec6bis ] into a form which is feasible for both macroscopic and mesoscopic systems in sec .
[ sec6ter ] .
the following sections deal with further applications in both macroscopic and mesoscopic systems : rayleigh - bnard convection ( sec .
[ sec5bis ] ) in a rotating shell ( sec .
[ rossby ] ) , a flashing ratchet ( sec .
[ sec6 ] ) and a brownian motor ( sec .
[ brown ] ) .
conclusions are drawn in sec .
detailed computations are to be found in appendices [ proofs2 ] , [ conf1app ] , [ qual ] , [ cgleapp ] and [ rossbyapp ] . appendix [ qual ] contains also a qualitative discussion of our necessary condition of stability starting from st .
the aim of this section is to review some useful results @xcite @xcite @xcite concerning the behaviour of under for a quite wide set of physical systems with @xmath37 degrees of freedom .
these systems are described by the following pde ( partial differential equation ) problem : @xmath38 here @xmath39 is a probability density current with @xmath37 components , the quantity @xmath40 too has @xmath37 components , @xmath41 refers to partial derivatives on the @xmath42 s ( @xmath43 ) , and @xmath44 is a positive scalar with @xmath45 and @xmath46 , which plays the role of a diffusion coefficient in a space with @xmath37 dimensions , so that acts as a simple balance equation for @xmath47 .
we invoke no information concerning the detailed physical nature of @xmath40 and @xmath44 .
the boundary condition corresponds to a system which remains confined at all times inside a region @xmath36 - with measure @xmath48 - of the state space , @xmath49 being the unit normal vector of the boundary surface of @xmath36 . here and in the following , we compute all integrals in @xmath9 on @xmath36 and denote with @xmath50 the average of the generic quantity @xmath51 on @xmath36 .
any physical system undergoing a markovian evolution dictated by a fokker - planck equation in thermal contact with an environment ( say , a thermal bath ) provides us @xcite with an example of a system described by ; generalization is discussed in @xcite .
another example of physical system described by is a single macromolecule @xcite in a fluid , which is composed by @xmath52 atoms and whose configuration is represented at any time by a point @xmath18 in a space with @xmath53 dimensions ( the fundamental equation describing the evolution of the distribution function of the system , namely equation ( 5 ) of @xcite , is a slight generalisation of - the main difference is that @xmath44 is a tensor ) .
finally , if @xmath54 , the system is made of many particles with the same mass and we identify @xmath40 with a velocity ( so that @xmath47 and @xmath39 are @xmath55 the mass density and the mass current density respectively ) , then in the limit of negligible @xmath44 the pde describes a system where mass is conserved with no net source and is exchanged nowhere across the boundary with the external world . for the purpose of our discussion
, we provide some preliminary information about .
quite generally , we assume the @xmath42 s to be curvilinear coordinates in a @xmath56dimensional riemannian variety ( like e.g. @xmath14 ) with state - space volume element @xmath9 .
we denote with @xmath57 and @xmath58 the i - th controvariant and covariant component of the generic @xmath56dimensional quantity @xmath59 respectively ( @xmath60 here and in the following ) .
moreover , we denote the jacobian @xmath61 of with @xmath62 . for simplicity , we assume to be an orientation - preserving diffeomorphism , so that @xmath63 . note that leaves time unaffected .
finally , we denote with @xmath64 the value of the generic quantity @xmath65 after the transformation .
first of all , it is shown in appendix [ proofs2 ] that : @xmath66 @xmath67 as for , basically a particular case of equation ( 19 ) of @xcite , it retrieves both equation ( 12 ) of @xcite and equation ( 7 ) of @xcite provided that @xmath68 in agreement e.g. with einstein s relation ( @xmath69 is a scalar quantity , hence it is invariant ) .
indeed , both @xmath70 and @xmath71 are supposed to vanish in both @xcite and @xcite , just like @xmath72 in our treatment .
this assumption is justified because of the interaction with a thermal bath in @xcite , and is explicitly stated in @xcite , where @xmath73 . finally , in sec .
[ sec4 ] we are going to discuss the behaviour of @xmath0 in the limit of negligible @xmath44 .
as for , it is a particular case of equation ( 20 ) of @xcite .
if @xmath40 is conservative , i.e. if @xmath74 where the scalar quantity @xmath75 is a differentiable function of @xmath18 , then our boundary condition makes @xmath15 to vanish after integration by parts in steady state ( where @xmath76 and @xmath77 ) ; implies therefore @xmath78 .
( an alternative definition of a conservative @xmath40 is @xmath79 , where the @xmath80-th component of @xmath81 is @xmath82 for the generic @xmath59 ) .
accordingly , if @xmath40 is conservative then thermodynamic equilibrium is the only possible steady state , and far - from - equilibrium steady states are possible only for other boundary conditions - think e.g. of a net amount of current flowing across the boundary .
far - from - equilibrium steady states with the boundary condition of are possible only if @xmath40 is not conservative .
not surprisingly , the latter statement is false when @xmath73 and @xmath83 @xcite @xcite , since @xmath84 in this case ; the question is briefly assessed in sec .
[ sec4 ] .
now , it comes to the discussion of invariance .
physical intuition dictates that and take the same form before and after the transformation . as for the quantities involved in , it is shown in appendix [ proofs2 ] that this requirement of invariance implies : @xmath85 @xmath86 @xmath87 @xmath88\ ] ] as for the quantities involved in , it is shown in appendix [ proofs2 ] that the values of both @xmath25 , @xmath2 and @xmath15 are not invariant ; in contrast , @xmath0 is invariant : @xmath89 accordingly , the only invariant quantities in are @xmath0 and @xmath44 .
the roles of @xmath0 and @xmath44 are discussed in sec .
[ sec3 ] and sec .
[ sec4 ] respectively .
we discuss in appendix [ conf1app ] a further aspect of . in spite of its simplicity ,
has a far - reaching consequence . for a sufficiently well - behaved @xmath47 ,
the theorem of the mean allows us to rewrite the invariant normalization condition @xmath12 as @xmath90 where @xmath91 belongs to @xmath36 .
together , invariance of the latter relationship and equation lead to : @xmath92 , a quantity which may differ from 1 .
this means that @xmath93 is usually not invariant .
it follows that if an invariant criterion of stability exists , then it holds regardless of the actual value of @xmath93 ; in particular , it may hold for both macroscopic and mesoscopic systems , as anticipated in [ sec1 ] .
[ sec6ter ] presents further , in depth - discussion of the issue . for future reference , here we discuss also the particular case of rescaling .
by rescaling we mean a diffeomorphism where @xmath94 , @xmath95 if @xmath96 and @xmath97 otherwise , and @xmath98 . equation is invariant under rescaling ; indeed , if @xmath98 then , and ensure that @xmath47 transforms just like @xmath41 , that @xmath39 is invariant , and that @xmath40 transforms like a vector .
the aim of this section is to assess the impact of on the stability of steady relaxed states of a system described by against small perturbations . under mild assumptions
, the system evolves - with a typical time - scale @xmath99 - towards a relaxed state where the current @xmath39 has no sinks and sources , i.e. @xmath77 . the relaxed state coincides with thermodynamic equilibrium when @xmath78 , hence @xmath100 , while non - equilibrium relaxed states are characterized by non - vanishing currents that circulate in the system s state space .
the words small perturbation refer to a perturbation with small measure in the riemannian variety of the @xmath42 s .
the word stability refers to the fact that if we apply a small perturbation to a steady state then the perturbation relaxes back to the initial state ; @xmath101 ( which is zero before the perturbation ) takes a non - zero value with a well - defined sign ( possibly oscillating in time ) , and after a while relaxes back to zero . to start with , we observe that the set @xmath102 of transformations , , , and maps the space @xmath103 onto itself . by construction
, @xmath102 leaves unaffected .
then , @xmath102 maps solutions of onto solutions of .
moreover , @xmath102 is both reflexive , symmetric and transitive ( as it can be shown with the help of both chain rule and the well - known properties of jacobians after cumbersome but straigthforward algebra ) , i.e. @xmath102 is a relationship of equivalence .
consequently , @xmath102 establishes a partition of the set of the solutions of in equivalence classes ; each class is unambiguously labelled by a value of @xmath0 because of .
furthermore , @xmath102 leaves @xmath8 unaffected .
then , @xmath102 maps steady states onto steady states .
different steady states of the same equivalent class correspond to different choices of @xmath104 and @xmath40 , which are mapped onto each other by @xmath102 .
outside steady states , invariance of ensures that the sign of @xmath101 is also invariant .
accordingly , @xmath102 maps stable steady states onto stable steady states .
finally , different choices of @xmath104 and @xmath40 may even correspond to different physical systems , as far as that the latter satisfy the same equation .
for example , if @xmath40 is a velocity then it is ruled by the equations of motion , which in turn contain the forces acting on the systems .
different systems are subject to different forces , and the behaviour of @xmath40 changes accordingly . however , the proof of relies on no invariance of the equations of motion .
correspondingly , @xmath102 may map different physical systems onto each other even if remains valid .
it follows that @xmath102 can map the stable steady state ( if any exists ) of a physical system which satisfies onto the stable steady state of another physical system which satisfies , both steady states having the same value of @xmath0 .
the basins of attraction of stable states may be different in different physical systems , as the equations of motion may be different ; but as far as these equations include , our argument remains unaffected .
for example , @xmath105 maps thermodynamic equilibria ( @xmath78 ) onto thermodynamic equilibria ( @xmath106 ) .
now , in sec .
[ sec5 ] we describe one particular ( pivotal ) physical system which satisfies and which enjoys the following property : a necessary condition for the stability of a steady state against small , slowly evolving perturbations is that this state satisfies : @xmath107 under suitable constraints .
we stress the point that differs from both minep and mepp , which involve @xmath2 and not just @xmath0 .
small perturbations of a stable steady state raise @xmath0 above its value in steady state ; after a while , @xmath0 relaxes back to the initial value .
minimization in is constrained by the requirement that the solutions of are also solutions of the equations of motion in steady state .
while steady states correspond to solutions of the equations of motion in steady state , stable steady state solutions satisfy also . indeed ,
the following facts occur : a ) when we apply then invariance of forces @xmath104 and @xmath40 to get transformed under @xmath102 ; b ) all steady states obtained from the stable , relaxed state of the pivotal system via @xmath102 have the same , constant value of @xmath0 , which is a minimum in the pivotal system ; c ) all unsteady , perturbed states obtained from the unsteady , perturbed states of the pivotal system have the same , time - dependent value of @xmath0 , which is larger than the minimum value of @xmath0 in the pivotal system ; d ) @xmath102 leaves the sign of @xmath101 unaffected , so that relaxing states are mapped onto relaxing states . together , these facts lead to the conclusion that once the pivotal system is given , the same necessary criterion for stability holds for all physical systems which satisfy and which are obtained from the pivotal system via . indeed , when one of such systems relaxes back to a stable steady state after a perturbation , then the value of @xmath0 in the relaxing state and in the relaxed state respectively are equal to the corresponding values of the pivotal system , where holds .
we conclude that invariance of and implies for steady , stable , relaxed states of systems described by provided that we find a pivotal system .
noteworthy , we invoke no onsager symmetry and no detailed knowledge of the physical nature of @xmath108 and @xmath40 . a qualitative argument which invokes no pivotal system
is provided in appendix [ qual ] .
the aim of this section is to assess the impact of on the stability of relaxed states .
since we are looking for a necessary condition of stability , we are free to select the perturbations which stability is to be checked against .
we focus our attention on slowly evolving perturbations @xcite and denote with @xmath109 the typical time - scale of their relaxation in the following . for large @xmath99
we assume that ( slow diffusion approximation ) : @xmath110 according to , and , is invariant .
the impact of on stability is no new topic in the literature .
vii of @xcite links and in the particular case of the relaxation of a physical system made of two coupled subsystems at different temperatures , where joule heating is the only dissipative process in the bulk of each subsystem , so that @xmath0 is just equal to @xmath111 times the joule heating power @xmath112 ( @xmath113 joule power density ) . in each subsystem
it is assumed that , @xmath114 and @xmath73 hold .
it turns out that if holds then the relaxed state minimizes the amount of heat produced per unit time in the bulk by irreversible phenomena .
this implies that holds , as @xmath114 in each subsystem .
in contrast , if @xmath115 then we are near thermodynamic equilibrium , i.e. in onsager s regime , and minep holds .
more generally , leads to the consequences listed below . * even in the extreme case @xmath116 equation gives @xmath117 provided that @xmath118 .
physically , a finite , positive amount of entropy is produced in the bulk per unit time even if @xmath44 provides just a higher - order correction to the equations of motion like . even if seemingly rather artificial , the latter statement applies to many physical systems where dissipative phenomena like the diffusion represented by @xmath44 in allow relaxation , but are not necessarily invoked when it comes to a detailed description of the relaxed states itself .
for example , friction damps small oscillations of a pendulum around its position of stable equilibrium until the pendulum stops , but provides no information about the final position itself .
this may be true even if no conservative forces occur , like e.g. in the case of stable equilibria of magnetically confined plasmas ; indeed , these equilibria are described by grad - shafranov equation of dissipation - free magnetohydrodynamics ( mhd ) even if the relaxation processes leading to them is ruled by dissipation @xcite .
this discussion plays a crucial role in appendix [ cgleapp ] below . * _ per se _ , does not imply @xmath119 ; rather , it is an useful simplification for the investigation of slowly evolving perturbations .
this issue turns out to be relevant in sec .
[ sec6 ] below . * admittedly , this simplification comes at a price .
when computing @xmath0 in a relaxed state , we can not straightforwardly rely on equation ; we need rather an explicit expression for @xmath0 as provided e.g. by the equation of motions of the particular physical system under investigation .
this is e.g. the case of the problems discussed in both secs .
[ sec5 ] and [ sec5bis ] below . * furthermore , implies that @xmath44 provides just a higher - order correction to the dynamics of the system implies that the same holds for @xmath72 too .
it follows that our results apply even if @xmath84 within a negligible error .
the case @xmath73 is of particular interest . in this case , if @xmath73 then both @xmath69 and @xmath70 affect through @xmath44 , and have therefore no impact in the slow diffusion approximation . according to ,
we expect therefore our results below to apply to both systems with @xmath114 and @xmath83 .
in particular , such systems may belong to the same equivalence class even if the equations of motion may differ ; for example , gravitating fluids with @xmath83 may exhibit convection , which can be absent when @xmath114 .
we anticipate here that @xmath114 in the pivotal system in sec .
[ sec5 ] ; all the same , according to our discussion we shall be able to apply even to @xmath83 systems in secs .
[ sec5bis ] and [ sec6 ] . *
the quantity @xmath44 describes diffusion ; in order to maintain a stable steady state satisfying , it is therefore required that the amount of heat produced per unit time within the bulk of the system is also small ; hence @xmath0 too is small .
for example , the macroscopic fluids dealt with in secs . [ sec5 ] and [ sec5bis ] below dissipate only weakly . in the general case , , and give @xmath120 and @xmath121 where @xmath122 and @xmath123 .
if @xmath0 is small enough , it follows that : + @xmath124 + all the way around , implies . * relationships , and make @xmath40 to transform like a vector and allow us to write : + @xmath125 + where @xmath126 is either zero or the gradient of an harmonic function with vanishing normal derivative at the boundary , so that implies : + @xmath127 + all the way around , if @xmath39 satisfies then allows @xmath39 to satisfy up to an additive gradient of an harmonic function .
the aim of this section is to describe a pivotal system .
here and in the following , we limit ourselves to the @xmath54 case and we take @xmath128 as the state space . we are going to show that the electrons in the pinch of a dense plasma focus ( dpf ) @xcite @xcite form a pivotal system .
a dense plasma focus ( dpf ) is a kind of discharge in which a high - pulsed voltage is applied to a gas between electrodes , generating a short - duration , high - density plasma region .
the pinch observed at the final stage of a well - behaved dpf discharge is made of a dense , hot plasma .
this plasma is an example of classical ( i.e. , non - degenerate ) , electrically neutral , viscous , resistive , electrically conducting , magnetised fluid , where the electrons carry electric current and the ( much heavier ) ions ensure charge neutralization , so that the ratio of ion and electron densities is fixed everywhere at all times ( we refer to both densities as to particle density below ) .
the inner structure of dpf pinch spontaneously exhibits typical , complex patterns , both in the form of filaments @xcite and of plasmoids @xcite - see fig .
[ pffilamentplasmoid ] .
mm , filament thickness @xmath129 plasmoid diameter in the range @xmath130 mm @xcite .
according to @xcite , the volume of the dissipating region in the pinch shrinks to small filaments and plasmoids , in order to minimize @xmath131 ; @xmath132 in plasmoids , unlike filaments . _
_ ] the life - time of the dpf pinch is @xmath133 s @xmath134 the decay time - scale @xmath135 s predicted by resistive mhd , so we are allowed to speak of a steady state. accordingly , @xmath0 is the sum of four terms , due to joule heating , viscous heating , heat transport ( due to conduction and radiation ) and particle diffusion @xcite .
the contributions of joule heating and viscous heating are the volume integrals of @xmath111 times the joule heating power density @xmath113 and the viscous heating power @xmath136 density respectively .
both radiation and turbulence ensure efficient transport of heat across the pinch , so we may safely assume both @xmath137 and vanishing gradient of particle density ; the contributions of heat transport and particle diffusion to @xmath0 are therefore negligible . correspondingly , is equivalent to simultaneous minimization of joule heating power @xmath138 ( kirchhoff s principle @xcite ) and viscous heating power @xmath139 ( kortweg and helmholtz principle @xcite ) , i.e. to minimization of @xmath131 where @xmath140 .
now , simultaneous validity of kirchhoff s and kortweg and helmholtz principles is precisely a necessary condition of stability of steady , relaxed states of viscous , electrically conducting fluids against small , slowly evolving perturbations @xcite @xcite @xcite . as discussed below
, the equations of motion in steady state constrain the minimisation . then
, we may say that is a necessary condition of stability for the relaxed state of this system .
indeed , it has been shown in sec .
8 of @xcite that the above minimization leads to : @xmath141 in the dpf pinch . here
@xmath142 is the electric current density , @xmath143 , @xmath144 , @xmath145 is the electric charge / mass ratio of the electrons ( in the notation of @xcite the quantity @xmath146 is utilised ) , and @xmath147 is the magnetic field with vector potential @xmath148 . both @xmath142 and @xmath149
are functions of the position vector @xmath150 inside our fluid , and @xmath151 refers to partial derivatives on the components of @xmath150 .
physically , plays the role of a london equation where london depth is equal to the collisionless skin depth of electrons .
it describes successfully @xcite the filamentary structure of electric currents flowing across the dpf pinch @xcite provided that the hartmann number @xmath152 ( the ratio of electromagnetic force to the viscous force ) is lower than a certain threshold @xmath153 ( in a turbulent plasma of pure hydrogen and with current @xmath154 , @xmath155 @xcite ) .
( experiments show that the impact of velocity , which @xmath136 depends on , is definitely not negligible in dpf @xcite filaments ) .
ultimately , it is this success which justifies in this system . the case @xmath156 is discussed below and corresponds to plasmoids .
we want to show that we are actually dealing with a pivotal system . to this purpose , we show that the electrons in the dpf pinch satisfy and , or , equivalently , and . then , we link - itself a consequence of - and . for the moment , we assume to hold ; this assumption is justified below .
we apply a rescaling @xmath157 to , where @xmath158 is the total mass of dpf electrons in the pinch and @xmath159 . as anticipated in sec .
[ sec1 ] , we are going to utilize dimensionless quantities . in dimensionless form , for example
, @xmath158 is the total number of electrons .
relationships and ensure that the impact of @xmath44 remains negligible . then , and ensure that @xmath160 and @xmath161 are @xmath162 times the electron mass density and mass density flow respectively .
the pde in is just the mass balance in a system - like the electrons in a dpf pinch - where no net source of mass exists .
the boundary condition in is discussed below . as for
, the facts that @xmath137 and that @xmath163 mass density , together with equations ( 5.2 ) and ( 6.1 ) of @xcite , ensure that the pressure gradient @xmath164 is @xmath165 ; in turn , this justifies , hence .
finally , @xmath166 agrees @xcite with the fact that the contribution of mass transport to @xmath0 is negligible , as anticipated .
moreover , is just a particular case of provided that @xmath167 ( the corresponding @xmath102 is an identity ) and that we take @xmath168 and @xmath169 .
together , the latter relationship and the fact that @xmath40 transforms like a vector ensure that the vector potential transforms like a vector , as expected .
gauge invariance allows addition of the gradient of an harmonic function @xmath170 to the vector potential to leave physics unaffected in steady state . as expected from sec .
[ sec4 ] , dissipation provides no information on the relaxed state ; indeed , contains no information concerning dissipation , and the latter plays a minor role in the momentum balance .
we are left to discuss both the role of the boundary condition in and the constraints on minimization in .
for simplicity of notation , we drop all superscripts below , so that @xmath18 is the position vector inside the fluid , etc . as for the boundary condition @xmath171
, it has been assumed in @xcite that the interaction of the system with fields due to external sources is much weaker than the interaction among currents internal to the pinch . in our language , this is equivalent to neglect the impact of the external world represented by non - vanishing net flux of charge carriers across the boundaries of the system , i.e. to @xmath172 .
admittedly , the virial theorem of mhd forbids the existence of magnetically confined plasmas without interaction with external currents .
however , sec . 6 of @xcite shows that the impact of such interaction is comparable to the impact of dissipation , which is considered to be weak in agreement with .
even so , and in spite of @xmath173 , if @xmath174 then @xmath175 is not conservative and a steady state far from thermodynamic equilibrium is possible ; @xmath176 is the _curvature_ in the language of @xcite .
in contrast with the 1d systems of @xcite , however , our systems are 3d . as for the constraints on the minimization @xcite , they are provided by the mass balance quoted above , maxwell s equation of electromagnetism and navier - stokes equation with a bulk force density corresponding to the electromagnetic force . for an electrically neutral , magnetized fluid maxwell s equation reduce to gauss and ampre s law , while the electromagnetic term in the momentum balance corresponds to the density of lorenz force .
the viscous term in navier - stokes equation is as small as the impact of interaction with external currents , as stated above .
moreover , the actual value @xmath177 of @xmath69 throughout the system provides minimization @xmath178 with the further constraint throughout the system @xcite , as @xmath114 .
we stress the point that - even if electrons are the electric charge carriers in a dpf pinch - the analysis of @xcite relies on the particular nature of the electric charge carriers nowhere . in other words ,
validity of does not depend on the actual value of @xmath145 .
accordingly , we allow @xmath145 to take arbitrary real values below . if @xmath179 , makes the electron velocity to be the sum of a potential term and a quantity @xmath180 times a rotational field ; in the limit @xmath181 , a potential flow is retrieved , the structure on the small spatial scale of the electron collisionless skin depth is lost and reduces to the identity @xmath182 .
unlike , however , follows from kirchhoff and kortweg and helmholz principles , which in turn agree with for @xmath114 only ; accordingly , relaxed states with @xmath83 may have non - potential flows even if @xmath183 - see e.g. secs .
[ sec5bis ] and [ brown ] below .
admittedly , there is still a weak point in our discussion . in ref .
@xcite , the proof of kirchhoff s and korweg and helmholz principles in the dpf pinch relies on lte .
thus , one could think that validity of is limited to systems where lte holds . in order to show that this is not the case ,
let us discuss what happens if @xmath156 .
according to the same constrained minimization invoked above , it turns out that the relaxed configuration is a plasmoid @xcite , with spheroidal geometry and @xmath132 as @xmath184 , @xmath185 and @xmath186 is very large ( @xmath187 t ) @xcite ; moreover , a plasmoid is approximately described by taylor s principle of minimum magnetic energy with fixed magnetic helicity @xcite , and no velocity affects taylor s principle - see @xcite , @xcite and refs . therein .
correspondingly , @xmath188 and follows from no more , even if the velocity of charge carriers ( @xmath189 the electric current density ) may still have non - zero vorticity .
since @xmath190 the electrical resistivity @xmath191 , we expect the @xmath192-minimizing relaxed state to correspond - all other things being equal - to some minimum of @xmath191 or of some monotonically increasing function of @xmath191 .
indeed , this is precisely what happens with plasmoids , according to the independent analysis leading to equations ( 1)-(4 ) of @xcite .
we may take this fact as a confirmation of our discussion , provided that we take into account the following three facts .
firstly , it starts from investigation of plasmoids in astrophysics , not in dpf .
secondly , it does not invoke lte ; validity of is therefore not limited to systems where lte holds .
finally , it relies on a postulate in the form of a fermat - like principle concerning the whole relaxation of a reconnection - affected magnetized fluid towards a plasmoid ; as such , the final outcome depends on the initial condition ( more precisely , on the initial inverse aspect ratio of the current sheet of interest ) : as a result , a whole class of relaxed plasmoids are obtain , each of them corresponding e.g. to a given value of embedded magnetic flux , an approximately constant quantity for weakly dissipating plasmas . in conclusion , electrons in a dpf pinch
are an example of a classical fluid with viscous and joule heating , negligible mass diffusion ( i.e. , holds ) , no net mass source and negligible mass flow across the boundary ( i.e. , holds ) , and @xmath114 .
if a steady , stable state of this fluid exists , then it satisfies also the variational problem .
then , we identify this fluid with the pivotal system .
we drop the restriction @xmath114 altogether in sec .
[ sec5bis ] .
in particular , if @xmath179 ( the pivotal system is magnetized ) , @xmath114 and @xmath156 then equation follows from and reduces to the london - like equation describing a filamentary pattern of electric currents .
if @xmath179 , @xmath114 and @xmath193 then a plasmoid - like structure arises .
until now , we have assumed that fluctuations occur on time - scales much shorter than the time - scale @xmath109 of the perturbation , in order to allow us to get rid of such fluctuations through suitable time averaging .
nothing is said about the detailed frequency spectrum of the fluctuations . in this section
we discuss what happens if the fluctuations oscillate in time with period @xmath194 .
we denote -as usual by now- the generic physical quantity by @xmath51 , write : @xmath195 and introduce the time - averaging @xmath196 } \left ( t \right ) \equiv \frac{1}{\tau_{aux } } \int_{t}^{t + \tau_{aux } } d\mbox{t } \emph{y } \left ( t ' \right ) $ ] with @xmath197 , so that @xmath196 } = \emph{y}_{0 } + \emph{y}_{20 } \cos \left ( \frac{2 \pi t}{\tau } \right ) + \ldots + o \left ( \frac{\tau_{aux}}{\tau } \right ) $ ] and @xmath198 } = \frac{d \widehat{\left[\emph{y}\right]}}{d t } + o \left ( \frac{\tau_{p}}{\tau_{aux } } \right ) $ ] .
the same result holds if we replace @xmath199 with a generic function of time evolving on a typical time - scale @xmath109 .
accordingly , time - averaging just erases all contributions of the dynamics on the fast time - scale @xmath200 and allows us to replace @xmath51 with @xmath196 } $ ] everywhere while leaving all relationships involving @xmath51 and its time derivative unaffected -up to a small error @xmath201 at least . to the best of author
s knowledge , this argument has been invoked for the first time in chap .
xv of @xcite for systems satisfying minep .
up to the same error , we are therefore allowed to say that a steady , stable ( unstable ) state is a configuration where the value of @xmath51 undergoes fluctuations with time - scale @xmath200 and where @xmath196 } $ ] is stable ( unstable ) against small perturbations evolving on a time - scale @xmath202 .
we retrieve physical examples of such configurations when investigating e.g. systems described by with @xmath203 periodic function of time .
these systems may relax to oscillating states , where both amplitude and period @xmath200 of the oscillations depend on the detailed structure of @xmath204 .
we may ask if these oscillating states are stable against perturbations evolving on time - scales much longer than @xmath200 .
now , both , , and remain valid even if @xmath203 depends on @xmath8 - as it can be shown with step - by - step repetition of the proof of equation ( 10 ) in @xcite .
maps @xmath105 still leave @xmath8 unaffected , and still holds .
it follows that @xmath205 } \right ) ' = \widehat{\left[\emph{y}'\right ] } $ ] ; if @xmath206 then @xmath207 } ' = \widehat{\left[\pi\right]}\ ] ] i.e. @xmath208 } $ ] too is invariant . of course , the instantaneous value of @xmath0 in a stable relaxed state coincides at all times with @xmath208 } $ ] if and only if the stable relaxed state is a steady state , i.e. in the limit @xmath209 . in the general case , the whole argument concerning the pivotal system still holds after replacement of @xmath0 with @xmath208 } $ ]
thus , and lead to the following necessary condition for stability : @xmath210 } = \min\ ] ] if @xmath203 becomes constant , @xmath211 and we retrieve as expected ; we limit therefore ourselves to the discussion of below .
after time - averaging of both its sides , gives @xmath208 } = \widehat{\left[\phi\right ] } $ ] because @xmath212 } = 0 $ ] as @xmath25 too is periodic in time ; @xmath213 } $ ] may differ from zero even if @xmath214 } = 0 $ ] because of correlations .
remarkably , the invariance property depends on no actual value of @xmath93 , and the same class of equivalence may contain both macroscopic systems ( like e.g. the pivotal one ) and mesoscopic systems , i.e. @xmath105 may map macroscopic and mesoscopic systems onto each other .
we may think e.g. of a rescaling with a suitably defined value of @xmath215 , so that @xmath93 changes as discussed in sec .
if we apply @xmath105 to a macroscopic system and , as a result , @xmath93 gets shrinked too much , the linear dimension of the resulting @xmath93 may become of the same order of magnitude of the de broglie length of the resulting system , and the latter is definitely mesoscopic . all the way around
, we could start from a mesoscopic system satisfying ( and also , as far as stability against slow perturnations is concerned ) and obtain a macroscopic fluid with no net mass source , no mass flow across the boundary and negligible particle diffusion ( in agreement with ) .
as far as @xmath0 is invariant , the problems of stability of a star and of a molecular motor may therefore be dual to each other , and share a common solution .
of course , a star produces an immensely larger amout of entropy per unit time ; but the actual value of @xmath0 is not relevant to the problem of stability , just its minimum property counts .
( analogously , stability of thermodynamic equilibrium requires just maximisation of @xmath25 , and does not depend on the actual value of @xmath25 ) .
( displayed by the central arrow in the figure ) of transformations given by , , , and maps a and b onto each other while leaving unaffected .
the only physical quantities entering the entropy balance which are preserved by @xmath105 are @xmath0 and @xmath44 .
both a and b may be either macroscopic or mesoscopic . in order to investigate thermodynamic properties of relaxed states
, we assume that for each system the set of solutions of ( dotted region ) includes a subset of relaxed configurations ( horizontal lines ) .
the actual state the system relaxes to depends on the detailed dynamics of the system , the initial conditions etc . in turn
, the subset of relaxed configuration may further include a subset ( vertical lines ) of configurations which are both relaxed and stable against slow perturbations , evolving on some time - scale @xmath109 .
for each system , relaxed states may either be steady states of oscillating states with period @xmath216 .
since we limit ourselves to discuss necessary conditions of stability against slow perturbations , we may invoke the slow diffusion approximation , which means that the impact of @xmath44 on is small - again , a condition invariant under @xmath105 .
any classical fluid with no net mass source , no mass flow across the boundary and negligible diffusion of mass is an example of macroscopic system satisfying and . as far as both a and b satisfy
, @xmath105 maps relaxed states onto relaxed states , and stable ones onto stable ones , regardless of the detailed dynamics of a and b ; for example , a and b can be a magnetized macroscopic fluid and a mesoscopic system respectively .
furthermore , if a pivotal system ( say , a ) exists where a necessary condition for stability of a relaxed state is constrained minimization of @xmath217 } } $ ] , the constraints being given by the equations of motion , then the same condition holds for b ; even if the equations of motion of a and b are different , minima of @xmath217 } } $ ] in a correspond to minima of @xmath217 } } $ ] in b. we show that a pivotal system exists , and apply the resuling conditions for stability to different systems , both macroscopic and mesoscopic . finally , a qualitative discussion in appendix [ qual ] suggests that the actual existence of a pivotal system is not even required . _
_ ] we may therefore ask ourselves if , given an oscillating relaxed state in a mesoscopic system , there is some reformulation of which applies to the stability of such state .
we need this reformulation for two reasons : a ) a mesoscopic system which belongs to an equivalence class may still satisfy and , but @xmath47 and @xmath39 are no more to be given the meaning of mass density and mass flow respectively ; b ) the impact of fluctuations should be explicitly taken into account .
if it exists , such reformulation may hold for both macroscopic and mesoscopic systems .
this is the topic of this section . to start with
, we stress the point that the proof of relies on . in turn , the proof of relies on the positiveness of @xmath0 . in average , this positiveness follows from the fact that at the trajectory level , realizations with positive entropy production are exponentially more likely than the corresponding realizations with negative entropy production @xcite .
the probability of observing a reverse behaviour , i.e. an entropy production opposite to that dictated by the second law of thermodynamics , decreases exponentially as the system size increases .
thus , careful discussion of the averaging procedure is required when discussing mesoscopic systems .
the quantity @xmath208 } $ ] depends on the path @xmath218 , i.e. the actual trajectory followed by the system in phase space starting from the configuration at @xmath219 and arriving at the configuration at the time @xmath8 ( with no loss of generality , we suppose time averaging is computed on the time interval from @xmath220 to @xmath8 here and in the following . ) .
we define @xcite @xcite the path ensemble average @xmath221 of the path - dependent , generic quantity @xmath222 where the sum is over the set of all paths connecting all possible initial and final states and the probability @xmath223 of the path @xmath218 depends on both the detailed dynamics of the system and the probability of the initial state .
the smaller the system , the more likely the probability of observing a reverse behaviour , the more relevant the distinction between @xmath208 } $ ] and @xmath217 } } $ ] . in contrast , we may safely write @xmath217 } } = \widehat{\left[\pi\right]}$ ] in macroscopic systems , where reverse behaviour is extremely unlikely . accordingly , and regardless of the detailed mechanics of fluctuations , we may provide the following definitions , just as anticipated in sec .
[ sec1 ] : * the relationship @xmath217 } } - \widehat{\left[\pi\right ] } = 0 $ ] defines a macroscopic system ; * the relationship @xmath217 } } - \widehat{\left[\pi\right ] } \neq 0 $ ] defines a mesoscopic system .
now , by definition a system which is initially in a stable state relaxes back to its initial configuration after a small perturbation has been applied .
this means that -after a sufficiently long time at least- both the @xmath218 s and the corresponding @xmath223 s are unaffected .
this property holds as far as the system is stable . since we are looking for necessary conditions of stability , we are free to choose the quantity to be perturbed in order to check stability . with no loss of generality ,
let us consider a quantity @xmath224 which is a function of @xmath18 such that the transformation of coordinates @xmath225 is a diffeomorfism with positive jacobian @xmath215 .
accordingly , any small perturbation @xmath226 of @xmath224 on a time - scale @xmath109 is equivalent to a small perturbation @xmath227 of @xmath215 on the same time - scale .
the perturbation @xmath228 corresponds to an infinitesimal @xmath105 which maps stable states onto stable states .
if we apply it to a stable state , then the @xmath223 s remain unaffected and @xmath229 .
this relationship remains valid even if we repeat the transformation @xmath228 again and again , so that we may take it to be valid even for a finite @xmath62 .
if @xmath230 } $ ] , then we obtain @xmath231 } } \right ) ' = \overline{\widehat{\left[\pi\right ] } ' } = \overline{\widehat{\left[\pi\right]}}$ ] where the last equality follows from .
if the system we start from is macroscopic , then @xmath217 } } = \widehat{\left[\pi\right]}$ ] .
stability implies , hence : @xmath232 } } = \min\ ] ] as anticipated , the crucial point is that -even if the starting system is macrosopic- the final system needs not to be macroscopic ; is a generalization of , valid for both mesoscopic and macroscopic systems . as for mesoscopic systems ,
an open question in st research is whether the probability distributions of work , heat and entropy production can be grouped into `` universality classes '' characterized , e.g. , by the asymptotics of such distributions , and , if yes , which specific features of a system determine this class @xcite . by now , it is temptative to identify each `` universality class '' with the equivalence class labeled by the value of @xmath217 } } $ ] which the mesoscopic system belongs to . as for macroscopic systems , @xmath217 } } = \widehat{\left[\pi\right]}$ ] and we retrieve as expected ; we limit therefore ourselves to the discussion of below .
it is clear , by now , that a suitable choice of the value of @xmath158 allows a rescaling to map any mesoscopic system which obeys and , i.e. and , onto a pivotal system for some value of @xmath145 , so that the system belongs to an equivalence class and its stable relaxed states satisfy . as for
, @xmath47 and @xmath39 are to be given the meaning of mass density and mass flow respectively no more , in contrast with what occurs in macroscopic systems , as anticipated in sec .
[ sec6ter ] . as for
, it is equivalent to take into account slowly evolving perturbations only , as anticipated in sec .
[ sec4 ] .
the proof of relies on the existence of a pivotal system , through and .
if , furthermore , @xmath114 everywhere and @xmath233 at all times ( as it happens e.g. when the system is in contact with just one heat reservoir at temperature @xmath69 ) , then an independent , qualitative confirmation of exists - see appendix [ qual ] - which invokes no pivotal system and shows its agreement with st ; see fig .
[ provaschema ] for an overall sketch of our results .
until now we have dealt with the pivotal system , where @xmath114 . generally speaking , if a macroscopic , @xmath114 system is a fluid with joule and viscous heating in relaxed , steady state then agrees with kirchhoff s and kortweg and helmholtz principles invoked in sec .
[ sec5 ] . remarkably , however , the arguments leading to invoke @xmath114 nowhere . in particular
, we invoke the argument presented in sec .
[ sec4 ] : if @xmath73 then @xmath69 affects through @xmath44 only ; then , in the slow diffusion approximation @xmath69 leaves the validity of unaffected - and the same holds for @xmath70 . in the following
, we are dealing with applications of to systems where @xmath234 . in these systems , and in in contrast with the pivotal system , the transport of energy is not so strong to flatten @xmath70 .
for example , different boundaries of the same system may be in thermal contact with heat reservoirs at different temperatures @xcite @xcite @xcite .
in such systems , far - from - equilibrium relaxed states are possible even if @xmath40 is conservative .
the aim of this section is to discuss a class of macroscopic systems , namely fluids with joule and viscous heating where rayleigh - bnard convection occurs , where @xmath234 and with negligible mass diffusion , no net mass sources and no mass flow across the boundary .
convection is triggered when the rayleigh number @xmath235 - basically , a dimensionless version of @xmath236 - exceeds a threshold @xmath237 .
we discuss mesoscopic systems with @xmath234 in sec .
[ sec6 ] .
physically , joule heating and viscous heating are related to the transport of different quantities , namely electric charge and momentum respectively .
correspondingly , they provide @xmath20 with different , additive contributions and different thermodynamic flows . heat transport provides a further , additive contribution with a further thermodynamic flow .
entropy production in our fluid is therefore a compound process @xcite , and validity of mepp is questionable - see appendix [ meppapp ] .
we show in appendix [ cgleapp ] that a @xmath238 steady , relaxed , possibly unmagnetized fluid belongs to the same class of equivalence of a particular pivotal system .
more to the point , we can map a rayleigh - bnard problem with its own boundary condition ( dirichlet , neumann or a linear combination of the two ) in a steady , relaxed fluid on the verge of convection onto the steady , relaxed filamentary configuration of a dpf , and vice - versa . differences between the dynamics of the two systems do not matter ; what is relevant here is that such mapping is possible .
the proof takes into account the fact - hinted at in sec .
[ sec4 ] - that the description of a stable relaxed state involves no dissipation ; the latter affects rather the description of the relaxation process , which is very slow in the slow diffusion approximation and may therefore safely being described as a succession of steady states .
this result , together with the above - quoted fact that the impact of non - zero @xmath70 on stability is negligible in the slow diffusion approximation , allows us to invoke , i.e. the version of for macroscopic systems , as a necessary condition of stability for relaxed states in rayleigh - bnard convection problems .
the problems of the stability of the relaxed steady states of the dpf filaments and the @xmath238 fluid are dual to each other . in the latter fluid @xmath40 plays a role of a fluid velocity , and @xmath239 as convection cells rotate , say , with period @xmath200 .
a far - from - equilibrium relaxed state is possible as @xmath234 .
rayleigh - bnard convection is the result of the simultaneous occurrence of gravitation ( which , _ per se _ , leaves the entropy balance unaffected ) and of a temperature jump across a fluid layer .
both are absent in the pivotal fluid of sec .
in contrast , these systems share the following , common features : a ) mass balance holds with no net mass source and no mass flow across the boundaries , i.e. holds with @xmath47 mass density , @xmath39 mass density flow ; b ) viscous and joule dissipation are the only heating processes of the bulk ( joule heating plays a role in magnetized fluid ) , while the impact of particle diffusion on entropy production is negligible - in agreement with and - so that @xmath0 includes a new term due to heat conduction .
the entropy balance of a small mass reads : @xmath240 where @xmath241 is the heat flow due to conduction and radiation @xcite .
if no mass flows across the boundary , reynolds transport theorem @xcite and integration of both sides of on the volume of the system gives back with @xmath242 and @xmath243 after invoking gauss theorem of divergence and after integration by parts .
both @xmath244 and @xmath245 are @xmath246 and the latter is an increasing function of the former ; then , implies minimization of @xmath247}$ ] .
we invoke this minization in order to discuss a rayleigh - bnard problem below .
we are going to retrieve some results , which include the well - known variational principle of chandrasekhar @xcite at @xmath248 .
our argument is a slightly modified version of the argument presented in sec .
vi.b of @xcite . here
the novelty is that we invoke lte no more ; we have replaced it with the assumption of no mass flow across the boundary .
it is clear , by now , that our pivotal system acts just as a trojan horse which allows us to smuggle the stability criterion into the world of the relaxed solutions of and . even in the framework of the slow diffusion approximation ,
the relevance of does not depend on the detailed physics .
rather , it follows from the very existence of a partition of the set of the solutions of in equivalence classes - in other words , on the invariance of @xmath0 under .
a far - reaching corollary follows . both gravitation and non - zero @xmath70 leave and and unaffected . then
, we may apply our discussion above to the relaxed state of any classical fluid with joule and viscous heating , negligible mass diffusion , no net mass source and negligible mass flow across the boundary .
regardless of @xmath70 , stability of the relaxed states of this fluid requires .
the detailed dynamics of the system appears in the constraints of minimization , and depend on the particular problem .
now , let us come back to rayleigh - bnard convection .
the contribution of heat conduction to @xmath0 increases with increasing @xmath236 , for not too large @xmath236 at least ( the precise meaning of not too large is given below ) ; it vanishes only if @xmath114 everywhere - see sec . 49 of @xcite and sec .
66 of @xcite .
accordingly , implies minimization of @xmath249 . as for the constraints on minimization
, we want to retrieve the constraints given by the equations of motion and by @xmath178 in the limit of vanishing @xmath70 .
it can be shown @xcite that the correct choice of constraints for the @xmath234 which satisfy this condition is given by the equations of motion and by a given value of @xmath131 .
the equations of motion include the mass balance with no net mass source and the navier - stokes equation ; in an electrically conducting fluid , maxwell s equation of electromagnetism are also included , and the density of lorenz force appears in the navier stokes equation .
no mass flows anywhere across the boundaries .
after time - averaging , the energy balance obtained from the equations of motion ensures that the time - averaged total heating power is equal to the time - averaged mechanical power supplied by buoyancy ( with density @xmath250 ) ; this relationship acts as a constraint on minimization .
under such constraint , minimization of @xmath251 } $ ] retrieves the result derived in secs .
33 and 43 of @xcite from the analysis of the equations of motion near the onset of rayleigh - bnard s convection cells in both unmagnetised and magnetised fluids .
indeed , the author keeps the value of @xmath70 fixed , so that the quantity actually minimized is @xmath236 : @xmath252 } = \widehat { \left [ \int p_{b } d\mbox{v } \right ] } \ ] ] since @xmath253 in the problems solved in @xcite , the words not too large above mean @xmath254 .
some consequences of are discussed in @xcite . here
we apply to the description of the relaxed state of a macroscopic , rotating spherical shell - with inner radius @xmath255 , outer radius @xmath256 , thickness @xmath257 , volume @xmath258 and rotation period @xmath259 - of a viscous , electrically conducting fluid where a non - vanishing temperature jump drives convective cells with typical linear size @xmath260 rotating with period @xmath200 and typical rotation velocity @xmath261 .
we are allowed to invoke as rotation affects unaffected .
this is e.g. the case of the subphotospheric , convection - ruled zone of rotating stars @xcite , where typical values of the rossby number @xmath262 @xcite are @xmath263 @xcite .
this convection zone is usually supposed to be a spherically symmetric shell where convection and dynamo effect @xcite sustain a temperature gradient and a magnetic field * b * respectively @xcite . in the following ,
we derive a scaling law for * b * from . , from fig .
19 of @xcite .
a tight dependence between magnetic field and @xmath264 is clearly observed in sun - like ( crosses ) and early- to mid - m type stars ( other dots ) .
the result is that magnetic activity is rising with decreasing @xmath264 as long as @xmath265 . at @xmath266 , activity saturates and
does not grow further with decreasing @xmath264 .
this behavior is interpreted as increasing dynamo efficiency with faster rotation in the @xmath265 regime where rotation is not yet dominating convection . at fast rotation ( @xmath267 )
the dynamo reaches a level of saturation that can not be exceeded even if the star is spinning much faster . _ _ ] the reciprocity principle of variational calculus - see sec .
ix.3 of @xcite - makes the solution of to solve also the variational problem : @xmath268 } = \max \qquad \mbox{with the constraint of fixed } \vert \nabla t \vert\ ] ] further discussion requires two auxiliary assumptions , to be dropped below .
firstly , we neglect viscosity , i.e. we replace @xmath269 with @xmath270 where @xmath271 is the electric field and @xmath272 is supposed to be a scalar .
faraday s law @xmath273 gives @xmath274 where @xmath275 .
secondly , we focus our attention on the case @xmath276 , so that we make a small error if we replace @xmath277 with @xmath278 everywhere inside the shell . for example , according to @xcite solar dynamo occurs in a narrow layer @xmath279 between the convection and the radiation zone ( @xmath280 is the solar radius ) . as a further , more fundamental hypothesis
, we assume small electromagnetic fluctuations , so that we may neglect contributions of dynamo to @xmath113 @xcite and write @xmath281 in the relaxed configuration of the shell .
thus , we write @xmath282 with @xmath283 , hence : @xmath284 } \approx \dfrac{v_{s } l_{c}^2 ro^2 < \vert \textbf{b } \vert > ^2}{\eta _ { * } \emph{p}^2}\ ] ] cowling s theorem @xcite forbids occurrence of dynamo in two - dimensional systems : convection cells in the dynamo - relevant zone are therefore likely to exhibit non - trivial three - dimensional structure .
azimuthal periodicity provides us with an upper bound @xmath285 .
then , maximization in requires that we replace @xmath260 with @xmath286 in , and the constraint in gives : @xmath287 where @xmath288 } } { v_{s } } } $ ] .
here we have implicitly assumed that maximization affects just @xmath260 and not directly @xmath186 , which in turn means that we neglect possible saturation of dynamo at very large rotation velocity , i.e. , very low @xmath264 .
if , furthermore , the centrifugal force is not too strong , then the dependence of @xmath289 on @xmath264 is weak .
in fact , both @xmath290 , @xmath291 , @xmath292 and @xmath250 depend rather - through @xmath255 , @xmath293 , @xmath294 and @xmath295 - on the rotation - independent chemical composition and mass of the star , as well as on its global balances of energy and momentum , i.e. the hydrostatic balance of gravity and pressure gradient , etc .
the same holds for the value of @xmath236 in .
moreover , @xmath296 is likely to be the same for the stars of the same spectral class , so that leads to : @xmath297 in appendix [ rossbyapp ] we discuss some independent estimates of @xmath298 which confirm and which do not rely on the auxiliary assumptions referred to above .
indeed , observations @xcite for f and g / k dwarf stars are in agreement with @xmath299 , where @xmath300 is the coronal field and @xmath301 . a review @xcite which includes both sun - like and m - type stars shows that fits data as far as @xmath302 , while the above - neglected dynamo saturation occurs at @xmath303 - see fig .
[ bvslogro ] .
we have discussed macroscopic systems so far .
in contrast , when the number of particles becomes small , like e.g. in molecular machines @xcite , the environment itself continuously drive the system away out of equilibrium and the fluctuations it generates over the system are very relevant , so that the probability of observing an apparent violation of the second principle becomes significant .
such small machines will spend part of their time actually running in reverse , i.e. now and then it is possible to observe that these small molecular machines are able to generate work by taking heat from the environment , in agreement with kuramoto s identity @xcite .
the generation of mechanical work out of thermal fluctuations requires a ) either a time - dependent modulation or another form of energy flux , b ) or some spatial asymmetry with a temperature gradient @xcite . in st , cases a ) and b ) correspnd to systems which are coupled to one ( or several ) heat bath(s ) of constant temperature each respectively .
the aim of this section is to apply in the description of stable operation of such machines in a particular class of problems in the case a ) .
[ brown ] below is devoted to b ) . in case a ) @xmath114 and particles
diffuse in a potential @xmath75 , which is periodic both in space and time @xcite @xcite @xcite .
the particle distribution satisfies a fokker - planck equation which is solved by a particle distribution @xmath47 and where @xmath304 , just as in @xcite , and the relaxed state ( if any ) may oscillate in time .
as for the boundary condition , the system remains confined at all times inside a region of the state space , just like in sec .
the equations of chemical kinetics acts as equations of motion and provide the constraints .
since @xmath114 and both the volume and the chemical composition are fixed ( particle diffusion is neglected in agreement with ) , the free helmholtz energy @xmath305 too is constant . since
@xmath114 and the slow diffusion approximation allow us to neglect the amount of entropy produced per unit time by heat transport and particle diffusion respectively , the minimization in reduces to minimization of the path - ensemble - averaged , time - averaged total dissipated power . indeed , it has been shown that the evolution of a particular kind ( flashing ratchet see fig . [ flashingratchet ] ) of molecular motor - where the potential oscillates in time between zero and a periodic , asymmetric function of position - follows a path of _minimum dissipation of the total energy_ with the constraint of given @xmath305 - see equation ( 4 ) of @xcite . admittedly , the treatment of ref .
@xcite deals mainly with the 1d case , but in its own words _many of our results apply when @xmath36 is a bounded domain in @xmath306_. the same principle has been applied to more complicated molecular motors - see equation ( 4.4 ) of @xcite ; on a qualitative base , generalization to a fokker - planck system with @xmath54 chemical species seems possible - see appendix [ qual ] . is both uniform and constant .
the distribution function @xmath47 provides information on the position of particles ( the small spheres in the figure , which displays a purely 1d motion for simplicity ) located inside a finite volume and subject to a potential @xmath75 . in turn , @xmath75 is controlled by the external world , is periodic both in space and time , and is asymmetric in space . in the figure
, time follows the blue arrows . at all times particles
undergo both thermal diffusion and drift motion ( under the effect of @xmath307 ) . at t = 0
( top ) particles prefer to stay near the minima of @xmath75 ( red and blue particles are displayed for clarity near distinct minima of @xmath75 , and @xmath75 is displayed as a grey , undulated surface ) .
as time goes by ( middle ) , the external world flattens @xmath75 , and particles undergo thermal diffusion only . after a period ( bottom )
, the external world restore the potential @xmath75 at t = 0 ; by now , however , particles may be shifted on average .
the overplay of @xmath75 and @xmath44 plays a crucial role even if @xmath44 is small .
maximum stability against thermal ( hence diffusion - related ) fluctuations corresponds to minimum averaged amount of power to be supplied by the external world in order to compensate losses - see text . _ _ ] mathematically , the analysis of @xcite relies on the results of @xcite .
physically , @xmath137 in the flashing ratchet ; moreover , the latter is compared to a overdamped mass - string - dashpot system where viscous dissipation only occurs ; the crucial point is that _molecular motors move in a very viscous environment_ in the words of sec .
2.2 of ref .
@xcite , so that minimization of the dissipated power resembles kortweg - helmholtz principle invoked in the @xmath137 dpf pinch - see sec .
[ sec5 ] .
crucially , @xmath44 never vanishes exactly , even if it is small ; smallness of diffusion means just that we are focussing on slowly relaxing perturbations , as anticipated in sec .
rather , the overall behaviour of the system turns out to be the outcome of the interplay between the tendency of particles tovards concentration at the minima of the potential and diffusion , which tends to spread and dissipate density .
if @xmath69 is uniform and constant , and if the ratchet keeps on oscillating in a stable way - so that the average amount of mechanical energy exchanged with the external world is also constant - then the same holds for @xmath305 - this explains the constraint . once the ratchet has started oscillating at a given period , physical intuition requires that the external world must do some work per period on the system in order to compensate the effect of thermal fluctuations , which tend to destroy the ordered oscillatory behaviour of the ratchet ; ratchet stability against fluctuations is maximum when the required time - averaged amount of power is a minimum . a final remark . in st ,
if a @xmath114 system is in thermal contact with a heat reservoir at temperature @xmath69 , equations ( 3 ) and ( 7 ) of @xcite - together with the fact that @xmath0 is the volume integral of @xmath20 - show that @xmath308 } } $ ] is equal to the time - averaged amount of work done on the system by the external world divided by @xmath309 .
the same equations ensure also that @xmath308 } } > 0 $ ] even if the discussion of @xcite takes explicitly into account the possibility that @xmath20 may not be positive - definite .
our result provides information on @xmath83 systems too . again , and in analogy with the discussion of macroscopic systems in sec .
[ sec5bis ] , if @xmath73 then @xmath69 affects through @xmath44 only ; then , in the slow diffusion approximation both @xmath69 and @xmath70 leave the validity of unaffected . for example
, feynman describes @xcite a ratchet coupled via an axle to a paddle wheel immersed in a fluid . because of the brownian motion of molecules , the paddle wheel has an equal probability to be rotated to the right or the left .
owing to the presence of the pawl preventing one rotation direction , say the left , the ratchet will rotate on the right , seemingly violating the second law of thermodynamics .
the apparent contradiction is solved and equal rotation probability restored by including thermal fluctuations of the ratchet . however ,
if its temperature is smaller than that of the fluid , unidirectional rotation becomes possible , enabling conversion of heat into work . rather than an applied ,
oscillating potential , it is therefore the non - uniformity of temperature which allows the motor to work .
no @xmath40 is required anymore ; for simplicity we are going to take @xmath310 altogether . the aim of this section is to apply to the description of stable , relaxed states of such mesoscopic system , henceforth referred to as brownian motor. to this purpose , we follow the treatment of @xcite and @xcite . in a system made by @xmath311 parts , the @xmath312th part ( @xmath313 ) being in thermal contact with a heat bath at given temperature @xmath314
, we are allowed to invoke as far as the system remains confined at all times inside a region of the state space and @xmath47 obeys a conservation equation like - see equations ( 3)-(4 ) of @xcite ; as for both the self - consistency of slow diffusion approximation and the physical meaning of @xmath47 , see below .
minimization in implies minimization of the contribution of each part to @xmath217 } } $ ] separately . in turn
, this implies minimization of the path - ensemble - averaged , time - averaged total dissipated power in the @xmath312th part for all @xmath315 s , because in the bulk of the @xmath312th part no entropy is produced by heat transport ( as @xmath316 across the @xmath312th part ) and the entropy produced by particle diffusion is negligible in the slow diffusion approximation . finally , and in analogy with our discussion of flashing ratchets above , minimization of
the path - ensemble - averaged , time - averaged total dissipated power follows - as the latter is just the sum of the minimized amounts of power dissipated in each part .
to fix the ideas , let us suppose that the @xmath312th part is immersed in a fluid heat bath and undergoes a viscous drag with friction coefficient @xmath317 , where the @xmath312th coefficient @xmath318 of proportionality depends on the detailed geometry and orientation in space of the @xmath312th part - see fig .
[ brownmotor ] , as well as equation ( 12 ) of ref .
@xcite and equation ( 34 ) of @xcite , where the fluids are perfect gases .
the @xmath312th part is free to change its own orientation in space during the operation of the motor .
the amount of power dissipated in the @xmath312th part is an increasing function of @xmath319 ; then , minimization of the path - ensemble - averaged , time - averaged total dissipated power in each part - required by - implies minimization of @xmath320 for all @xmath80 s . this minimization occurs as all the parts adjust their own orientation in space .
this orientation is unambiguosly described by a set of euler angles , which are the @xmath321 s of the state space of the system , and @xmath47 , the probability distribution of orientation , depends on both the @xmath321 s and time .
if only 1 part out of @xmath311 is free to change orientation then there are 3 euler angles ; the state space is @xmath322 $ ] and can be linked with @xmath128 by a diffeomorphism .
it is useful to define the quantity @xmath323 ; it is an increasing function of fluctuation amplitude - see equation ( 2 ) of @xcite . since @xmath324 , minimization of @xmath320 implies minimization of @xmath325 .
parts ( here @xmath326 for simplicity ) . during the motion of the system ( not displayed in the figure )
, it is assumed that the @xmath312th part ( @xmath313 ) is in thermal contact at all times with a heat bath at temperature @xmath314 , and may undergo a viscous drag with a orientation - dependent viscous coefficient @xmath320 .
the values of the @xmath314 s are given , and differ from each other .
moreover , the @xmath312th part has its own shape ( here , a triangle and a rectangle are displayed for @xmath327 and @xmath328 respectively ) . here
, part 1 is free to change its own orientation ; for example , the tip of triangle @xmath329 may point either horizontally ( as displayed in the figure ) or vertically . on average ,
dissipation gets spontaneously minimized through changes in this orientation .
_ _ ] we have seen that minimization of @xmath325 is a necessary condition for the validity of in our system .
we show that is also a sufficient condition .
indeed , the facts that @xmath324 and that @xmath330 make the minimization of @xmath325 to imply the minimization of viscous heating power . since the latter is just the sum of the amounts of viscous heating power dissipated in each gas and each amount is non - negative , minimization of each amount separately follows . this minimization is achieved through changes in the detailed geometry of each part @xcite , hence all @xmath331 s remain unaffected . as a consequence , minimization of the amount of viscous heating power in each part implies minimization of the amount of entropy produced in each part ;
after summing on all parts and averaging , follows . indeed , it has been shown - in the particular case where all euler angles but one are blocked at least - that : a ) a molecular motor whose parts are in thermal contact with perfect gases acting as heat baths at different temperatures tends to minimize @xmath325 ; b ) this minimization implies also the minimization of the diffusion coefficient , in agreement with the slow diffusion approximation ; c ) the relaxed state may violate minep - see equations ( 5 ) , ( 7 ) and the comment to equation ( 18 ) of @xcite .
reults c ) holds in analogy with the results of @xcite concerning joule heating and referred to in sec .
the second principle of thermodynamics provides information concerning a physical system at thermodynamic equilibrium ; such information do not depend on the choice of coordinates with the system is described with .
far from thermodynamic equilibrium , the same holds for the description provided by glansdorff and prigogine s minimum entropy production principle ( minep ) for the stable , relaxed state of a macroscopic system which satisfies onsager symmetry @xcite .
( conservatively , here we stick to the familiar gibb s entropy ) . to date , no proposed extremum property for far - from - equilibrium , stable relaxed states of systems which violate onsager symmetry seems to enjoy such invariance .
for example , a first - principle proof @xcite of the so - called maximum entropy production principle ( mepp ) @xcite seems to rely precisely on violation of this invariance .
physical intuition , in contrast , makes us to look for extremum properties - for stable , far - from - equilibrium relaxed states - which are invariant under change of coordinates .
the price to be paid is likely to be the loss of generality .
indeed , the results of @xcite suggest that variational principles concerning particular contributions to the entropy balance may hold for selected classes of problems .
another kind of invariance - namely , invariance of microscopic physics under time reversal - plays a crucial role in stochastic thermodynamics ( st ) , which provides a framework for extending the notions of classical thermodynamics to the level of individual trajectories across the space state of a selected class of ( usually mesoscopic ) physical systems , namely those systems where a non - equilibrium process occurs which is coupled to one ( or several ) heat bath(s ) of constant temperature . remarkably , st is often concerned with oscillating relaxed states , in contrast with minep and mepp , which deal with steady relaxed states . rather than to conditions of stability for stable relaxed states , st leads to a large number of exact , experimentally confirmed results - usually referred to as fluctuation theorems ( fts ) - starting from the invariance of the microscopic dynamics under time - reversal .
fts allow to quantify the probability of observing occasional reverse behaviour , i.e. i.e. an entropy production opposite to that dictated by the second law of thermodynamics ; it decreases exponentially as the system size increases @xcite @xcite @xcite .
the success of fts suggests to take advantage of the properties of invariance of a system while looking for some property of its stable , far - from - equilibrium relaxed states .
in particular , it seems worthwhile to investigate the consequences of the invariance under diffeomorphisms involving the coordinates of the space state . generally speaking
, these diffeomorphisms can map macroscopic systems ( where the volume of the system in the space state is large enough and fluctuations can be neglected in some cases at least ) and mesoscopic systems ( where the volume is small and fluctuations , possibly of quantum origin , can not be neglected altogether ) onto each other .
correspondingly , if such a property actually exists and is invariant under change of cooordinates then it provides information on the stability of both macroscopic and mesoscopic systems .
moreoverm given the fact that relaxed states of mesoscopic systems are often oscillating states , this - so far hypothetical - property should describe oscillating relaxed states - and steady states as a particular , zero - frequency case - in order to cope with both macroscopic and mesoscopic systems .
a connection could be established between different problems of stability .
finally , of course , this property should agree with st .
the aim of this paper is twofold : a ) to write down this invariant property for the relaxed states of a selected class of physical systems , and b ) to discuss some applications to stability .
we limit ourselves to systems described by a pde problem made of a simple fokker - planck equation and a boundary condition ( the system remains confined at all times within a given region of the state space ) .
it has been shown @xcite @xcite @xcite that if a generic , differentiable transformation @xmath332 of both the coordinates which the probability distribution function @xmath47 depend upon and of the quantities involved in leaves both and time unchanged , then the amount @xmath0 ( often referred to as @xmath16 in the literature ) of entropy produced per unit time by irreversible processes inside the bulk of the system is invariant under @xmath102 .
we have investigated the consequences of this invariance . to start with ,
each @xmath332 is a relationship of equivalence , hence it establishes a partition of the set of the solutions of in equivalence classes ; each class is unambiguously labelled by a value of @xmath0 .
secondly , given the possibility of writing in dimensionless form ( and of avoiding therefore problems with dimensional analysis ) , it follows that different elements of the same class of equivalence may represent different physical systems with the same value of @xmath0 , as far as all of them satisfy even if the other details of the dynamics are different . for further discussion ,
firstly we discuss the necessary condition for the stability of a steady relaxed state of a macroscopic system ; generalization follows .
if relaxation occurs towards some final , relaxed , steady state which is stable against small perturbations then unsteady , relaxing states are mapped onto unsteady , relaxing states and relaxed , stable states are relaxed onto stable , relaxed states . for example , thermodynamic equilibria are mapped onto thermodynamic equilibria ( @xmath78 ) . in particular ,
if at least one ( pivotal ) of the systems which satisfy exists where constrained minimization of @xmath333 ( the constraints being provided by the equations of motion ) is a necessary condition for the stability of relaxed states against small , slowly evolving perturbations , then the same property is shared by the stable , relaxed states of all physical systems which are obtained from the pivotal system via a map @xmath102 ; even if the detailed equations of motion may be different , the minimized quantity is the same .
this corollary follows ultimately from the invariance of @xmath0 , not from validity of onsager symmetry nor from the detailed nature of the involved systems .
further discussion takes advantage of the fact that we are looking for necessary conditions for stability , hence we are free to select the perturbation which stability is to be checked against . for simplicity , we assume that the perturbation is slow , i.e. that its typical time scale is longer than other typical time - scales of the dynamics of the system ( for example , the oscillation period in a oscillating relaxed state ) .
this is equivalent to assume that the diffusion term in the fokker - planck equation is relatively small ( diffusion stands for irreversible phenomena which damp perturbations of stable states ) ; this approximation is referred to as slow diffusion approximation below .
now , we show that the electrons in the pinch of a dense plasma focus ( dpf ) @xcite form a pivotal system . depending on the relative relevance of electromagnetic and viscous force @xcite ,
the spatial structure of the system may either filamentary or plasmoid - like @xcite .
stable , steady relaxed configurations minimize @xmath333 .
( experiments @xcite @xcite suggest that these structures are indeed exceptionally stable against resistive decay as predicted by magnetohydrodynamics ( mhd ) , so that the words stable and steady seem to be justified ) .
generalization from steady to oscillating relaxed states of macroscopic systems requires replacement of @xmath333 with its time - average @xmath208 } $ ] , the time - average being taken on a time scale which is both much longer than one period of oscillation of the relaxed state and much shorter than the typical relaxation time scale of the perturbation @xcite .
analogously , generalization to mesoscopic systems requires further replacement of @xmath208 } $ ] with its path - ensemble average @xcite @xmath217 } } $ ] .
these generalizations rely on the fact that the map commute with both time - averaging and path - ensemble - averaging , near a stable relaxed state at least .
when it comes to mesoscopic systems , we could say that the equivalence classes referred to above act as `` universality classes '' which these systems can be grouped in .
a striking consequence of the generalisation to mesoscopic systems is that the problems of stability of macroscopic and mesoscopic systems may be dual to each other , i.e. share a common solution .
the problem of stability against slow perturbations of the relaxed state of a mesoscopic system described by which remains confined at all times within a given region of the state space ( e.g. , a brownian motor ) and the problem of stability of the relaxed state of a weakly dissipative fluid with no net mass source , no mass flow outside the boundary and negligible particle diffusion ( e.g. , the subphotospheric , convection - ruled , dynamo - affected layer of a g - class rotating star ) are dual to each other . even if the star produces an immensely larger amount of entropy per unit time , it is not the actual , averaged value of @xmath0 which is relevant to the problem of stability , but its minimum property .
( analogously , it is the maximum property of entropy @xmath25 , not its actual value , which is relevant to the stability of thermodynamic equilibrium ) .
qualitative arguments even suggest that constrained minimization of @xmath217 } } $ ] are indeed in agreement with st , even if no pivotal system is available - see appendix [ qual ] .
this minimization is therefore the looked - for invariant property of stable relaxed states .
[ provaschema ] displays a general overview of the results of our discussion . in a nutshell ,
invariance bridges the gap between the problems of stability of some selected macroscopic and mesoscopic systems . in order to discuss applications ,
we start with the following remark . as far
as we investigate their stability against slowly evolving perturbations , a macroscopic , classical fluid - with joule and viscous heating , negligible mass diffusion , no net mass source and negligible mass flow across the boundary - and a mesoscopic system satisfying - with @xmath128 as the state space - enjoy the same property : both can be mapped onto a pivotal system . in the macroscopic fluid , equation reduces to a simple mass balance ; in the mesoscopic system , investigation of stability against slow perturbations allow us to invoke the slow diffusion approximation .
consequently , they share also the same necessary condition of constrained minimization of @xmath217 } } $ ] for the stability of relaxed states . the equations of motion provide the constraints , which depend therefore on the particular system of interest .
it does not matter what the detailed dynamics is like ; what is relevant here is that such mapping is possible .
examples of the macroscopic fluids quoted above are the filamentary configuration of the electrons in the dpf pinch and a ( possibly unmagnetized ) fluid near the onset of rayleigh - bnard convection ; indeed , it is shown in appendix [ cgleapp ] that the problems of stability of the two systems are dual to ech other , even if the dynamics is utterly different . beyond thermodynamic equilibrium ( where of course @xmath217 } } \equiv 0 $ ] ) , in the following we are going to discuss four particular problems as particular cases of the macroscopic and the mesoscopic systems hinted at above . in each case , we have derived from the constrained minimization of @xmath217 } } $ ] a necessary condition of stability , not necessarily in the form of a variational principle .
the criterion of stability we obtain depends on the particular problem .
noteworthy , both minep and mepp apply in no case . in particular , according to the language of ref .
@xcite all macroscopic systems we deal with in the following are compound systems , where mepp fails .
this is not surprising , as both minep and mepp are concerned with the entropy produced per unit time by all irreversible processes in the system , while @xmath0 is the amount produced per unit time by the irreversible processes only which occur in the bulk . in all cases we retrieve results previously available in the literature , independently derived from either observations or dedicated proofs from first principles .
._classification of some necessary criteria of stability of relaxed states against slow perturbations for systems which obey the equation @xmath334 with @xmath172 at the boundary . here
@xmath47 , @xmath335 and @xmath49 are a distribution function , a probability current density and the unit normal vector perpendicular to the boundary of the region @xmath336 ( with measure @xmath93 ) where the system is confined ; moreover , @xmath44 and @xmath40 are a diffusion coefficient and a vector field respectively .
we denote by @xmath196 } $ ] and @xmath337 the time average and the path - ensemble - average @xcite respectively of the generic quantity @xmath51 .
following @xcite , we introduce the amount @xmath338 of the boltzmann - gibbs entropy @xmath25 of the system produced in the bulk of the system , where @xmath5 is boltzmann s constant .
all necessary criteria are retrieved as particular examples of the constrained minimization of @xmath217 } } $ ] , the constraints being provided by the equations of motion in each case .
stable , relaxed states of macroscopic ( mesoscopic ) systems correspond to @xmath217 } } - \widehat{\left[\pi\right ] } = 0 \left ( \neq 0 \right ) $ ] .
stable , relaxed , steady ( oscillating ) states correspond to @xmath208 } - \pi = 0 \left ( \neq 0 \right ) $ ] .
macroscopic systems of interest have no net mass source , no mass flow across the boundary and negligible mass diffusion , provided that @xmath47 and @xmath39 are mass density and flow respectively . _ [ cols="<,^,^,^,^",options="header " , ] [ tabellasinottica ] results are collected in table [ tabellasinottica ] .
the list is not - and is not meant to be - complete .
each column of table [ tabellasinottica ] refers to a different problem far from thermodynamic equilibrium , where each of the quantities on the left may either vanish ( @xmath339 ) or differ from zero ( @xmath28 ) . for each problem
, results are discussed below ; we start from the minimization of @xmath217 } } $ ] and retrieve a corresponding necessary criterion of stability , which finds its confirmation in the refs . referred to in the first line of the table . for comparison
, @xmath23 is the necessary criterion of stability of thermodynamic equilibrium ( @xmath340 ) .
\(a ) electrons in a dpf pinch @xcite . rather than the identically vanishing @xmath70
, it is @xmath341 which keeps the stable , steady relaxed state far from thermodynamic equilibrium in this macroscopic system .
here @xmath40 is a suitably defined linear combination of velocity and vector potential and @xmath163 mass density .
( steady means here with lifetime much longer than predicted by mhd ) .
the necessary criterion of stability of the relaxed state is @xmath342 with fixed @xmath177 @xcite , where @xmath269 is the sum of viscous heating power and joule heating power density , and @xmath177 is the value of the temperature @xmath69 at the boundary .
if viscous heating is ( not ) negligible in comparison with electromagnetic force , then a plasmoid - like configuration ( a filamentary structure ) arises . in the plasmoid case @xcite , th enecessary criterion of stability
a monotonically increasing function of resistivity achieves a minimum - in agreement with the results of @xcite ; moreover , taylor s principle of minimum magnetic energy for given minimum helicity is approximately satisfied @xcite . in the filamentary case @xcite , current filaments are similar to those of a superconductor @xcite
; the relaxed state belongs to the same class of equivalence of a ( possibly unmagnetized ) fluid near the onset of rayleigh - bnard convection , in spite of the different equations of motion and of the different values of @xmath70 .
\(b ) rayleigh - bnard convection in a rotating shell of magnetized fluid affected by dynamo .
the subphotospheric , convection - ruled zone of a rotating star @xcite provides an example .
here @xmath40 and @xmath47 are @xmath55 the velocity and the mass density repsectively . both @xmath83 and @xmath341
keep the oscillating relaxed state of this macroscopic system far from thermodynamic equilibrium ; rotation of both the shell and the convective cells rule the oscillation frequencies .
it turns out that the absolute value of @xmath70 is minimized , with the constraint that the time - averaged dissipated power is equal to the time - averaged mechanical power provided by buoyancy ; this result had been previously found in @xcite under the more restrictive assumption of local thermodynamic equilibrium @xcite , and is in agreement with @xcite . as a consequence ,
the averaged magnetic field is inversely proportional to the rossby number , in agreement with observations for f and g / k dwarf stars @xcite @xcite .
\(c ) flashing ratchet .
an externally applied , oscillating potential keeps this mesoscopic system far from thermodynamic equilibrium , even if @xmath114 and @xmath79 ( @xmath304 , where @xmath75 is an externally applied potential , which is periodic both in time and in space , and is asymmetric in space ; @xmath47 is a particle distribution ) . in order to sustain regular ratchet oscillations ,
the external world must do some work per period on the system in order to compensate the effect of thermal fluctuations , which tend to destroy the ordered oscillatory behaviour of the ratchet ; ratchet stability against fluctuations is maximum when the required time - averaged amount of power is a minimum .
it turns out that stable relaxed states minimize the dissipated power with the constraint of given helmholz free energy .
this is in agreement with the results of ref .
@xcite .
\(d ) brownian motor .
this mesoscopic system is made of @xmath311 parts ( @xmath313 ) , where the @xmath315-th part is immersed in a fluid heat bath at temperature @xmath314 and viscous coefficient @xmath320 . rather than an applied ,
oscillating potential , it is therefore the non - uniformity of temperature which allows the motor to work , by enabling conversion of heat into work .
the differences between the constant @xmath314 s keep the system far from thermodynamic equilibrium .
stable relaxed , oscillating states minimize a a quantity ( the effective temperature @xmath325 ) - which has the dimension of a temperature and which depends on the @xmath320 s - under the constraint of given @xmath314 s .
this is in agreement with the results of ref .
@xcite , where @xmath310 and @xmath47 refers to the probability of orientations in space ; this treatment stems from the seminal analysis of feynman @xcite .
we have shown that invariance against changes of coordinates plays a crucial role when it comes to understand many different criteria for stability for dissipative macroscopic and mesoscopic systems . in a nutshell , invariance allows us to see that the problem of the stability of the far - from - thermodynamic - equilibrium relaxed state of a physical system may be dual to the corresponding problem in another system , even if the two systems may differ considerably .
then , once the problem has been solved for one system , the solutions of the problems for a whole class of physical system is at hand .
here we have discussed invariance of suitably averaged @xmath0 for a wide class of systems .
there is no fundamental reason to think that invariance of other quantities for other classes of systems is impossible , and that it leads to no further criteria for stability . finally , it is clear , by now , that our pivotal system acts just as a trojan horse which allows us to smuggle the stability criterion into the world of the relaxed solutions of under the slow diffusion approximation .
dpf pinch electrons ( column ( a ) in table table [ tabellasinottica ] ) are by no means special : once the existence of the pivotal system is proven , its detailed dynamics is not relevant , as expected in a thermodynamic approach .
we could as well start e.g. from the systems described in columns ( b , ( c ) or ( d ) , where the stability criteria have been speparately and independently established .
it is even possible that no pivotal system is really required - as suggested by the qualitative discussion of appendix [ qual ] - and that a deeper connection exists between invariance and stability exists .
thus , further extension of the method outlined here to other problems is conceivable , and will be the matter of future work .
fruitful discussions with prof .
m. polettini ( univ . luxembourg ) and dr .
e. cosatto ( ansaldo energia , genova ) are warmly acknowledged .
widest - label p. glansdorff , i. prigogine _ thermodynamic theory of structure , stability and fluctuations _ wiley , london ( 1971 ) p. glansdorff , i. prigogine , physica * 30 * , 351 ( 1964 ) u. seifert , rep .
. phys . * 75 * , 126001 ( 2012 ) d. j. evans , d. j. searles , advances in physics , * 51 * , 7 , 1529 - 1585 ( 2002 ) t. tom , braz .
* 36 * , 1285 ( 2006 ) .
t. tom , m. j. de oliveira , phys .
e * 82 * , 021120 ( 2010 ) g. a. casas , f. d. nobre , e. m. f. curado , phys .
e * 86 * , 061136 ( 2012 ) d. m. carberry , s. r. williams , g. m. wang , e. m. sevick , d. j. evans , j. chem .
* 121 * , 17 , 8179 - 8182 ( 2004 ) s. r. williams , d. j. evans , e. mittag , comptes rendus physique , * 8 * , 5 , 620 - 624 ( 2007 ) g. e. crooks , phys .
* e * * 61 * , 3 , 2361 ( 2000 ) i. neri , e. roldan , f. juelicher , _ statistics of infima and stopping times of entropy production and applications to active molecular processes _ arxiv:1604.04159v3 [ cond-mat.stat-mech ] 4 oct 2016 l. d. landau , e. lifshitz , _ fluid mechanics _ pergamon , oxford ( 1960 ) r. j. tykodi , _ thermodynamics of steady states _
macmillan , new york , ( 1967 ) c. van der broeck , physics of complex colloids * 184 * , 155 - 193 ( 2013 ) l. m. martyushev , v. d. seleznev , phys .
rep . * 426 * , 1 ( 2006 ) l. m. martyushev , v. d. seleznev , phys .
a * 233 * , 410 , 17 - 21 ( 2014 ) a. di vita , phys .
e * 81 * , 041137 ( 2010 ) i. gyarmati , _ non - equilibrium thermodynamics _
springer , berlin ( 1970 ) r. balescu , trans .
fusion technol .
* 25 * , 105 ( 1994 ) e. barbera , continuum mechanics and thermodynamics * 11 * , 5 , 327 ( 1999 ) h. ziegler , j. appl .
( zamp ) * 34 * , 832 ( 1983 ) m. polettini , entropy * 15 * 2570 ( 2013 ) doi : 10.3390/e15072570 g. grinstein , r. linsker , j. phys .
40 * , 9717 ( 2007 ) r. dewar , j. phys . a : math
. theor . * 36 * , 631 ( 2003 ) e. t. jaynes , phys . rev . *
106 * , 620 ( 1957 ) g. p. beretta , reports on mathematical physics volume * 64 * , 1 - 2 , 139 - 168 ( 2009 ) g. c. li , m. r. von spakovsky _ steepest - entropy - ascent quantum thermodynamic modeling of heat and mass diffusion in a far - from - equilibrium system based on a single particle ensemble _ arxiv:1601.01344v2 [ cond-at.stat-mech ] g. p. beretta , entropy * 10 * , 160 ( 2008 ) doi : 10.3390/entropy - e10030160 y. sawada , prog .
. phys . * 66 * , 68 ( 1981 ) j. s. kirkaldy , metall .
a * 16a * , 10 , 1781 ( 1985 ) a. hill , nature * 348 * , 426 ( 1990 ) l. m. martyushev , v. d. seleznev , i. e. kuznetsova , zh .
. fiz . * 118 * , 149 ( 2000 ) j. a. sekhar , j. mater .
sci . * 46 * , 6172 ( 2011 ) a. kleidon , naturwissenschaften * 96 * , 653 ( 2009 ) a. kleidon , k. fraedrich , t. kunz , f. lunkeit , geophys .
* 30 * , 23 , 2223 ( 2003 ) doi : 10.1029/2003gl018363 r. d. lorenz , phil . trans .
b * 365 * , 1349 ( 2010 ) doi : 10.1098/rstb.2009.0312 n. virgo , entropy * 12 * , 107 ( 2010 ) doi : 10.3390/e12010107 z. yoshida , s. m. mahajan , phys .
plasmas * 15 * , 032307 ( 2008 ) l. m. martyushev , phil
b * 365 * , 1333 ( 2010 ) u. lucia , physica a * 393 * , 256 ( 2013 ) e. rebhan , phys .
a * 42 * , 781 ( 1990 ) j. b. marston , annual reviews of condensed matter physics * 3 * , 285 ( 20128 ) j. j. qian , phys .
* 29 * , 1305 ( 1996 ) m. polettini , eur .
. lett . * 97 * , 30003 ( 2012 ) d. jou , j. casas - vazquez , g. lebon , rep .
. phys . * 51 * , 1105 ( 1988 ) a. di vita , j. thermodynam . cat . * 3 * , 4 ( 2012 )
doi : 10.4172/2157 - 7544.1000e108 m. polettini , f. vedovato _ of dice and men , subjective priors , gauge invariance , and non - equilibrium thermodynamics _
european thermodynamics conference , brescia , italy , july 1 - 5 ( 2013 ) @xmath344 v. i. smirnov , _ a course in higher mathematics
_ mir , paris ( 1977 ) d. andrieux , p. gaspard , j. chem .
phys . , * 121 * , 6167 ( 2004 ) d. andrieux , p. gaspard , j. stat .
phys . , * 127 * , 1 , 107 ( 2007 ) j. schnakenberg , revies of modern physics , * 48 * , 4 , 571 ( 1976 ) j. p. freidberg , _ ideal mhd _ cambridge univ . press , cambridge , ( 1987 ) h. qian ,
e * 65 * , 021111 , 571 ( 2002 ) y. nakayama , k. kawaguchi , phys .
. soft matter phys . * 91 * 1 , 012115 ( 2015 ) a. di vita , eur .
j. d * 54 * , 451 ( 2009 ) a. bernard , h. bruzzone , p. choi , h. chuaqui , v. gribkov , j. herrera , k. hirano , a. krejc , s. lee , c. luo , f. mezzetti , m. sadowski , h. schmidt , k. ware , c.s .
wong , w. zoita , j. moscow phys .
* 8 * , 93 ( 1998 ) l. soto , plasma phys .
fus . * 47 * , a361 ( 2005 ) h. herold et al . ,
fusion * 29 * , 1255 ( 1989 ) l. jakubowski et al .
* 32 * , 1871 ( 2002 ) w. h. bostick , w. prior , j. plasma phys .
* 8 * , 1 , 7 - 20 ( 1972 ) a. di vita , eur .
j. d * 56 * , 205 - 208 ( 2010 ) l. soto , c. pavez , f. castillo , f. veloso , j. moreno and s. k. h. auluck , phys .
plasma * 21 * , 7 , 072702 ( 2014 ) j. b. taylor , phys .
lett . * 33 * , 1139 ( 1974 ) l. comisso , m. lingam , y. m. huang , a. bhattacharjee , phys .
plasmas * 23 * , 100702 ( 2016 ) j. b. taylor , b. mcnamara , phys .
fluids * 14 * , 1492 ( 1971 ) l. d. landau , e. lifshitz , _ electrodynamics of continuous media _ pergamon , oxford ( 1960 ) a. di vita , eur . phys .
j. d * 67 * , 191 ( 2013 ) e. t. jaynes , ann .
phys . chem .
* 31 * , 579 ( 1980 ) h. lamb , _ hydrodynamics _ cambridge university press , cambridge , uk ( 1906 ) b. andrae , j. cremer , t. reichenbach , e. frey , phys .
lett . * 104 * , 21 , 218102 ( 2010 ) m. polettini , a. wachtel , m. esposito . the journal of chemical physics *
143*.18 ( 2015 ) : 184103 .
m. polettini , phys .
e * 84 * , 051117 h. froehlich , proc .
. soc . * 87 * , 330 ( 1966 ) l. d. landau , e. lifshitz , _ statistical physics - theory of the condensed state _
pergamon , oxford ( 1960 ) i. s. aronson , l. kramer , rev .
modern phys . *
74 * , 1 , 99 ( 2002 ) j.e .
wesfreid , y. pomeau , m. dubois , ch .
normand , p. berge , j. physique * 39 * 725 - 731 ( 1978 ) s. chandrasekhar , _ hydrodynamic and hydromagnetic stability _
clarendon press , oxford , ( 1961 ) e. r. priest , _ solar magnetohydrodynamics _ reidel publ .
co , dordrecht , holland , ( 1982 ) c. jordan , b. montesinos , mon . not .
. soc . * 252 * , short communication , 21p26p ( 1991 ) d. j. bercik , w. p. abbett , g. h. fisher and y. fan , _ turbulent magnetic field generation in rotating stars _ american geophysical union , fall meeting 2003 a. reiners , living rev . solar phys .
* 8 * , 1 ( 2012 ) n. a. sinytsin , j. phys .
theor . * 42 * 193001 ( 2009 ) r. jordan , d. kinderlehrer , f. otto , siam j. math .
29 * , 1 , 1 - 17 ( 1998 ) k. moffatt , _ dynamo theory _ the encyclopedia of astronomy and astrophysics ( institute of physics publishing and nature publishing group ) 666 - 671 g. belvedere , m. r. e. proctor and g. lanzafame , nature * 350 * 481 - 483 ( 1991 ) i. v. elsgolts , _ differential equations and variational calculus _ mir ,
paris , ( 1981 ) a. h. boozer , j. plasma phys . * 35 * , 1 , 133 - 139 ( 1986 ) j. w. connor and j. b. taylor , nucl . fus . * 17 * , 5 1047 - 1055 ( 1977 ) u. r. christensen and j. aubert , _ scaling laws for dynamos in rotating spherical shells and application to planetary magnetic fields _ gtp workshop on modeling mhd turbulence ; applications to planetary and stellar dynamos at ncar , 27 - 30 june 2006 , boulder co , usa j. dolbeault , d. kinderlehrer , m. kowalczyk , _ the flashing ratchet : long time behavior and dynamical systems interpretation _ , technical report 0244 , ceremade ( 2002 ) j. m. r. parrondo , phys . rev .
e * 57 * , 6 , 7297 - 7300 ( 1998 ) a. feigel , a. rozen , _ self - assembly of brownian motor by reduction of its effective temperature _ arxiv:1312.5279v1 [ cond-mat.stat-mech ] 18 dec 2013 m. chipot , d. kinderlehrer , m. kowalczyk , meccanica , * 38 * , 5 , 505 - 518 ( 2003 ) p. meurs , pascal , ch .
van den broeck , a. garcia .
review e * 70 * , 5 ( 2004 ) 051109 .
r. p. feynman , b. r. leighton , m. sands , _ the feynman lectures on physics _
vol . 1.46 ( addison - wesley , 1964 ) .
mepp meets no general consensus yet . for example
, ziegler s version @xcite of mepp has _its statistical substantiation only if the deviation from equilibrium is small_ @xcite and is criticised in @xcite . according to ref .
@xcite , ziegler s mepp can be considered just as a working hypothesis for nonlinear nonequilibrium thermodynamics .
moreover , validity of mepp is questionable in the so called compound processes. by definition , the entropy production of a compound process is a sum of functions dependent only on some ( but not on all ) of the thermodynamic fluxes , and the fluxes which one of these functions depends upon do not affect the other functions .
simultaneous chemical and thermal processes can be the simplest example of compound processes : the former are scalar and the latter are vector ; and according to curie s principle , these processes can not influence each other .
finally , the mepp - supporting results of @xcite , which rely on the so - called information thermodynamics approach @xcite , are criticized in @xcite . in particular , it has been observed @xcite that information thermodynamics provides the fundamental tenet of equiprobability of microstates with no physical basis similar to what is provided e.g. by liouville s theorem in equilibrium thermodynamics ; in the harsh words of @xcite , the author of @xcite _swept the dirt under the carpet_. according to a further , steepest entropy ascent approach @xcite @xcite , irreversibility in nature is a fundamental microscopic dynamical feature and as such it must be built into the fundamental laws of time evolution . in this approach ,
mepp ultimately arises from a fundamental non - unitary extension of standard schroedinger unitary dynamics . in the words of @xcite ,
_every trajectory unfolds along a path of steepest entropy ascent compatible with the constraints _ [ ... ] .
_ for an isolated system , the constraints represent constants of the motion . _
[ ... ] _ the qualifying and unifying feature of this dynamical principle is the direction of maximal entropy increase . _
[ ... ] _ the challenge with this approach is to ascertain if the intrinsic irreversibility it implies at the single particle ( local , microscopic ) level is experimentally verifiable , or else its mathematics must only be considered yet another phenomenological tool_. recent research aims precisely at comparison of predictions based on steepest descent approach and experiments where relaxation is postulated to follow a particular law - we refer to @xcite and in particular to its equation ( 12 ) . when it comes to classical physics , the constraints acting on the relaxation identify a hyper - surface in the phase space of the system . in the words of @xcite , _on
this surface we can identify contour curves of constant entropy , generated by intersecting it with the constant - entropy surfaces .
every trajectory _ [ ... ] _ lies_ on this surface _and is at each point orthogonal to the constant entropy contour passing through that point . in this sense ,
the trajectory follows a path of steepest entropy ascent compatible with the constraints_. however , this conclusion relies on the postulate that relaxation obeys a particular law -see equation ( 25 ) of @xcite , retrieved also in equation ( 12 ) of @xcite .
admittedly , general arguments supporting mepp are discussed in @xcite .
remarkably , however , this work seems to identify a property of @xmath15 , rather than of @xmath2 .
in fact , equation ( 7 ) of @xcite suggests maximisation of the amount per unit time of the entropy of the universe due to exchange between the relaxed system and the external world - in our language , maximisation of @xmath345 rather than of @xmath2 . the same limitation seems to apply to many alleged applications of mepp , including metallurgy @xcite , crystallization @xcite and crystal growth @xcite , solidification @xcite , physics of the atmosphere @xcite @xcite @xcite @xcite and plasma physics @xcite .
for example , the growth of a crystal is obviously affected by what happens at its boundaries .
correspondingly , the selection criterion discussed in @xcite is related to the _conjecture that each lamella must grow in a direction which is perpendicular to the solidification front_. analogously , the results of @xcite are related to the growth velocity via the _rate of entropy production per unit area_ , which is a surface quantity like @xmath15 and unlike @xmath0 .
finally , in @xcite a crystal grows and changes its surface so as to select a flow which maximises the entropy production , @xmath345 being the only contribution to the entropy balance which depends on both a flow and the surface of the system . as for solidification ,
the relevant quantity is the _entropy generation term which depends on the solid - liquid region_ @xcite , i.e. on an interfacial boundary . as for the atmosphere , the only contribution to the entropy balance investigated in @xcite , @xcite , @xcite and @xcite is due to energy fluxes across boundaries
- see equations ( 5 ) and ( 14 ) of @xcite , equations ( 1 ) and ( 2 ) of @xcite , fig . 1 of @xcite and equation ( 1 ) of @xcite - and is therefore to be identified with @xmath345 ;
in contrast , the agreement of mepp with models of a stratified atmosphere including viscosity ( so that @xmath0 is not negligible ) seems to be poor @xcite . as for plasmas , what is actually maximized in equation ( 7 ) of @xcite is the inhomogeneity of temperature in a boundary layer , i.e. a quantity related to @xmath15 .
in contrast with mepp , different extremum properties for @xmath0 and @xmath15 seem to hold for relaxed states of different physical systems under different boundary conditions -see e.g. sec . 3 of @xcite for a particular example and @xcite for a more general discussion , which includes some of the results quoted above .
for instance , the latter discussion involves also an extremum property similar to what is proposed in @xcite , but concerning @xmath345 rather than @xmath2 ; such property is suggested in @xcite for a shock wave in the limit of large mach number , where the volume integral @xmath0 is as vanishingly small as the volume occupied by the shock front .
we follow the treatment of @xcite and @xcite . the results listed below
have been derived in @xcite in the framework of the research concerning the models @xcite @xcite based on schnakenberg s treatment of master equation @xcite .
equations , and give : @xmath346 after integration by parts - for details , see the proof of equation ( 10 ) in @xcite for a boltzmann - gibbs entropy . the first term on the r.h.s .
of is the only term which is @xmath246 in all cases , because @xmath347 .
then , comparison of and leads to and .
according to @xcite and @xcite , is invariant whenever , and hold .
indeed , @xmath44 is a scalar , then holds .
moreover , the definition of @xmath215 allows us to write : @xmath348 now , invariance of the normalization condition @xmath12 and equation imply , i.e. @xmath47 transforms like a scalar density .
finally , invariance of @xmath349 and equation imply , i.e. @xmath39 transforms like a vector density . in turn , equations , and have a number of consequences .
firstly , invariance of our boundary condition follows from . secondly , and imply @xmath350 , i.e. @xmath25 is not invariant -see equation ( 8) of @xcite .
thirdly , and reynolds transport theorem @xcite give @xmath351 \mbox{d}v $ ] . together with , integration by parts and gauss theorem of divergence lead to @xmath352 . in turn , the latter result , together with and , leads to : @xmath353 i.e. @xmath2 is not invariant - see equation ( 11 ) of @xcite .
fourth , , , , and lead to - see equation ( 10 ) of @xcite .
finally , after cumbersome algebra equations , , , and lead to .
this result can also be obtained straightforwardly by observing that both @xmath44 and the quantity @xmath354 in are invariant according to , , and .
in contrast with @xmath0 , the lack of invariance of @xmath2 , together with equations and , show that @xmath15 too lacks invariance .
we can rewrite in a sligthly different way for systems which exchange only heat with a thermal bath at temperature @xmath69 . in such systems , @xmath15 is equal to the amount of heat flowing per unit time from the system towards the bath divided by @xmath69 and is therefore equal to the time derivative @xmath355 of bath entropy , provided that - as usual - we neglect all irreversible processes occurring inside the bulk of the bath , i.e. we assume @xmath356 in the bath . then , @xmath357 , so that and give @xmath358 .
integration on time from @xmath219 to a time @xmath359 gives : just like and which is based upon , holds for both steady states and unsteady states . as for steady states , @xmath76 ; moreover , we denote with @xmath361 the value taken by @xmath362 . as for unsteady states ,
@xmath363 may differ from @xmath220 and we define @xmath364 .
finally , we write for both steady and unsteady state , subtract term by term and obtain now , equation is just equation ( 11 ) of @xcite ( but for a slightly different definition of @xmath25 ) ; the latter relationship deals with the invariance properties of the entropy balance in systems undergoing markovian evolution according to an equation which generalises under transformations which include also rescaling of time - scales . after multiplying @xmath366 by @xmath367 and exponentiating
, the reationship implies invariance of the dimensionless quantity @xmath368 where @xmath369 . in other words
, @xmath370 does not depend on the jacobian @xmath215 defined in sec .
[ sec2 ] . after deriving @xmath370 once and twice on @xmath215
, we obtain therefore : until now , we have assumed @xmath215 to be constant .
nothing essential changes , however , if we let it to oscillate mildly and slowly in time , say @xmath375 - think e.g. of a self - similar rescaling with slow , alternating expansion and compression .
a relaxed , stable state will be also oscillating with period @xmath376 , at least as far as the external world keeps on driving such slow oscillations .
if no dissipation occurs inside the system , then @xmath377 and @xmath378 at all times .
if dissipation occurs , our assumption of constant and uniform @xmath69 leads to @xmath379 , where @xmath380 is the amount of dissipated work from @xmath219 to the time @xmath8 .
it follows that @xmath381 and @xmath382 .
let us denote with @xmath383 and @xmath384_{r}$ ] the quantities obtained by @xmath222 and @xmath337 via time reversal respectively .
now , we may invoke the identity since our stable relaxed state is permanently oscillating , however , we may drop the subscript @xmath389 altogether in , as time reversal leaves oscillating quantitites unaffected but for a change of phase .
accordingly , leads to : now , we invoke the argument of sec .
[ sec6ter ] .
the small perturbation @xmath228 of a stable , relaxed state leaves the @xmath223 s remain unaffected , and we are allowed to write @xmath391 all along the perturbation .
accordingly , for a stable , relaxed state relationships and the definition of @xmath392 lead to : i.e. , @xmath394 is a minimum in stable relaxed states with respect to a small perturbation of @xmath215 .
the assumption of oscillating @xmath215 is relaxed in the @xmath395 limit ( i.e. @xmath215 is constant ) .
it is relaxed too , provided that we write the generic , slowly evolving @xmath396 as a sum of harmonic functions of time with different periods , provided that each term is so small that we may consider the response of @xmath394 to @xmath396 as a sum of responses to each term .
admittedly , our argument relies on an identity , , whose validity is limited by a number of additional assumptions @xcite : a ) the energy of the system is finite ; b ) the dynamics are markovian ; c ) if the system is already in equilibrium and is unperturbed , then it must remain in equilibrium ; d ) the system starts in thermal equilibrium and is driven from that equilibrium by an external perturbation . assumption
a ) and b ) are satisfied in many systems described by .
assumption c ) abd d ) may be satisfied e.g. if the system is in contact with one heat bath at temperature @xmath69 .
in particular , assumption d ) is satisfied e.g. if the external world raises slowly @xmath397 starting from a zero value at @xmath398 .
however , these assumptions are rather artificial , and their validity in many applications is questionable , to say the least . accordingly
, our argument is just a qualitative support , rather than a rigorous proof , of . in conclusion
, we discuss a result which seems to agree - qualitatively at least - with the validity of for stable relaxed states of a system which describes cyclic population dynamics of @xmath54 species , which undergoes a markovian stochastic process and which is described by an approximate fokker - planck equation @xcite .
the space of the concentrations of 3 chemical species is our @xmath399state space .
it turns out that when the external world changes a control parameter and drives the system across a supercritical hopf bifurcation from large limit - cycle - like oscillations to small erratic oscillations around a fixed point , then @xmath2 peaks very much near the transition , and decreases away from it . in our language , it is the amplitude of the oscillation which is in steady state in both configurations ; the limit cycle and the fixed point are equivalent to a stable oscillation with constant amplitude @xmath28 and @xmath339 respectively , once small - scale , fast fluctuations have been averaged out . the words
approximate fokker - planck are justified because the diffusion coefficient is not exactly constant ; however , once the system has relaxed then fluctuations of diffusion coefficient too can be averaged out .
no exchange with the external world is considered , then @xmath400 and gives @xmath401 .
as the system is driven from one stable configuration to the other one , firstly the entropy production increases , then decreases : this is consistent with our result that each stable state corresponds to a local minimum of entropy production .
admittedly , this is no rigorous result , as the value of @xmath0 computed when fluctuations are taken into account differs from the value of @xmath0 computed when fluctuations are neglected ( but in particular cases ) . remarkably , however , fluctuations may even _ lower _
@xmath0 @xcite , so that our arguments are not weakened .
independent estimates of @xmath298 confirm .
the plasma of our shell is made of has three remarkable features @xcite : a ) inter - particle collsions are far from negligible , as @xmath486 ; b ) the pressure is not negligible in comparison with the magnetic energy density @xmath487 ; c ) radiative transport of heat is relatively weak . in such plasmas , a rescaling @xmath488 by a dimensionless factor
@xmath489 leaves physics unaffected provided that the collisional vlasov high beta scaling ( cvhbs ) holds @xcite : we have neglected viscosity heating so far .
numerical simulations of an extensive set of dynamo models in rotating spherical shells @xcite take into account also viscous heating and cover a wide range of control parameters . generally speaking ,
these simulations hint at lack of explicit dependence of the magnetic field on either resistivity or rotation period , and suggest rather a strong dependence on buoyancy , in agreement with our discussion .
results are fitted by the following scaling laws : where @xmath496 , @xmath497 } $ ] and @xmath498 are a modified rossby number , a modified nusselt number and a modified rayleigh number representing the cell rotation , the heat transport and the ( suitably normalized ) amount @xmath499 of heat advected by buoyancy per unit volume and unit time respectively .
together , and the definitions of @xmath500 and @xmath501 give :
a pivotal system satisfies , and , as a consequence , ( or , equivalently , if @xmath179 ) . in order to fix the ideas , we start with the dpf filamentary structure discussed in sec . [ sec5 ] . at the end
, we shall see that a suitable rescaling maps this structure onto somewhat completely different . to start with ,
let us introduce a complex - valued function @xmath402 of position where @xmath403 is @xmath55 the ( bohm ) diffusion coefficient @xmath404 @xcite .
we neglect @xmath405 ; this assumptions is justified below .
we invoke the following two results of ref .
firstly , ampre s law ensures that both and @xmath406 are satisfied provided that @xmath407 .
secondly , both and the momentum balance ( including lorenz force density ) are satisfied , provided that : a ) a contribution @xmath408 to the pressure of the fluid ( which appears in equation ( 5 ) of @xcite and sec . 8 of @xcite ) is neglected , in agreement with ; b ) viscous terms in the momentum balance are of the same order of the interaction with external currents in sec .
[ sec5 ] and are neglected , in agreement with equations ( 3 ) of @xcite and ( 8.3 ) of @xcite ; c ) @xmath409 solves the pde problem : where @xmath411 is the speed of sound and @xmath412 , @xmath413 is the value of @xmath47 in the limit @xmath414 , @xmath415 , @xmath416 .
a non - zero value for the quantity @xmath417 ( which has the dimension of @xmath418 ) takes into account possible dissipation - free dispersion of waves across the system ; it leaves therefore our discussion concerning @xmath0 unaffected . as for the equation in ,
it is just equation ( 9 ) of @xcite , and contains no viscous term . as for the boundary condition , we refer to sec .
45 of @xcite which deals with a @xmath419 case and which corresponds precisely to vanishing flow across the boundary ; this makes sense as leads to both and to the balance of momentum regardless of the actual value of @xmath417 , thus allowing a complete description of the pivotal system . in particular , if @xmath420 - like e.g. in @xcite - then according to @xcite the equation in is formally similar to the ginzburg - landau ( gl ) equation for a superconductor @xcite .
remarkably , dissipation in a superconductor occurs mainly near the axis of a filament , i.e. where @xmath421 .
in contrast , the opposite , far - from - the - filament - axis limit corresponds to weak dissipation , in agreement with as anticipated .
( even if @xmath179 the main entropy - raising mechanisms in the bulk of of the pivotal system , namely joule heating and viscous heating , are related to diffusion of magnetic field lines and to diffusion of particles respectively .
both processes lead to dispersion of matter across space .
then , it is self - consistent to neglect dissipation - related terms in the equations of motion when holds ) . as shown in sec .
[ sec4 ] , this does not prevent @xmath0 from being @xmath422 . in the same limit , terms
@xmath423 become small - as anticipated above - and @xmath424 is ruled by the contribution of the external world , which is negligible in the pivotal system .
( gradients of @xmath417 , @xmath425 and @xmath145 vanish everywhere ) . finally , if @xmath181 , then @xmath424 plays no role and we retrieve an unmagnetized fluid where euler equation plays the role of momentum balance .
now , let us further discuss in the far - from - filament - axis region , i.e. in the slow diffusion approximation .
to this purpose , we need some intermediate relationships .
firstly , we introduce the quantities @xmath426 , @xmath427 , @xmath428 , @xmath429 and @xmath430 ( so that @xmath431 ) .
secondly , the definition of @xmath409 and imply @xmath432 and : @xmath440 \eta + \\ & + \left [ \dfrac{\vert \textbf{c } \vert ^2}{2 } - 1 - \kappa \mu \textbf{a } \cdot \textbf{j } - \dfrac{\beta r^2}{2 \gamma ^2 } \vert \nabla_{y } \ln \left ( f \right ) \vert ^2 - \dfrac{2 \beta r^2}{\gamma ^2 } \vert \nabla_{y } \left(\dfrac{\chi}{\gamma}\right ) \vert ^2 \right ] \eta \\ \end{split}\end{aligned}\ ] ] equation is equivalent to a reduced form of the equation in in the slow diffusion approximation .
it enjoys two relevant properties .
firstly , in a pivotal system the fact that @xmath0 depends explicitly on neither @xmath424 , @xmath417 , @xmath403 and @xmath411 allows the latter quantities to take arbitrary values independently from each other , and allows us to neglect their gradients across the system ; by doing so , we make an error @xmath448 . secondly , if @xmath449 then @xmath450 and the scaling @xmath451 gives @xmath452 , @xmath453 and @xmath454 .
then , reduces to a particular case of nonlinear schroedinger equation , i.e. @xmath455 .
not surprisingly , equations ( 6 ) , ( 53 ) and ( 54 ) of @xcite show that this equation leads precisely to the euler equation for unmagnetized fluids , as a term @xmath456 ( the so called quantum pressure ) is discarded because of . at last , here it comes the fundamental result of this section .
let us take @xmath457 with no loss of generality .
then , in steady state ( @xmath458 ) is the steady state version of a gl equation with real coefficients .
now , the latter equation , which we write in the form describes also the steady state of a fluid near the onset @xmath460 of rayleigh - bnard convection - see equation ( 5 ) of @xcite for the case @xmath461 and equation ( 3 ) of @xcite for the general case respectively ; here @xmath462 and @xmath463 are positive constants , @xmath464 and @xmath465 is the laplacian on some rescaled spatial coordinate @xmath241 . depending on the values of @xmath145 and @xmath403 , the boundary condition on @xmath466 may be either dirichlet ( @xmath467 ) , neumann ( @xmath468 ) or a linear combination of the two ( @xmath469 ) .
together , the fact that @xmath470 and equation allow us to find a rescaling which maps in steady state and onto each other ( gradients of @xmath471 are negligible in the slow diffusion approximation , because of ) .
in particular , once both @xmath472 , @xmath473 , @xmath474 and @xmath475 are known then in steady state reduces to provided that we identify @xmath476 and @xmath241 with @xmath477 , @xmath478 , @xmath479 , @xmath480 and @xmath481 respectively .
the reverse is also true . all the way around ,
once @xmath462 and @xmath463 are known for a fluid described by the quantity @xmath482 in the @xmath483space near the convection onset and a known boundary condition ( dirichlet , neumann or a linear combination of the two ) .
then , it is always possible to map onto a steady state version of of a pivotal system through a rescaling with jacobian @xmath215 , because we have 5 unknown quantities ( @xmath215 , @xmath472 , @xmath473 , @xmath474 and @xmath475 ) against the 4 = 3 + 1 constraints given by the 3 known values of @xmath462 and @xmath463 and by the 1 boundary condition ( which gives the value of @xmath484 ) .
there is still room for 5 - 4 = 1 degree of freedom , namely @xmath215 . in other words ,
a steady state fluid on the verge of rayleigh - bnard convection is in the same class of equivalence of a particular pivotal system , namely the steady , relaxed filamentary configuration of a dpf pinch .
it is precisely the fact that dissipation is not required for the description of the steady relaxed state - see sec .
[ sec4 ] - which allows the inviscid equation to provide information on the steady relaxed configurations of fluids with convection .
rather , dissipation describes relaxation ; the latter is very slow in the slow diffusion approximation and may therefore safely be described as a slow succession of steady states .
of course , detailed dynamics is completely different in the two systems : dpf filaments are steady only when compared with mhd time - decay , the temperature gradient across the fluid is not negligible , radiation transport of heat occurs across the dpf , and is it is not even required that a magnetic field occurs in the rayleigh - bnard case . but dynamics provides just constraints on minimization of @xmath0 ; it does not change the fact that if a steady state is stable in one system - hence @xmath485 - then the corresponding steady state is also stable in the other system - i.e. , @xmath0 remains a minimum .
this is in analogy to what happens at thermodynamic equilibrium , where the fact that @xmath25 attains a maximum does not depend on the choice of the @xmath18 s .
we may say that the problems of stability of the electric current filaments in a dpf pinch and of a fluid near rayleigh - bnard convection are dual to each other . | recently , it has been shown [ m. polettini et al . , 12th joint european thermodynamics conference , brescia , italy , july 1 - 5 , 2013 ] that diffeomorphisms in the state space of a wide class of physical systems leave the amount @xmath0 of entropy produced per unit time inside the bulk of the system unaffected . starting from this invariance , we show that if the boundary conditions allow the system to relax towards some final ( relaxed ) state , then the necessary condition for the stability of the relaxed state against slowly evolving perturbations is the same for all the systems of this class , regardless of the detailed dynamics of the system , the amplitude of fluctuations around mean values , and the possible occurrence of periodic oscillations in the relaxed state .
we invoke also no onsager symmetry , no detailed model of heat transport and production , and no approximation of local thermodynamic equilibrium .
this necessary condition is the constrained minimization of suitably time- and path - ensemble - averaged @xmath0 [ g. e. crooks , phys .
rev .
* e * * 61 * , 3 , 2361 ( 2000 ) ] , the constraints being provided by the equations of motion .
even if the latter may differ , the minimum property of stable relaxed states remains the same in different systems .
relaxed states satisfy the equations of motion ; stable relaxed states satisfies also the minimum condition .
( thermodynamic equilibrium corresponds to @xmath1 and satisfies trivially this condition ) .
our class of physical systems includes both some mesoscopic systems described by a probability distribution function which obeys a simple fokker - planck equation and some macroscopic , classical fluids with no net mass source and no mass flow across the boundaries ; the problems of stability of these systems are dual to each other .
as particular cases , we retrieve some necessary criteria for stability of relaxed states in both macroscopic and mesoscopic systems . retrieved criteria include ( among others ) constrained minimization of dissipated power for plasma filaments in a dense plasma focus [ a. di vita , eur .
phys .
j. d * 54 * , 451 ( 2009 ) ] , of electrical resistivity in plasmoids [ l. comisso et al . , phys . plasmas * 23 * , 100702 ( 2016 ) ] , of time- and path - ensemble - averaged dissipated power in flashing ratchets [ j. dolbeault et al .
, technical report 0244 , ceremade ( 2002 ) ] and of the so - called effective temperature in brownian motors [ a. feigel et al .
, arxiv:1312.5279v1 cond-mat.stat-mech 18 dec 2013 ] , as well as the scaling of the average magnetic field to the reciprocal of rossby number in dynamo - affected , convection - ruled shells of rotating stars [ c. jordan et al . ,
mon . not .
r. astr
. soc .
* 252 * , short communication , 21p26p ( 1991 ) ] . finally , we argue that the validity of the popular `` maximum entropy production principle '' [ l. m. martyushev , v. d. seleznev , phys . rep . *
426 * , 1 ( 2006 ) ] is questionable in these systems . * on the invariant quantities in an entropy balance and their impact on the stability of some physical systems * a. di vita , pacs : 05.70.ln keywords : non - equilibrium thermodynamics , entropy production |
it is known that fusion reactions within a plasma are enhanced .
this is due to the fact that ions in a plasma are not `` naked '' , but are surrounded by electrons in the plasma that form a shielding cloud around them .
thus , an ion is not repelled , as is a `` naked '' ion , since the `` dressing '' of electrons shield the repulsive potential of the ion for distances greater than the debye radius , @xmath4 .
the presence of the plasma is taken into account , in what is called the _
screening factor_@xcite , in estimating cross sections of nuclear reaction rates . in the static case ,
when the velocities of the ions are assumed to be very slow compared to those of the screening particles , the screening factor used is the salpeter factor @xcite ( valid in the weak screening limit , when @xmath5 electromagnetic fluctuations are present in a plasma , even when in thermal equilibrium . from these fluctuations ,
we evaluate the screening of the interacting particles .
their effect is to enhance the reaction rates , just as a negative potential reduces the coulomb repulsion .
we use the formalism of the fluctuation - dissipation theorem@xcite to describe the electromagnetic fluctuating fields . using the fluctuation - dissipation theorem ,
it is possible to study all the fluctuations existent in a plasma .
that is , it is possible to also include the effect of screening modes that are not eigenmodes ( i.e. , not at the plasma frequency ) .
these modes are highly damped , but although they have a small amplitudes , their overall effect turns out to be great because their phase space is very large . in particular
, they have non - negligible amplitudes for @xmath6 , where @xmath7 is the plasma frequency . in estimating the effect of electromagnetic fluctuations on the screening in the limit where only frequencies on the order of @xmath7 are important ( @xmath8 ( @xmath9 the usual salpeter factor is obtained . in a non - magnetized plasma ,
the longitudinal eigenmodes are langmuir waves with @xmath10 for usual laboratory and space plasmas , this frequency is much less than @xmath11 and the assumption that @xmath12 is reasonable .
however , if we are interested in the effect of all the existent modes in the plasma on the screening , it is important not to be limited only to modes where @xmath13 we show , in section ii , that the screening can be described in terms of the electromagnetic fluctuations . the change in the reaction rates due to the screening is calculated and our screening factor is compared with the salpeter factor .
a discussion of our results is given in section iii .
the expression for the reaction rate between ions 1 and 2 is @xcite @xmath14 where @xmath15 is the cross section and @xmath16 and @xmath17 are the thermal distributions of the particles .
the cross section is @xmath18 , where @xmath19 @xmath20 is the penetration factor , @xmath21 is the kinetic energy of the ions at large separations in the center of mass reference frame and @xmath22 is a purely nuclear factor that varies slowly with @xmath21 .
@xmath19 @xmath20 is proportional to @xmath23 @xmath24}^{1/2}dr ] $ ] , where @xmath25 is the reduced mass .
the integration is performed from the nuclear radius @xmath26 to the turning point radius @xmath27 .
@xmath28 is the bare coulomb interaction @xmath29 and @xmath30 the potential energy due to the screening . in the case of weak screening
, @xmath31 is pratically independent of @xmath32 in the integration interval .
the reaction rate @xmath33 is then found to be increased by a factor @xmath34 @xcite . in the static case , the weak screening limit , the enhancement factor is @xmath35 , the salpeter factor . in a plasma ,
electromagnetic fluctuations are always present , acting as a background for the interacting ions .
let us consider the effect of the fluctuating electric fields .
@xmath36 , the electric potential acting in the region from @xmath27 to @xmath26 , creates a potential energy @xmath37 .
the enhancement factor is , then , @xmath38 , where @xmath39 is the potential energy of the ions when they are at infinity , @xmath40 . averaging @xmath41 in time and using a taylor expansion ,
we obtain terms which are proportional to @xmath42 , @xmath43 and @xmath44 , as well as cross terms . because the fields @xmath36 , @xmath45 and @xmath46 are random with respect to one another , the only terms that remain to first order are @xmath42 , @xmath47 and @xmath48 . assuming that they are equal to one another , @xmath49 where @xmath50 .
@xmath36 is assumed to be changing slowly in the integrand of the penetration factor in the region from @xmath27 to @xmath26 .
therefore , we consider fields with frequencies @xmath51 and wavenumbers @xmath52 , where @xmath53 is the velocity of the reduced mass in the center of mass system .
we find that @xmath42 is relatively insenstive to the exact upper limits of the integrations , @xmath54 and @xmath55 .
@xmath42 is related to the fluctuations of the longitudinal electric field in a plasma : @xmath56 . therefore , @xmath57 > from the fluctuation - dissipation theorem@xcite , the expression for the intensity of the fluctuations of the longitudinal electric field is @xmath58 where @xmath59 is the longitudinal dielectric permittivity of the plasma .
we note that this is an exact quantum mechanical relation . given a dielectric permittivity that describes the plasma , it is possible , from eq .
( 4 ) , to obtain the fluctuations in the longitudinal field due to the eigenmodes ( i.e. , the electrostatic langmuir waves ) , as well as to all the other fluctuations existent in the plasma .
we show below that the overall effect of fluctuations with @xmath60 is to substantially increase the reaction rates .
although the amplitudes of these fluctuations are small , they are not negligible . due to the very large range of their frequencies ,
their overall effect is to increase the reaction rates appreciably . because the reaction rates are exponentially dependent on the screening potential , its increase is appreciable for reactions with large values of @xmath61 such as @xmath62 and @xmath63 . in the limit where only fluctuations with frequencies on the order of the eigenfrequency @xmath64 are important , eq .
( 4 ) can be easily intergrated over frequency with the aid of the kramers - kronig relations , obtaining @xmath65 this is the expression which is generally found in text books @xcite . for an electron plasma in the collisionless case , the longitudinal dielectric permittivity is @xmath66 , where @xmath67 and @xmath68 . with this dielectric permittivity , @xmath69 [ eq.(3 ) ] and the enhancement factor f , in eq.(2 ) , is the salpeter enhancement factor ,
@xmath70 we use the model described in detail in opher & opher @xcite , which includes thermal and collisional effects , to obtain the longitudinal dielectric permittivity .
it uses the vlasov equation in first order with the bgk ( bhatnagar - gross - krook ) collision term , which is a model equation of the boltzmann collision term @xcite .
the inclusion of collisions changes the results very little and the effect on @xmath71 for the plasmas studied , is almost negligible . for example , in the case analysed in opher and opher @xcite , where @xmath72 and @xmath73 , the difference in the correlation energy , with or without collisions , is less than @xmath74 .
the dielectric permittivity that we therefore use , is almost identical to the collisionless dielectric permittivity given after eq .
( 5 ) in the text ( the only difference is the inclusion of other shielding species besides the electron ) . from this model , the longitudinal dielectric permittivity for an isotropic plasma
is found to be @xmath75 where @xmath76 is the label of the species of the particles , @xmath77 is the thermal velocity of the species and @xmath78 , the fried and conte function . by using eq .
( [ el ] ) in eq .
( [ phi ] ) and eq .
( [ elf ] ) , we obtain @xmath79 the enhancement factor is obtained from eq .
we note that using a relation which is similar to eq .
( 2 ) for the transverse field @xmath80 in conjunction with the transverse dielectric permittivity instead of eq . ( 7 ) , the correct black body spectrum is obtained [ 5,6 ] .
we estimate the enhancement factor @xmath81 for the reaction rates @xmath82 , @xmath83 and @xmath63 for an electron - proton plasma at temperatures @xmath84 and densities ranging from @xmath85 choosing the maximum density for each temperature so that the plasma parameter is less than @xmath86 .
when @xmath87 approaches such high values , non - linear effects begin to be important .
( note that @xmath88 has an extra factor @xmath89 due to the inclusion of protons . )
the gamow energy , the most effective energy for thermonuclear reactions , is @xmath90 , where @xmath91 is the reduced atomic weight and @xmath92 .
therefore , for each of the temperatures that we choose , the gamow energy changes , with a consequent change in the limits of the integration , @xmath55 and @xmath54 . as a particular case , we assume the center of mass energy to be zero .
@xmath93 , where @xmath94 is the salpeter enhancement factor and @xmath87 is the plasma parameter , is plotted for @xmath95 in fig .
1 . in fig .
2 , we plot @xmath96 vs @xmath87 for @xmath97 we also estimate the enhancement factor @xmath41 for the reaction @xmath98 in an electron - deuteron plasma for temperatures @xmath84 and densities @xmath99 again , the maximum density for each temperature is chosen to be @xmath100 . @xmath96 vs @xmath87 is shown in fig .
we see that the enhancement factor @xmath41 is larger than the salpeter factor . in order to see how the fluctuating potential is affected by these modes , we plot the potential @xmath101 vs @xmath102 for the reaction @xmath63 , with @xmath95 and @xmath103 ( in this case , @xmath104 ) in fig .
the solid curve is plotted for @xmath105 , the dashed curve for @xmath106 and the dotted curve for @xmath107 .
we normalize the curves and plot @xmath108 divided by @xmath109 . when @xmath108 is integrated for all wavenumbers @xmath110 , but only for frequencies up to @xmath111 ( @xmath112 ) , we obtain the salpeter potential .
we see that for @xmath107 , the potential @xmath113 has non - negligible values only at low frequencies .
however , for @xmath106 and @xmath105 , the potential @xmath114 spreads out to high frequencies . for these high values of @xmath115 instead of dropping abruptly at @xmath116
the curves decrease slowly with frequency .
there is thus a substantial contribution for @xmath60 .
thus for large wavenumbers , ( i.e. , @xmath117 ) fluctuations make a substantial contribution at frequencies higher than the plasma frequency . for example , for @xmath118 and @xmath119 @xmath120 is not zero , but 0.002 .
similarly , for @xmath121 and @xmath122 @xmath123 is 0.01 .
by using the _ fluctuation - dissipation theorem _ to estimate the screening of the reacting ions due to the electromagnetic field fluctuations in plasma , we calculate the enhancement factor for the reaction rates .
the enhancement is found to be greater than that for the adiabatic case ( @xmath12 ) , where the salpeter factor is valid .
it is important to note that this effect is not due to a specific choice of the dielectric permittivity .
the model that we use includes the standard dielectric permitivity and gives almost the same results as does a collisionless thermal description .
a higher enhancement factor is obtained because we include the high frequencies and do not assume that @xmath124 . using this method in a previous study to derive the transverse magnetic fluctuations , the standard black body spectrum was obtained . for the plasmas studied here , electron - proton plasmas with temperatures @xmath125 and densities 1@xmath126 the reaction rates for @xmath82 , @xmath127 and @xmath63 are increased up to @xmath128 in the regime @xmath129 .
the reaction @xmath98 for an electron - deuteron plasma is increased by @xmath130 .
we find that these fluctuations have small , but not negligible amplitudes , compared to those at low frequencies .
they exist , however , in a much larger range of phase space , up to very high frequencies .
their overall effect is to increase the screening potential . because the reaction rates are exponentially dependent on the screening potential , this increase is especially strong for reactions with relatively large @xmath61 , such as @xmath62 and @xmath63 . when summing the fluctuations only up to the plasma eigenfrequency @xmath116 the salpeter enhancement factor is obtained .
we show here that the fluctuating potential has a large contribution from highly damped modes with frequencies @xmath131 .
there has been some controversy about the existence of a screening factor , which is different from the salpeter factor , due to dynamic screening @xcite .
the effect described here is due to quasi - fluctuations that are heavily damped and that always exist in a plasma .
they alter the salpeter screening factor , although not necessarily in the same way as does dynamic screening .
this work shows that by taking into account all the modes , including those which are heavily damped , with frequencies higher than the plasma frequency , the screening factor of the nuclear reaction rates is altered and is larger than the salpeter factor .
the consequences of this work are left for future studies .
60 d. d. clayton , _ principles of stellar evolution and nucleosynthesis _
( new york : mcgraw - hill , 1968 ) .
e. e. salpeter , aust .
j. phys . ,
* 7 * , 373 ( 1954 ) .
a. g. sitenko , _ electromagnetic fluctuations in plasma _
( academic press , ny , 1967 ) .
a. i. akhiezer , i. a. akhiezer , r. v. plovin , a. g. sitenko , and k. n. stepanov , _ plasma electrodynamics _ vol .
2 ( pergamon press , oxford , 1975 ) .
m. opher and r. opher , , 2628 ( 1997 ) .
m. opher and r. opher , , 3296 ( 1997 ) .
m. opher and r. opher , , 4835 ( 1999 ) .
p. c. clemmow and j. p. dougherty , _ electrodynamics of particles and plasmas _ , ( addison - wesley , redwood city , ca , 1990 ) .
c. carraro , a. schfer , and s. e. koonin , , * 331 * , 565 ( 1988 ) ; a. v. gruzinov , , * 496 * , 503 ( 1998 ) ; m. opher and r. opher , ( in press ) ( 2000 ) . | the _ fluctuation - dissipation theorem _ is used to evaluate the screening factor of nuclear reactions due to the electromagnetic fluctuations in a plasma .
we show that the commonly used saltpeter factor is obtained if only fluctuations near the plasma eigenfrequency are assumed to be important ( @xmath0 ( @xmath1 ) ) . by taking into account all the fluctuations , the highly damped ones , with @xmath2 as well as those with @xmath3 we find that nuclear reaction rates are higher than those obtained using the saltpeter factor , for many interesting plasmas . |
in the classical theory of hypergeometric functions , a prominent role is played by the euler integral formula @xmath3 which yields an analytic continuation of the gauss hypergeometric series @xmath4 from the unit disk @xmath5 to the larger domain @xmath6 . however , this euler integral is not symmetric in @xmath7 and @xmath8 , even though the function @xmath4 enjoys such symmetry .
following erdlyi @xcite , one can introduce another variable of integration and obtain the symmetric formula [ eq : sym 2f1 ] _
2f_1(s ; t ; u ) = g(s , t ) _
0 ^ 1 _ 0 ^ 1 x^s_1 - 1y^s_2 - 1(1-x)^t - s_1 - 1(1-y)^t - s_2 - 1(1-uxy)^-tdxdy , where @xmath9 after making the substitutions @xmath10 , @xmath11 , and @xmath12 , one finds that the double integral in takes the simple form [ eq : double int intro ] _ 0^ _ 0^ , which restricted to @xmath13 is a twofold mellin transform of @xmath0 , where @xmath14 . in this paper
we introduce a generalization of the mellin transform of a rational function @xmath0 , which we call an euler mellin integral .
the general form of an euler mellin integral can be found in below .
euler mellin integrals are closely related to @xmath2-hypergeometric euler type integrals ( see @xcite ) .
the most notable difference between these previously studied functions and the euler
mellin integrals we introduce here is the domain of integration .
we integrate over explicit , simply connected , but non - compact sets , whereas previous authors used compact yet rather elusive cycles .
we show that the simple connectivity of our domain of integration allows us to handle the multivaluedness of the integrand ; however , to achieve convergence , the non - compactness initially restricts the values of the parameters @xmath15 , see theorem [ thm : expgro ] . in theorem
[ thm : mel ] , we remove the restrictions on @xmath15 through an explicit meromorphic continuation of an euler mellin integral .
this yields a meromorphic function whose singular locus is contained in certain families of hyperplanes . taking these into account , we obtain a function which is entire in the parameters @xmath15 . a further generalization of the results of [ sec : em integrals ] is achieved by considering the coamoeba of the polynomial @xmath1 , as defined in . in practice
, we rotate the domain of integration of the euler
mellin integral to @xmath16 , for appropriate choices of @xmath17 ( see [ sec : coamoebas ] ) . in the previous example
, this means replacing the integral by @xmath18 the @xmath2-hypergeometric approach considers a polynomial @xmath1 with general coefficients on a fixed set of monomials , which are identified with a matrix @xmath2 .
we show in theorem [ thm : gkz ] that for a generic choice of coefficients of @xmath1 , the corresponding euler mellin integral with parameters @xmath15 satisfies an @xmath2-hypergeometric system of differential equations with parameter @xmath19 . in particular , the meromorphic continuations of euler mellin integrals obtained through [ sec : em integrals][sec : coamoebas ] provide a family of @xmath2-hypergeometric functions that are entire in the parameter @xmath20 ( see definition [ def : gkz ] ) .
a key open problem in the study of an @xmath2-hypergeometric system is to explain how its solution space varies with the parameter @xmath20 .
after meromorphic continuation , euler mellin integrals provide a new tool to approach this task . as illustrated in
[ sec : gkz ] , the explicit nature of the domain of integration in an euler mellin integral makes it possible to perform computations in examples where the variation of @xmath2-hypergeometric solutions is not yet fully understood .
we also show in proposition [ prop : indep ] that under certain conditions on @xmath2 , euler mellin integrals provide a full basis of solutions for generic @xmath20 .
an example in [ sec:0134 ] provides evidence that euler
mellin integrals explain the appearance of special solutions at rank - jumping parameters .
barnes integrals are another class of @xmath2-hypergeometric integrals previously considered in the literature @xcite ( see definition [ def : mb ] ) ; in particular , such integrals are used by beukers to compute elements in the local monodromy group of an @xmath2-hypergeometric system . in
[ sec : mellinbarnes ] , we compare these to euler mellin integrals .
a consequence of theorem [ thm : em - mb ] is that there are at least as many linearly independent euler
mellin integrals as mellin
barnes integrals ; this provides additional examples in which euler mellin integrals are linearly independent at generic @xmath20 ( see corollary [ cor : lopsidedintegralindependence ] ) . in [ sec : em integrals ] , we introduce euler mellin integrals , show their convergence , and perform their meromorphic continuation , which is our main result . in [ sec : coamoebas ] , we employ coamoebas to extend the results of the previous section to include more general domains of integration .
mellin integrals are shown to be @xmath2-hypergeometric functions in [ sec : gkz ] , providing a basis of solutions to @xmath2-hypergeometric systems in certain cases . in [ sec : mellinbarnes ] , we relate euler mellin integrals to mellin
barnes integrals , obtaining further insight into the linear independence of both sets of integrals . finally , [ sec:0134 ] contains an example of the behavior of euler mellin integrals at a rank - jumping parameter of an @xmath2-hypergeometric system .
this work was done in connection with the program algebraic geometry with a view toward applications " at institut mittag - leffler , and the authors are thankful to the institute for its hospitality .
we are also thankful to alicia dickenstein and laura felicia matusevich for helpful conversations .
this section contains our main result , theorem [ thm : mel ] , which provides an explicit presentation of a meromorphic continuation of the euler mellin integral of a polynomial @xmath1 of several variables @xmath21 .
given a polynomial @xmath22 , the _ euler mellin integral _ is a natural generalization of the mellin transform of a rational function @xmath0 of several variables and is given by [ eq : eulermellin ] m_f(s , t ) : = _ _ + ^n = _ ^n dx_1 dx_n , where @xmath23 denotes the positive orthant in @xmath24 . here
we employ multi - index notation for variables @xmath25 and polynomials @xmath26 ; that is , for @xmath27 and @xmath28 , we write @xmath29 and @xmath30 .
whenever there is no risk of confusion , we use the notation @xmath31 . in order for such an integral to converge , restrictions must a priori be placed on both the exponent vector @xmath15 and the polynomial @xmath1 ; it is not enough to demand only that each @xmath32 is nonvanishing on @xmath33 . we next provide such a domain of convergence for the euler
mellin integral , generalizing ( * ? ? ?
* thm . 1 ) .
[ def : nonvng ] if @xmath34 is a face of the newton polytope @xmath35 of @xmath1 , then the _ truncated polynomial with support @xmath34 _ is given by @xmath36 the polynomial @xmath1 is said to be _ completely nonvanishing _ on a set @xmath37 if for each face @xmath34 of @xmath35 , the truncated polynomial @xmath38 has no zeros on @xmath37 . in particular , the polynomial @xmath1 itself does not vanish on @xmath37 . for a vector @xmath39 , we denote by @xmath40 the weighted minkowski sum @xmath41 of the newton polytopes of the @xmath32 with respect to @xmath42 .
note that with this notation , the newton polytope of @xmath1 satisfies @xmath43 .
[ thm : expgro ] if each of the polynomials @xmath44 are completely nonvanishing on the positive orthant @xmath33 ( as in definition [ def : nonvng ] ) , then the euler mellin integral @xmath45 of converges and defines an analytic function in the tube domain @xmath46 it suffices to prove that for any @xmath15 with all @xmath47 and @xmath48 , there exist positive constants @xmath49 and @xmath50 such that @xmath51 in fact , it is enough to show that this inequality holds outside of some ball @xmath52 in @xmath24 . since @xmath53
, we can expand it as a sum @xmath54 of @xmath55 vectors such that @xmath56 .
it is shown in the proof of ( * ? ? ?
1 ) that for each @xmath57 there are positive constants @xmath58 and @xmath59 such that for @xmath60 outside of some ball @xmath61 , @xmath62 note that it is essential in ( * ? ? ?
1 ) that @xmath32 is completely nonvanishing on the positive orthant . thus for @xmath60 outside of @xmath63
, we have [ eq : bound ] |f(e^x)^t| e^-,x = _ i=1^m(|f_i(e^x)| e^-,x)^_i _ i=1^mc_i^_ie^_i k_i|x| = ce^k|x| , where @xmath64 and @xmath65 are the desired positive constants .
[ ex:2f1classical ] by a classical integral representation of the gauss hypergeometric function @xmath66 , [ eq : classical 2f1 ] _ 0^ = _ 2 f_1(t_2,t_1+t_2-s;t_1+t_2;1-c ) for @xmath67 and @xmath68 , where @xmath69 denotes the principal branch of the argument mapping .
note that @xmath68 is equivalent to @xmath70 being completely nonvanishing on @xmath71 . since @xmath72 $ ] , the condition that @xmath73 is the same as @xmath74 .
we also note that the right hand side of is analytic in this domain .
further , since @xmath75 and @xmath76 , the convergence domain given in theorem [ thm : expgro ] is not optimal ; however , being full - dimensional , it is large enough for our goal of meromorphic continuation .
as the right hand side of is a meromorphic function in @xmath77 and @xmath78 , it provides a meromorphic extension of the corresponding euler
mellin integral . on this right side
, we have the regularized @xmath4 as one factor , thus the polar locus of the meromorphic extension is contained in two families of hyperplanes given by the polar loci of the gamma functions .
our main result shows that this kind of meromorphic continuation is possible for all euler mellin integrals . to obtain the strongest form of this result
, we choose a specific presentation for @xmath40 . to begin , each newton polytope @xmath79 can be written uniquely as the intersection of a finite number of halfspaces [ eq : newton fi ] _
f_i=_j=1^n_i \{^n ^i_j,^i_j } , where the @xmath80 are primitive vectors .
fixing an order , let @xmath81 be equal to the set @xmath82 , where we assume that @xmath83 for all @xmath84 .
we now extend the definitions of @xmath85 from to each @xmath86 ; namely , for each @xmath50 , let @xmath87 with @xmath88 and set @xmath89 .
it now follows from the definition of the @xmath90 that [ eq : newtonpol2 ] _ f = _
k=1^n \{^n _ k,_k , } and @xmath91 . we are now prepared to state our main result , which provides a meromorphic continuation of , generalizing ( * ? ?
* thm . 2 ) .
in [ sec : coamoebas ] , we obtain a stronger form of the result by relaxing the condition that the @xmath32 be completely nonvanishing on @xmath33 .
[ thm : mel ] if the polynomials @xmath44 are completely nonvanishing on the positive orthant @xmath33 ( as in definition [ def : nonvng ] ) and the newton polytope @xmath92 is of full dimension @xmath93 , then the euler
mellin integral @xmath94 admits a meromorphic continuation of the form @xmath95 where @xmath96 is an entire function and @xmath97 are given by . by theorem [ thm : expgro ] , the original euler mellin integral @xmath45 of converges on @xmath98 which is a domain since @xmath35 is of full dimension .
our goal is to expand the convergence domain of the integral , at the cost of multiplication by terms corresponding to the poles of the gamma functions appearing in .
we do this iteratively , integrating by parts in the direction of a vector @xmath86 at each step .
this expands the domain of convergence in the opposite direction of @xmath86 by a distance @xmath99 , which we determine explicitly . to begin , we set notation for the first iteration in one direction . fix @xmath50 between @xmath100 and @xmath101 , and let @xmath34 be the face of @xmath79 corresponding to @xmath86 and @xmath90
. for @xmath102 , set @xmath103 since @xmath104 , it follows that @xmath105 . in particular , since there is a decomposition @xmath106 with @xmath107 , we see that @xmath108 if and only if @xmath109 for all @xmath110 .
fixing @xmath110 , the truncated polynomial @xmath111 has the homogeneity @xmath112 , where @xmath113 is any nonzero complex number and @xmath114 .
hence the coefficients of the scaled polynomial @xmath115 are independent of @xmath50 and the @xmath113 . in particular , we have that the newton polytope of @xmath116 is disjoint from @xmath34 .
this fact allows us to extend the domain of convergence of over the hyperplane defined by @xmath117 as follows .
since @xmath94 is independent of @xmath113 , we have @xmath118.\ ] ] thus differentiating with respect to @xmath113 and setting @xmath119 yields the identity @xmath120 where @xmath121 is the polynomial @xmath122 note that @xmath123 is contained in @xmath124 ; moreover , since @xmath34 is the face of @xmath35 corresponding to @xmath86 and @xmath125 is disjoint from @xmath126 , we see that @xmath123 is disjoint from @xmath34 .
in other words , @xmath127 for each @xmath128 . rewriting as the sum [ eq : int rewrite ] m_f(s , t ) = _
( g_k)__+^n , for some linear polynomials @xmath129 , noting that each term of is a translation of the original euler mellin integral . by theorem [ thm :
expgro ] , the term corresponding to @xmath130 converges on the domain given by @xmath131 and @xmath132 where the latter is equivalent to @xmath133 the sum converges on the intersection of these domains , which is given by @xmath134 where @xmath135 . since @xmath99 is by definition strictly greater than 0 , has a strictly larger domain of convergence than ; we say that it has been extended by the distance " @xmath99 in the direction determined by @xmath86 . before iterating this procedure , we set some notation .
let @xmath136 be the semigroup generated by the integers @xmath137 .
let @xmath138 be an ordered @xmath139-tuple with @xmath140 for each @xmath110 .
we sometimes write @xmath20 as an exponent of @xmath141 , viewing @xmath142 .
similarly , set @xmath143 .
now after @xmath139 iterations , let @xmath144 denote the direction of the extension in the @xmath110th iteration .
let @xmath145 be the sum of the distances of the first @xmath146 components of @xmath20 in the direction @xmath144 .
then there is a rational function of the type [ eq : first hs ] l_(s , t ) = _ i=1^q , where @xmath147 is an ordered @xmath139-tuple of linear polynomials such that @xmath148 can be expressed as a finite sum of translations of the original euler mellin integral : [ eq : induction ] m_f(s , t ) = _ l_(s , t ) _ _ + ^n .
fixing @xmath50 , we next expand the domain of convergence of in the direction determined by @xmath86 .
this is achieved through simultaneous expansion of the domains of convergence of all terms , arguing as above .
this yields the expression @xmath149 where @xmath150 , @xmath151 , and the resulting rational function @xmath152 is given by @xmath153 since the convergence domain of each term in is extended by the distance @xmath99 in the direction determined by @xmath86 , the convergence domain of the sum is similarly extended .
in addition , since @xmath154 , we have that @xmath155 ; therefore , the products @xmath156 will never repeat factors in their denominators .
as is in the same form as , we may iterate this procedure to extend the domain of convergence .
finally , note that after @xmath139 iterations that have extended the domain of convergence of @xmath94 in the direction determined by @xmath157 for @xmath158 of the @xmath139 steps , we obtain a meromorphic function on the tube domain given by @xmath159 such that @xmath160 and @xmath161 continuing , @xmath94 can be extended to a meromorphic function on @xmath162 as in . we note that because the denominator of the products of the rational functions @xmath163 never has repeated terms , all poles of the extended euler mellin integral are simple .
therefore by the removable singularities theorem , @xmath96 in is an entire function , as desired .
the entire function @xmath96 is of great interest to the study of @xmath2-hypergeometric functions .
the gamma functions appearing in might have introduced some unnecessary zeros in the meromorphic continuation of the euler
mellin integral , which hinder @xmath2-hypergeometric applications .
[ rm : polesemigroups ] in the proof of theorem [ thm : mel ] , we see that the linear form @xmath164 appears in the denominator of some rational function @xmath165 if and only if @xmath166 . hence if @xmath167 , then our meromorphic continuation has introduced unnecessary zeros into the entire function @xmath96 .
[ rm : onepolyzeros ] if @xmath168 , then @xmath169 for some constant @xmath170 , where @xmath171 is as in .
therefore each @xmath165 is divisible by @xmath172 , which can thus be factored outside of the sum .
in particular , @xmath173 is an entire function .
we conclude this section with examples to illustrate theorem [ thm : mel ] and our recent remarks .
consider the case of @xmath174 linear functions of one variable , [ eq : ex mf ] m_f(s , t ) = _ 0^ .
note that we have reindexed @xmath78 for this example .
when @xmath175 , is the beta function . here
@xmath176 , or with the notation of remark [ rm : onepolyzeros ] , @xmath177 . when @xmath178 , we showed in example [ ex:2f1classical ] that @xmath179 this equality is obtained by the change of variables @xmath180 and application of the generalized binomial theorem . by similar calculations for @xmath181 , @xmath182 where @xmath183 denotes the first appell series . for arbitrary @xmath55 and @xmath184 , @xmath185 where @xmath186 , @xmath187 , and @xmath188 .
[ ex : mellinlinearpoly ] finally , consider the case of one linear function of @xmath93 variables , @xmath189 we claim that @xmath190 and hence @xmath176 .
this is clear when @xmath191 because we again have the beta function . for @xmath192
one can argue by induction , making the change of variables given by @xmath193 and @xmath194 for @xmath195 . to generalize this example to an arbitrary simplex , consider the euler
mellin integral @xmath196 where the exponent vectors @xmath197 are the columns of an invertible matrix @xmath198 . by the change of variables
@xmath199 we find that @xmath200
for theorems [ thm : expgro ] and [ thm : mel ] to hold , each @xmath201 must be completely nonvanishing on the positive orthant .
this is a strong restriction that many polynomials will not fulfill .
however , the goal of this section is to modify this hypothesis by considering the coamoeba of @xmath202 .
the amoeba @xmath203 and the coamoeba @xmath204 of a polynomial @xmath1 are defined to be the images of the zero set @xmath205 under the real and imaginary parts of the coordinate - wise complex logarithm mapping , @xmath206 and @xmath207 , respectively .
more precisely , if @xmath208 and @xmath209 , then the amoeba and coamoeba of @xmath1 are , respectively , @xmath210 the amoeba @xmath203 lies in @xmath24 ; however , since the argument mapping is multivalued , the coamoeba @xmath204 can be viewed either in the @xmath93-dimensional torus @xmath211 or as a multiply periodic subset of @xmath24 .
amoebas were introduced by gelfand , kapranov , and zelevinsky in @xcite , while the term coamoeba was first used by the third author in 2004 at a conference at johns hopkins university .
the following results provide the connection between coamoebas and euler mellin integrals .
[ prop : coamoeba ] for @xmath212 , a polynomial @xmath202 is completely nonvanishing on the set @xmath213 if and only if @xmath214 .
the claim is equivalent to the statement @xmath215 where @xmath38 is the truncated polynomial with support @xmath34 .
this has been proven by johansson @xcite and independently by nisse and sottile @xcite .
thus for polynomials @xmath216 such that the closure of the coamoeba of @xmath217 is a proper subset of @xmath218 , there is a @xmath214 for which the euler mellin integral with respect to @xmath17 is well - defined : [ eq : theta eulermellin ] m_f^(s , t ) : = _ ^-1 ( ) .
note that after fixing the matrix @xmath2 , and hence the set of monomials of the polynomials @xmath32 , any choice of coefficients for the @xmath32 with positive real part ensures that @xmath219 .
in particular , this means there is always a choice of coefficients for the @xmath32 so that @xmath220 is a proper subset of @xmath218 . for another discussion on the existence of components of the complement of @xmath220 ,
see remark [ rem : lopsided ] . as differs from our earlier definition of the euler
mellin integral in only by a change of variables , it is immediate that @xmath17-analogues of theorems [ thm : expgro ] and [ thm : mel ] hold . in addition , a slight perturbation of @xmath17 does not impact the value of .
the euler
mellin integral @xmath221 of is a locally constant function in @xmath17 .
thus it depends only on the choice of connected component @xmath222 of the complement of @xmath220 , and we thus write @xmath223 .
accordingly , there is a meromorphic continuation of @xmath224 , denoted @xmath225 .
first consider the case @xmath191 , and suppose that @xmath226 and @xmath227 lie in the same connected component of the complement of @xmath220 ; in fact , assume that the interval @xmath228\subseteq \tt^n\setminus\overline{\ca'_f}$ ] .
in other words , @xmath202 has no zeros with arguments in this interval , and hence @xmath229 is analytic in the corresponding domain .
connecting the two rays @xmath230 and @xmath231 with the circle section of radius @xmath232 yields a closed curve , and the integral of @xmath229 over this ( oriented ) curve is zero by residue calculus . by the proof of theorem [ thm : expgro ]
, the integral over the circle section tends to 0 as @xmath233 , so the two euler mellin integrals @xmath234 and @xmath235 are equal . in arbitrary dimension ,
we obtain the desired equality by considering one variable at a time while the remaining variables are fixed . [ ex:2f1 first ] revisiting the polynomial @xmath236 from the introduction , we see that if we choose @xmath237 , then @xmath238 where @xmath222 is the component of the complement of @xmath220 containing @xmath17 . in accordance with remark [ rm : onepolyzeros ] , we may ignore one of the factors @xmath239 in the denominator , while @xmath240 is the regularized gauss hypergeometric function .
we now fix a connected component @xmath222 of the complement of @xmath220 and study the entire function @xmath241 from . in particular , we consider its dependence on the coefficients @xmath242 of the polynomials @xmath32 , where @xmath243 and @xmath244 ( so @xmath245 ) . in order to emphasize this dependence , we write @xmath246 rather than @xmath96 .
generalizing @xcite , we show that @xmath247 is an @xmath2-hypergeometric function in the sense of gelfand , kapranov , and zelevinsky .
more precisely , @xmath248 satisfies the @xmath2-hypergeometric system of partial differential equations , where the exponents @xmath249 of the @xmath32 provide a matrix @xmath2 via the cayley trick , [ eq : a ] a = ^(m+n)(r_1++r_m ) , and the desired homogeneity parameter is @xmath19 .
set @xmath250 .
we first recall the definition of an @xmath2-hypergeometric system . for a vector @xmath251 , denote by @xmath252 and @xmath253 the unique vectors in @xmath254 with disjoint support such that @xmath255 .
[ def : gkz ] let @xmath256 be a matrix . define the differential operators @xmath257 and @xmath258 to be @xmath259 the _ @xmath2-hypergeometric system _ @xmath260 at @xmath261 is given by @xmath262 a local multivalued analytic function @xmath263 that solves this system is called an _ @xmath2-hypergeometric function _ with homogeneity parameter @xmath20 . the ideal @xmath264 cuts out an affine variety @xmath265 , which has an action of an algebraic torus , @xmath266 . to understand the role of the euler operators @xmath267 ,
note that a germ of a analytic function at a nonsingular point @xmath268 that is annihilated by @xmath269 is homogeneous , in the usual sense , of degree @xmath270 .
in general , the euler operators in @xmath271 force solutions to have weighted homogeneities . from this point of view , it becomes natural to fix @xmath2 and view @xmath20 as a parameter of @xmath271 .
a key problem in the study of an @xmath2-hypergeometric system is to describe the variation with the parameter @xmath20 of its solution space of germs of analytic functions at a nonsingular point . to begin this study
, one must first find solutions of @xmath260 , that vary nicely with @xmath20 .
saito , sturmfels , and takayama presented an algorithm to compute such a basis for arbitrary @xmath20 , called _
canonical series solutions _ @xcite
. however , because this algorithm uses grbner degeneration , the solutions it produces are not well - suited to the variation of @xmath20 . for nonresonant @xmath20 ,
gelfand , kapranov , and zelevinsky computed a basis of euler - type integral solutions @xcite .
these are also unsuitable for understanding parametric behavior , as their domains of integration are not explicit and the family integrals formed there are not guaranteed to be linearly enough at nongeneric @xmath20 .
in contrast , since our meromorphic continuations of euler mellin integrals are entire in @xmath20 , they provide a new tool for describing the parametric variation of @xmath2-hypergeometric solutions . to this end
, we now consider the behavior of the entire function @xmath272 , as described in theorem [ thm : mel ] , when @xmath49 is viewed as a variable .
let @xmath273 denote the singular locus of all @xmath2-hypergeometric functions , which is the hypersurface defined by the principal @xmath2-determinant ( also known as the full @xmath2-discriminant ) @xcite .
[ thm : gkz ] let @xmath274 and let @xmath222 be a connected component of @xmath275 , where @xmath1 is the polynomial @xmath243 and @xmath244 .
then the analytic germ @xmath276 , for any @xmath277 , has a ( multivalued ) analytic continuation to @xmath278 that is everywhere @xmath2-hypergeometric ( in the variables @xmath49 ) with homogeneity parameter @xmath19 .
let us first consider the case @xmath279 and @xmath280 , where we have the integral representation [ eq : phi hyp ] _
f^(s , t , c)= _ ^-1 ( ) .
fix a representative @xmath277 .
as @xmath17 is disjoint from @xmath220 for polynomials @xmath1 with coefficients @xmath49 near the original ones , say in a small ball @xmath281 , the integral in does indeed define an analytic germ @xmath282 . by theorem [ thm : mel ]
, @xmath247 can be extended to an entire function with respect to the variables @xmath77 and @xmath78 . in other words
, @xmath247 has been analytically extended to the infinite cylinder @xmath283 . to see that @xmath247 is an @xmath2-hypergeometric function with homogeneity parameter @xmath20 as given , we fix @xmath77 and @xmath78 under the above condition , noting that the product of gamma functions in @xmath247 is simply a nonzero constant .
thus it is enough to show that the integral itself is @xmath2-hypergeometric at @xmath20 .
this is accomplished through the argument of ( * ? ? ?
5.4.2 ) , which applies since differentiation and integration may be interchanged because euler mellin integrals are uniformly convergent by the bound in .
see also ( * ? ? ?
* remark 2.8(b ) ) .
having established that @xmath247 is an @xmath2-hypergeometric function in the product domain given by @xmath284 in @xmath285 , it follows from the uniqueness of analytic continuation that its extension to the cylinder @xmath286 will remain @xmath2-hypergeometric .
now for each fixed @xmath15 , there is a ( typically multivalued ) analytic continuation of @xmath287 from @xmath281 to all of @xmath288 .
as these continuations still depend analytically on @xmath77 and @xmath78 , we have now achieved the desired analytic continuation to the full product domain @xmath289 .
the uniqueness of analytic continuation again guarantees that @xmath247 will everywhere satisfy the @xmath2-hypergeometric system with the homogeneity parameter @xmath20 , as desired . in @xcite , for generic ( nonresonant ) parameters @xmath20 , a basis of solutions of @xmath260
are given via generalized euler integrals .
the set of nongeneric , or resonant , parameters is given by the union over all faces @xmath34 of the polyhedral cone @xmath290 over the columns of @xmath2 of @xmath291 .
the dimension of the solution space in this case is given by @xmath292 , which is @xmath293 times the euclidean volume of the convex hull of @xmath2 and the origin .
under certain conditions on @xmath2 , a similar result holds for our meromorphic continuations of euler mellin integrals .
we consider here the case that @xmath1 is a univariate polynomial , followed by another example for which linear independence of these integrals can be seen directly by analyzing coamoebas .
such linear independence will also be addressed later in corollary [ cor : lopsidedintegralindependence ] .
[ prop : indep ] if @xmath20 is nonresonant and @xmath191 , then the extended euler mellin integrals @xmath276 , where @xmath222 ranges over the components of @xmath294 , provide a basis of solutions to @xmath260 . for a generic choice of coefficients @xmath49 , @xmath1 will have distinct roots @xmath295 with distinct arguments @xmath296 .
that is , @xmath297 will have @xmath292-many components in its complement in @xmath298 . for convenience , we use the convention that @xmath299 .
now fix @xmath300 so that the circles @xmath301 , viewed as 1-chains oriented counterclockwise , have disjoint supports for @xmath302 .
by @xcite , the integrals @xmath303 are linearly independent , forming a basis for the solution space of @xmath260 . by the convergence of @xmath304 from theorems [ thm : expgro ] and [ thm : mel ]
, we see that as non - compact chains , @xmath305 and @xmath306 are homologous , providing the second statement . for a nongeneric choice of coefficients @xmath49 for @xmath1 , which still lie away from @xmath307
, a similar argument implies linear independence of the euler
mellin integrals given by the distinct components of the complement of @xmath220 . [ ex : gauss with @xmath78 ] in example [ ex:2f1 first ] , it was shown that if @xmath309 and @xmath17 is near @xmath310 , then @xmath311 by theorem [ thm : gkz ] , this is a solution of @xmath260 for @xmath312 \quad\text{and}\quad \beta = -(t , s).\ ] ] now consider points @xmath49 of the form @xmath313 , where @xmath314 is near 1 .
define the polynomials @xmath315 as shown in figure [ fig : gauss coamoebas ] , the coamoeba for @xmath316 has two connected components : one containing @xmath317 and another containing @xmath318 .
these yield two solutions to @xmath260 at @xmath319 by theorem [ thm : gkz ] , namely , @xmath320 and @xmath321 .
in addition , @xmath322 and @xmath323 for all @xmath324 , so we let @xmath325 and @xmath326 denote the entire functions corresponding to @xmath327 and @xmath328 , respectively .
ccc ( 40,40 ) ( 20,20 ) ( 20.5,21)@xmath317 ( 0,0)(1,0)40 ( 0,0)(0,1)40 ( 40,0)(0,1)40 ( 0,40)(1,0)40 & ( 40,40 ) ( 1,0 ) ( 1,1 ) ( 3,1 )
( 3,0) cycle ; ( 0,1 ) ( 1,1 ) ( 1,3 ) ( 0,3) cycle ; ( 3,1 ) ( 3,3 ) ( 4,3 ) ( 4,1) cycle ; ( 1,3 ) ( 3,3 ) ( 3,4 ) ( 1,4) cycle ; ( 20,20 ) ( 20.5,21)@xmath317 ( 39,39 ) ( 30.25,34.75)@xmath318 & ( 40,40 ) ( 39,39 ) ( 30.25,34.75)@xmath318 ( 0,20)(1,0)40 ( 20,0)(0,1)40 ( 80,40 ) ( 2.5,2.3 ) arc ( 180:5:20 mm ) ; ( 2.5,2.3 ) arc ( 180:355:20 mm ) ; ( 25,23 ) ( 20,20 ) ( 17,17)@xmath329 ( 40,20 ) ( 37,17)@xmath100 ( 60,20 ) ( 57,17)@xmath330 ( 65,23 ) ( 67,24)@xmath331 ( 27,38)@xmath332 ( 27,08)@xmath333 let @xmath334 be the loop in the coefficient space given by first following the coefficients in the reverse of @xmath327 and then those in @xmath328 . since @xmath316 and @xmath335 , we can explicitly perform an analytic continuation of @xmath320 along the loop @xmath334 , see figure [ fig : gauss continuation ] .
therefore if the monodromy of @xmath336 is irreducible , then @xmath320 and @xmath321 form a basis for the solution space of @xmath336 of analytic germs at @xmath313 .
as the monodromy irreducibility of @xmath336 is equivalent to the nonresonance of @xmath20 @xcite , the conclusion of proposition [ prop : indep ] also holds in this case .
an alternative approach for constructing integral representations of @xmath2-hypergeometric functions is through _ mellin barnes integrals _ @xcite .
the main result of this section is theorem [ thm : em - mb ] , which identifies the set of mellin
barnes integrals of a system @xmath260 with a certain subset of the set of euler mellin integrals .
this in turn implies corollary [ cor : lopsidedintegralindependence ] , which asserts the linear independence of certain collections of meromorphic continuations of euler mellin integrals at generic parameters @xmath20 .
the _ gale dual _ of @xmath2 is an integer @xmath337-matrix @xmath338 with relatively prime maximal minors such that @xmath339 .
typically we will not require that the condition on maximal minors holds ; in this case , the matrix @xmath338 is called a _ dual matrix _ of @xmath2 . given a dual matrix @xmath338 of @xmath2 , let @xmath340 $ ] denote the sublattice of @xmath341 generated by the rows of @xmath338 . also , consider the zonotope @xmath342 where @xmath343 denotes the @xmath110th row of the matrix @xmath338 .
for connections between camoebas and gale duals , see @xcite .
[ def : mb ] fix a gale dual @xmath338 of @xmath2 , and let @xmath344 be such that @xmath345 .
then for @xmath346 , the _ mellin barnes integral _ has the form @xmath347 where @xmath348 .
we say that @xmath349 is _ totally nonresonant _ for @xmath2 if the shifted lattice @xmath350 has empty intersection with any hyperplane spanned by any @xmath351 linearly independent columns of @xmath2 .
given @xmath352 and @xmath353 , let @xmath354 the following mellin
barnes result summarizes corollary 4.2 , theorem 3.1 , and proposition 4.3 of @xcite .
[ thm : beukers ] fix @xmath352 and @xmath353 . 1 . if @xmath355 , then the integral @xmath356 converges absolutely .
2 . if @xmath355 and @xmath357 for each @xmath110 , then @xmath356 satisfies the @xmath2-hypergeometric system @xmath260 as a germ of an analytic function at @xmath49 .
[ item3 ] suppose that @xmath20 is totally nonresonant for @xmath2 and @xmath358 are such that the @xmath359-tuples @xmath360 are distinct and contained in @xmath361 .
then as germs at @xmath49 , the mellin
barnes integrals @xmath362 , , @xmath363 are linearly independent .
it follows from results in @xcite that each @xmath359-tuple @xmath364 is contained in the translated lattice @xmath365 $ ] .
thus for each set of points @xmath366)$ ] , there is an associated set of linearly independent solutions to @xmath260 , given by the corresponding mellin barnes integrals . to relate mellin
barnes integrals to euler mellin integrals , we first recall the definition of the lopsided coamoeba of a polynomial @xmath202 .
the polynomial @xmath367 , with coefficients @xmath49 also viewed as variables , has coamoeba @xmath368 contained in @xmath369 . given @xmath370 with fixed coefficients @xmath49 ,
the _ lopsided coamoeba _ of @xmath1 , denoted by @xmath371 , is the intersection of @xmath368 with the sub-@xmath218-torus of @xmath369 obtained by fixing @xmath372 as prescribed by @xmath1 .
the lopsided coamoeba @xmath371 is viewed as a subset of @xmath218 .
there are several ways to define the lopsided coamoeba ; for a more exhaustive treatment , we refer the reader to @xcite .
the name lopsided coamoeba " can be misleading ; the lopsided coamoeba is not a actually a coamoeba itself , but it can be viewed as a crude approximation of the coamoeba .
see figure [ fig : lopsided ] for a comparison . there is a natural inclusion @xmath373 , and in particular , each component of @xmath374 is contained in a component of @xmath375 .
the induced map on components of the closures of these coamoebas is injective but in general , is not surjective .
.,title="fig:",width=151 ] .,title="fig:",width=151 ] [ rem : lopsided ] the main result of @xcite states that the set of connected components of the complement of @xmath376 is equipped with an _ order map_.
fix @xmath377 , and let @xmath378 denote the principal branch of the argument mapping
. denote by @xmath379 $ ] the row vector of length @xmath232 with entries as shown , indexed by @xmath380 in lexicographic order .
this indexing corresponds to the labeling of the columns of @xmath2 , as shown in .
define the function @xmath381 to be @xmath382b.\ ] ] the properties of this function is given in detail in ( * ? ? ?
4.1 and prop .
4.4 - 5 ) and in ( * ? ? ?
2.3.10 ) . there
it is shown that @xmath383 is well - defined on @xmath218 ; and furthermore , it is constant on each component of @xmath374 .
in fact , the map from the set of connected components of @xmath374 to @xmath366)$ ] given by @xmath384 is surjective and independent of choice of @xmath385 . if the maximal minors of @xmath2 are relatively prime , then this map is also injective .
we are now prepared to approach the first result of this section , which states explicitly how the order map for the set of components of the complement of the closure of the lopsided coamoeba lifts to a bijection between the set of mellin
barnes integrals arising from the points in the translated lattice @xmath365 $ ] and the set of euler mellin integrals arising from the components of the complement of the lopsided coamoeba of the polynomial with coefficients @xmath49 . to show this
, we consider a dual matrix @xmath338 of @xmath2 with a special form .
[ conv : dual ] without loss of generality , we may assume that @xmath2 is of the form @xmath386,\ ] ] where @xmath387 is a nonsingular @xmath388-matrix .
let @xmath338 denote the dual matrix of the form @xmath389d,\ ] ] where @xmath390 is defined by the property that each column sum of @xmath338 is zero and @xmath391 is an integer diagonal matrix chosen so that @xmath338 becomes an integer matrix .
let @xmath392 ( respectively , @xmath393 ) denote the greatest common divisor of the maximal minors of the matrix @xmath2 ( respectively , @xmath338 ) .
[ lem : em - mb ] following convention [ conv : dual ] , there is an equality @xmath394 we may assume that @xmath395 .
following ( * ? ? ?
4.2 ) , this equality holds precisely when @xmath2 can be extended to a @xmath396 unimodular matrix @xmath397 \
\text{with inverse } \
\ , { \widetilde a}^{-1 } = \left [ \begin{array}{cc } * & \widetilde b \end{array } \right ] = \left [ \begin{array}{cc } * & \tilde b_0 \\ * & { \widetilde b}_1\\ * & { \widetilde b}_2 \end{array}\right].\ ] ] it follows that @xmath398 is a gale dual of @xmath2 , and by the schur complement formula , @xmath399 .
as @xmath400 for some affine transformation @xmath198 , the desired equality follows .
[ thm : em - mb ] for each @xmath401 , there is an equality @xmath402
. we will give the proof after rewriting the homogeneity parameter as @xmath403 .
in this case , the euler mellin integral is @xmath404 where @xmath405 and @xmath406 denotes the @xmath110th column of @xmath338 .
denote by @xmath407 the function given by the integral in , and let @xmath408 .
note that @xmath401 is equivalent to @xmath409 , which in turn is equivalent to the convergence of the integral @xmath410 however , this integral is precisely the mellin transform with respect to @xmath60 of @xmath407 . consequently , @xmath411 see example [ ex : mellinlinearpoly ]
. turning to the mellin
barnes integral , fix parameters @xmath412 and @xmath413 . assuming we have that @xmath414 for all @xmath415 , that @xmath416 , and that @xmath417 , it follows that @xmath418 for all @xmath50 .
furthermore , with @xmath419 denoting the @xmath420-st column of @xmath2 , @xmath421.\ ] ] for ease of display , set @xmath422
. then @xmath423 the bounds in the proof of theorem [ thm : expgro ] implies that we can apply the mellin inversion formula , which yields the equality @xmath424 combining this with lemma [ lem : em - mb ] provides the desired result . [ cor : lopsidedintegralindependence ] if @xmath20 is totally nonresonant for @xmath2 , then , when viewed as analytic germs at some @xmath274 , the @xmath2-hypergeometric functions given by meromorphic continuations of euler mellin integrals arising from the distinct components of @xmath425 are linearly independent .
let @xmath426 be representatives for the components of @xmath425 .
suppose that @xmath427 are constants providing a vanishing linear combination of @xmath428 , so that @xmath429 it then follows from theorem [ thm : beukers].[item3 ] that @xmath430 . in @xcite , it is noted that it is not always possible to construct a basis of the solution space of @xmath260 only considering mellin
barnes integrals of the form . by corollary [ cor : lopsidedintegralindependence ] ,
the same is true for constructing bases of solutions from meromorphic continuations of euler mellin integrals arising only from the set of components of @xmath374 .
however , @xmath375 has in general has more connected components than @xmath374 , and in many cases it is possible to use euler mellin integrals to construct a basis of solutions even though mellin
barnes integrals do not suffice .
consider the matrix @xmath431 $ ] .
the @xmath2-discriminant is @xmath432 whose coamoeba covers @xmath433 .
the maximal number of points in @xmath366)$ ] is five by @xcite , and hence there is no mellin
barnes basis of solutions for @xmath260 for any @xmath20 . however , for a generic choice of coefficients @xmath49 the coamoeba of @xmath434 has six components in its complement , and for nonresonant @xmath20 these provide a euler mellin basis of solutions for @xmath260 by proposition [ prop : indep ] .
we note that there exist matrices @xmath2 whose associated coamoeba has fewer than the desired @xmath292-many number of connected components in the complement of its closure for all choices of coefficients @xmath49 @xcite . on the other hand ,
if @xmath2 is a circuit , then when @xmath20 is totally nonresonant , there always exists a mellin
barnes , and hence an extended euler mellin , basis of integral representations for solutions of @xmath260 @xcite .
we conclude with an example first studied in @xcite , where it was first seen that some parameters @xmath20 admit a higher - dimensional solution space for @xmath260 than expected .
we illustrate how meromorphic continuations of euler mellin integrals capture these extra solutions at non - generic parameters @xmath20 , offering a new tool to understand how these special functions arise .
consider the system @xmath260 given by @xmath435 $ ] and the unique parameter @xmath436 which yields a solution space whose dimension is one larger than expected .
for this @xmath2 , the euler
mellin integral is [ eq:0134 em int ] m_f^ ( s , t , c ) = _ ^-1 ( ) for the polynomial @xmath437 and @xmath277 for a fixed connected component @xmath222 of @xmath438 . in order to calculate the corresponding @xmath439 ,
we first expand five times in different directions , so that it converges for @xmath440 . upon expansion
, @xmath441 is equal to @xmath442 where all integrals are taken over @xmath213 and @xmath443 is the pochhammer symbol .
this shows that when @xmath440 , the entire function @xmath439 falls into the situation noted in remark [ rm : onepolyzeros ] , and we thus ignore the factor @xmath444 in . to be explicit , @xmath445\\\ h_2(z )
= & \ \frac{36 c_1 c_3 ^ 2 z^6}{(s+3 ) ( 4t - s+2)}+\frac{48 c_1 c_3 c_4 z^7}{(s+3)(4t - s+1)}+\frac{48 c_1c_3 c_4 z^7}{(s+4 ) ( 4t - s+1 ) } + \frac{64 c_1 c_4 ^ 2 z^8}{(s+4 ) ( 4 t - s ) } \\ & \quad + \frac{c_2 ^ 3 z^3}{(s+1 ) ( s+2)}+\frac{3c_2 ^ 2 c_3 z^5}{(s+1 ) ( s+2)}+\frac{4 c_2 ^ 2 c_4 z^6}{(s+1 ) ( s+2 ) } + \frac{27 c_2 c_3 ^ 2 z^7}{(s+3 )
( 4t - s+2 ) } \\ & \quad + \frac{36 c_2 c_3 c_4 z^8}{(s+3 ) ( 4t - s+1)}+\frac{36 c_2 c_3 c_4 z^8}{(s+4 ) ( 4t - s+1 ) } + \frac{48c_2 c_4 ^ 2 z^9}{(s+4 ) ( 4 t - s ) } \\ & \quad + \frac{9 c_3 ^ 3 z^9}{(s+3 ) ( 4t - s+2)},\displaybreak[0]\\\ h_3(z ) = & \ \frac{48 c_1 c_3 ^ 2 c_4 z^{10}}{(s+3 ) ( 4t - s+1 ) ( 4t - s+2)}+\frac{48 c_1 c_3 ^ 2 c_4 z^{10}}{(s+4 ) ( 4t - s+1 ) ( 4t - s+12)}\\ & \quad + \frac{64 c_1 c_3 c_4 ^ 2 z^{11}}{(s+4 ) ( 4t - s+1)^2}+\frac{36 c_2 c_3 ^ 2 c_4 z^{11}}{(s+3 ) ( 4t - s+1 ) ( -s+4 t+2 ) } \\ & \quad + \frac{36 c_2 c_3 ^ 2 c_4 z^{11}}{(s+4 ) ( 4t - s+1 ) ( 4t - s+2)}+\frac{48 c_2 c_3 c_4 ^ 2 z^{12}}{(s+4 ) ( 4 t - s ) ( 4t - s+1 ) } \\ & \quad
+ \frac{12 c_3 ^ 3 c_4 z^{13}}{(s+3 ) ( 4t - s+1 ) ( 4t - s+2)}+\frac{12c_3 ^ 3 c_4 z^{13}}{(s+4 ) ( 4t - s+1 ) ( 4t - s+2)},\displaybreak[0]\\\ \text{and}\quad h_4(z ) = & \
\frac{64 c_1 c_3 ^ 2 c_4 ^ 2 z^{14}}{(s+4 ) ( 4 t - s ) ( 4t - s+1 ) ( 4t - s+2 ) } + \frac{48 c_2 c_3 ^ 2 c_4 ^ 2 z^{15}}{(s+4 ) ( 4 t - s ) ( 4t - s+1 ) ( 4t - s+2 ) } \\ & \quad + \frac{16 c_3 ^ 3 c_4 ^ 2 z^{17}}{s ( s+4 ) ( 4
t - s ) ( 4t - s+1 ) ( 4t - s+2)}.\end{aligned}\ ] ] each term in corresponds to a translation of the original integral and converges at @xmath440 .
in addition , the lack of a degree 2 term in @xmath1 means that no term of any @xmath446 has both @xmath447 and @xmath448 as factors in its denominator .
thus there are entire functions @xmath449 , @xmath450 , and @xmath451 in @xmath77 and @xmath78 such that @xmath452 from this expression we see that while @xmath453 independently of @xmath49 and @xmath222 , we also obtain two functions @xmath449 and @xmath450 that are also solutions of @xmath260 .
explicit calculation reveals that @xmath454 for any choice of @xmath222 .
these span the laurent series solutions of the system @xmath455 , which has dimension two only at this parameter @xcite .
the vanishing of @xmath456 at @xmath436 together with the appearance of @xmath449 and @xmath450 illustrates the first direct relationship between the computation of the local cohomology of the commutative ring @xmath457 / \
< \square_u\mid au=0\>$ ] with respect to @xmath458 and the laurent polynomial solutions of @xmath455 .
bernd sturmfels and nobuki takayama , grbner bases and hypergeometric functions , grbner bases and applications ( linz , 1998 ) , 246258 , _ london math .
lecture note ser_. * 251 * , cambridge univ .
press , cambridge , 1998 . | we consider integrals that generalize both the mellin transforms of rational functions of the form @xmath0 and the classical euler integrals .
the domains of integration of our so - called euler
mellin integrals are naturally related to the coamoeba of @xmath1 , and the components of the complement of the closure of the coamoeba give rise to a family of these integrals . after performing an explicit meromorphic continuation of euler
mellin integrals , we interpret them as @xmath2-hypergeometric functions and discuss their linear independence and relation to mellin
barnes integrals
. = 1ex |
recent wasa experiments at celsius @xcite and cosy @xcite on the basic double - pionic fusion to deuterium identified a narrow isoscalar resonance structure with mass @xmath2 2.37 gev and width @xmath3 70 mev in the total cross section of the reactions @xmath4 and @xmath5 .
the differential distributions are consistent with a spin - parity assignment of @xmath6 to this structure .
subsequent measurements of two - pion production reactions , where the two colliding nucleons do not fuse to deuterium , but stay unbound , also show this resonance effect , if the reaction contains isoscalar parts @xcite . the final proof for this structure to represent a genuine @xmath7-channel resonance has been achieved by polarized @xmath8 scattering in the energy region of interest .
the obtained analyzing power data produce a pole in the coupled @xmath9 partial waves at @xmath10 mev , if included in the said data base with subsequent partial - wave analysis @xcite . henceforth , this state has been denoted by @xmath0 following the convention used for nucleon excitations .
the golden reaction channel for the observation of @xmath0 turned out to be @xmath4 , since there the background from conventional processes due to @xmath11-channel roper and @xmath12 excitations is smallest . since wasa has been the only detector with a nearly full solid angle coverage for both charged and neutral particles , which was placed at a hadron accelerator , it is of no surprise that there were no data for this channel from previous measurements .
thus it was left to the wasa collaboration to reveal the pronounced lorentzian energy dependence sitting upon an only small background in the total cross section of this channel . the experimental dalitz plot at resonance in this golden channel @xmath13 points to a @xmath12 excitation in the intermediate state @xcite , which means that @xmath0 behaves asymptotically like a bound @xmath12 system .
we note that already in 1964 dyson and xuong @xcite predicted exactly such a state , based on su(6 ) symmetry considereations , with a value for the mass remarkably close to the one of the now observed dibaryon resonance . later - on kamae and fujita @xcite as well as goldman _
_ @xcite predicted such a state , though the latter with a markedly lower mass .
only recent modern quark - model calculations see this state properly near the experimental mass @xcite .
also new relativistic faddeev - type calculations carried out by gal and garcilazo by use of hadronic interactions find this state at the correct mass @xcite .
if we account just for the well - known momentum dependence of the width of the @xmath14 resonance , then we expect for a conventional @xmath12 system bound by 80 mev a decay width of about 160 mev @xcite .
this is more than twice that observed for @xmath0 .
hence it is of no surprise that theoretical calculations predict a too large width for this state . until very recently gal and garcilazo @xcite came closest with about 100 mev for the width , if they allow for all decay channels discussed in the next section .
this discrepancy in the width might indicate some exotic contribution , which hinders the decay of @xmath0 such as hidden color as discussed in refs .
@xcite . in fact , very recently it has been shown in ref .
@xcite that the experimentally observed small width can be reconciled theoretically , if the hidden color aspect is taken into account .
to clarify the experimental situation , we examine in the following , whether all major hadronic decays of @xmath0 have been identified and understood or whether a substantial decay branch has escaped detection so far .
we consider the following reaction scenario as suggested by the data on two - pion production @xcite : @xmath15 where @xmath0 denotes an @xmath7-channel resonance both in @xmath16 and @xmath12 systems . by this scenario ,
we neglect a possible direct decay @xmath17 , but we shall come back to this point in the next to next section .
note that an intermediate @xmath18 configuration is excluded by isospin .
first , we consider the possible decay channels in the scenario of eq .
in particular we estimate the partial decay width into the elastic @xmath16 channel .
the cross section of the isoscalar two - body resonance process @xmath20 is given by the relativistic breit - wigner formula @xcite : @xmath21 where @xmath22 denotes the initial center - of - mass momentum .
as best estimates for mass and width of the resonance , we take the average over the results from elastic scattering and two - pion production , _
i.e. _ @xmath23 = 2.375 gev and @xmath24 = 75 mev . with @xmath25 = 3 and @xmath26 = 1/2 , the peak cross section at @xmath27 = 2.375 gev ( @xmath22 = 0.73 gev/_c _ ) is then @xmath28 with @xmath29 since we also have @xmath30 we get from ( 3 ) and ( 5 ) : @xmath31 in general , partial and total widths are momentum dependent quantities .
therefore , branching ratios , as we will derive below for d*(2380 ) , have been defined by convention to be quoted at the breit - wigner resonance mass or at the resonance pole , see particle data group recommendations @xcite . in case of d*(2380 ) both coincide within uncertainties .
hence , we need to consider the resonance cross sections only at its peak position .
also , in all relevant wasa experiments the energy resolution has been similar , in the range of 10 - 20 mev , which is small compared to the resonance width .
hence the smearing effect on the resonance curve is small and in particular comparable in all measured channels , so that this effect tends to cancel in the branching ratios to large extent . to estimate @xmath32 consider the total cross sections of all channels , where the isoscalar @xmath12 system can decay into : * \(i ) @xmath33 and @xmath34 : + due to isospin rules we expect @xmath35 however , due to the isospin violation in the pion mass , the available phase space is somewhat smaller for charged pion production than for the production of the lighter neutral pions . in ref . @xcite it has been shown that this results in a resonance cross section , which is lower by about 20@xmath36 in case of the @xmath34 channel .
hence we have @xmath37 + the peak cross section of the @xmath4 reaction at @xmath38 = 2.37 gev has been measured to be 0.27 mb @xcite .
this includes the contributions of the @xmath11-channel @xmath12 and roper excitations .
accounting for these background effects , the pure resonance cross section in this channel amounts to about 0.24 mb , _
i.e. _ , @xmath39 0.62 mb . since the cross sections of the three fusion reactions @xmath4 , @xmath40 and @xmath41 measured in ref .
@xcite are closely connected by isospin , only a small uncertainty of about 10@xmath36 in absolute normalization has been quoted .
this leads to @xmath42 * \(ii ) @xmath43 , @xmath44 , @xmath45 and @xmath46 isoscalar parts : + first , we consider the @xmath43 channel .
though both the @xmath47 pair and the @xmath48 pair are isovector pairs , together they may couple to total isospin @xmath49 .
hence the isoscalar resonance @xmath0 may also decay into the isoscalar part of the @xmath43 channel .
in fact , the decay of the resonance into the @xmath43 channel proceeds via the same intermediate @xmath50 system as the @xmath33 channel does .
from isospin coupling we expect that the resonance decay into the @xmath43 system should be half of that into the @xmath45 system , which is in the order of 0.2 mb
see next to next paragraph .
in fact , a recent measurement @xcite of this channel by wasa - at - cosy is in agreement with a resonance contribution of 0.10(1 ) mb in the total cross section at @xmath38 = 2.37 gev .
+ the @xmath44 channel is just the isospin mirrored one to the @xmath43 channel .
hence it has to have the same resonance contribution .
+ in a recent paper @xcite , fldt and wilkin present an estimate of the resonance cross section in the @xmath51 reaction . according to their calculation based on final state interaction theory , the expected peak cross section in the deuteron breakup channel @xmath52 is about 85@xmath36 that of the non - breakup channel @xmath33 , _ i.e. _ about 0.2 mb .
recently also albaladejo and oset @xcite estimated the expected resonance cross sections in @xmath51 and @xmath53 using a more elaborate theoretical procedure .
their result for the @xmath51 channel is compatible with that from ref .
in fact , a recent measurement at wasa @xcite shows that the data are in accordance with a contribution of @xmath0 with a strength of 0.20(3 ) mb .
+ the resonance effect in the isoscalar part of the @xmath46 channel is composed of the configurations , where either both @xmath1 and @xmath54 pairs couple each to @xmath55 or both pairs each to @xmath56 .
the first case gives just twice the contribution in the @xmath45 channel .
the latter case provides the same situation as in the @xmath43 channel .
hence , we have + @xmath57 + note that in these non - fusion channels there is no abc effect , _
i.e. _ no low - mass enhancement in the @xmath58-invariant mass spectra @xcite .
hence the phase - space reduction due to different masses of charged and neutral pions as discussed above for the @xmath34 channel does not play a significant role here .
+ our estimate for the resonant @xmath53 cross section is in good agreement with that of ref .
@xcite , where , however , only the contributions with @xmath55 coupled nucleon and pion pairs were considered .
our result also is compatible with available measurements for this channel , see ref .
@xcite , including the newest results from hades @xcite .
+ in total we have from these four reactions @xmath59 altogether we get as an estimate @xmath60 putting this into eq .
( 6 ) and selecting the minus sign in front of the root ( see discussion below ) we obtain @xmath61 which in turn corresponds to a resonance cross section in the elastic @xmath16 channel of only @xmath62 if we apply eq .
( 3 ) for the incident channel . from the said partial - wave analysis of elastic @xmath1 scattering including the new wasa data on polarized @xmath8 scattering in the energy region of @xmath0 an elastic partial width of @xmath63 = 10(2 )
mev has been derived corresponding to a branching ratio of 12(3)@xmath36 @xcite .
this result agrees reasonably well with the value obtained above in eq .
( 13 ) .
we note in passing that the other solution of eq .
( 6 ) the one with the plus sign leads to the complementary result , namely @xmath63 = 67 mev thus implying that the resonance would be predominantly elastic , _ i.e. _ , mainly decay into the elastic channel and only weakly into the pion - production channels .
this solution is at obvious variance with elastic @xmath1 scattering data . from the peak cross sections given under ( i ) and ( ii ) as well as from eqs .
( 3 ) - ( 12 ) we may readily calculate the branching ratios br : = @xmath64 for the decay of the resonance into the individual @xmath19 channels .
the results are listed in table 1 .
.experimental branching ratios ( br ) of the @xmath65 resonance into its decay channels based on eqs .
( 1 - 3 ) and ( 12 ) and the peak cross sections given under ( i ) and ( ii ) . [ cols="<,<,<,<,<",options="header " , ] + , @xmath40 ( isoscalar part ) , @xmath66 and @xmath67 ( from top to bottom ) as obtained from wasa measurements @xcite and refs .
@xcite by subtraction of the conventional background as given in refs .
, title="fig : " ] , @xmath40 ( isoscalar part ) , @xmath66 and @xmath67 ( from top to bottom ) as obtained from wasa measurements @xcite and refs . @xcite by subtraction of the conventional background as given in refs .
, title="fig : " ] , @xmath40 ( isoscalar part ) , @xmath66 and @xmath67 ( from top to bottom ) as obtained from wasa measurements @xcite and refs . @xcite by subtraction of the conventional background as given in refs .
title="fig : " ] , @xmath40 ( isoscalar part ) , @xmath66 and @xmath67 ( from top to bottom ) as obtained from wasa measurements @xcite and refs .
@xcite by subtraction of the conventional background as given in refs .
, title="fig : " ] the experimental branching ratios listed in table 1 add up to 103(15)@xmath36 , which is consistent within the assumption in eq .
( 1 ) that the @xmath65 decay proceeds solely via the intermediate @xmath12 system and the @xmath1 channel . from the data on differential observables , in particular dalitz plots and @xmath68-invariant mass spectra , we know that these are the dominant decay channels .
but this does not exclude that there might be other small decay channels .
_ e.g. _ , it has been proposed that there might be a 5@xmath36 decay branch @xmath69 with an unusual light and narrow @xmath70 particle , in order to describe the low - mass enhancement ( abc effect ) , see refs .
@xcite . from the fact that we have two independent results for the elastic partial width , namely the one obtained from the said partial - wave analysis and the one derived in eq .
( 8) , we may obtain an upper limit for decay branchings not covered by eq .
if we account for such branching by a partial width @xmath71 , then @xmath24 in eq.(6 ) is to be replaced by @xmath72 .
assuming then , e.g. , @xmath73 to be in the order of 10 @xmath36 leads subsequently to a reduction of our result for the elastic partial width from 8 to 7 mev , i.e. the agreement with the said results fades away .
thus we estimate that the sum of branchings missed by eq .
( 1 ) can not be larger than 10 @xmath36 .
as an alternative to the consideration of peak cross sections we may also exam the full resonance effect in the various reaction channels . since the signal - to - background ratio is largest at the resonance peak , an integration over the full resonance region is naturally more sensitive to the assumed background and also to the negligence of interferences of the resonance with the background .
however , statistics improves that way and this method is largely independent of the experimental energy resolution .
hence this method may serve as a cross - check of the results obtained by the peak cross section method .
1 shows the background subtracted cross sections for the reactions @xmath4 , @xmath40 ( isoscalar part ) , @xmath74 and @xmath75 ( from top to bottom ) as obtained from wasa measurements @xcite and refs .
@xcite by subtraction of the conventional background given in refs .
@xcite . in order to keep the background problem as small as possible , but to account still sufficiently for any energy smearing due to the experimental resolution , we do not integrate over the full resonance region , but only from @xmath76 to @xmath77 . if we normalize then to the branching ratio of 14 @xmath36 for the @xmath65 decay into @xmath33 we obtain 21(1 ) @xmath36 , 11(1 ) @xmath36 and 7(1)@xmath36 for the branching ratios the @xmath65 decay into @xmath34 , @xmath52 and @xmath43 channels , respectively .
these results agree very well with those from the peak cross section method given in table 1 . one such channel , which has not yet been investigated experimentally , is @xmath79 , _
i.e. _ concerns isoscalar single - pion production . since
single - pion production in @xmath80 collisions is either purely isovector or isospin mixed , the isoscalar part has to be obtained by combination of various cross section measurements .
most favorable appears the combination @xcite : @xmath81 which derives from isospin decomposition of single - pion production @xcite .
experimentally the most difficult part is the measurement of the @xmath82 reaction , since it affords either neutron beam or target .
technically this may be achieved by use of deuteron beam or target and measurement of the above reaction in the quasi - free mode . since this necessitates ,
however , exclusive and kinematically complete measurements , in order to obtain reliable results , the data base on that is sparse , see ref .
in particular , there are no data in the region of the @xmath0 resonance . also theoretically , it is difficult to construct a process of sizable cross section , where @xmath0 can decay into the single - pion channel . as mentioned at the beginning of this section ,
an intermediate @xmath83 system is isospin forbidden .
hence the next simple candidate configuration would be @xmath84 .
however , spin - parity @xmath85 of @xmath0 would require a @xmath86-wave between @xmath87 and @xmath88 .
since the resonance energy is just at the @xmath89 threshold , the probability for such a decay must be tiny already from the kinematical point of view @xcite .
another possibility would be , if @xmath65 decays via the resonance structure @xmath90 with @xmath91 at the @xmath18 threshold , as proposed in refs .
@xcite , which in turn can decay into the @xmath80 system .
this would give then the scenario @xmath92 .
also , the following process @xmath93 appears to be possible in principle .
however , aside from the fact this affords another reaction step beyond elastic scattering via the @xmath65 resonance , this process has to proceed with the @xmath94 partial wave , which is not dominant in single pion production . whatsoever , a careful experimental investigation of this issue appears to be appropriate . since wasa at cosy has finished its experimental program , dedicated measurements on that issue are no longer possible .
however , the existing data base at wasa taken for various purposes contains data also on the desired single - pion production channels .
a corresponding data analysis is in progress .
an electromagnetic excitation of the deuteron groundstate to the @xmath0 resonance is highly informative , since its transition formfactor gives access to size and structure of this resonance .
judging just from the electromagnetic coupling constant , we expect electromagnetic decays to be suppressed already by two order of magnitudes as is borne out , _ e.g. _ in the decay of the @xmath95 resonance
. a technical feasible excitation of @xmath0 would start by photo or electro excitation from the deuteron groundstate .
a real or virtual photon would need then to transfer two units of angular momentum , _ i.e. _ be of e2 or c2 multipolarity , which lowers the transition probability further .
in addition , the overlap in the wavefunctions of @xmath86 and @xmath0 enters profoundly .
we are aware of two theoretical calculations dealing with such a scenario @xcite , where cross sections in the range pb / sr - nb / sr are predicted for the forward angular range .
these are two orders lower than conventional processes and hence it appears very difficult to sense the resonance excitation under usual conditions . in this respect
the reaction @xmath96 appears to be attractive , since the conventional processes there are expected to be particularly small @xcite .
a way out could be polarization measurements .
the situation looks similar to the one in elastic @xmath1 scattering .
as we have shown above , the @xmath0 resonance contribution is about 0.17 mb , which is more than two orders below the total elastic cross section .
however , with help of the analyzing power , which consists only of interference terms in partial waves , it was possible to filter out reliably the resonance contribution .
the analogous case in electro or photo excitation of @xmath0 constitute measurements of the polarization of the outgoing proton in the reactions @xmath97 and @xmath98 , respectively , where @xmath99 stands for a virtual photon created in inelastic electron scattering on the deuteron . as in the analyzing power of @xmath1 scattering the angular dependence of the resonance effect in the polarization of the outgoing proton should be proportional to the associated legendre polynomial @xmath100 @xcite . therefore , the maximal resonance effect is expected to be at a scattering angle of @xmath101 = 90@xmath102 .
in fact , such an effect has already been looked for previously by kamae
_ et al . _ in corresponding data from the tokyo electron synchrotron @xcite .
in order to describe the observed large polarizations in the region of @xmath0 they fitted a number of resonances to the data , among others also a @xmath85 state . however , presumably due to the limited data base they only obtained very large widths for these resonances in the order of 200 - 300 mev as one would expect from conventional @xmath12 excitations .
recently , new polarization measurements from jlab appeared @xcite .
their lowest energy point is just in the @xmath0 region and is compatible with a maximal polarization of @xmath103 .
it confirms thus the old tokyo results in the sense that in this region there is a build - up of a very large polarization , which rapidly decreases both towards lower and higher energies , see fig . 1 in ref .
of course , a dedicated measurement over the region of interest is needed , in order to see , whether a narrow structure with the width of @xmath0 can be observed in this observable .
we have considered all @xmath19 channels , into which the isoscalar dibaryon resonance @xmath0 can decay . for all of these channels there
exist meanwhile experimental data , which show the @xmath0 resonance contribution and thus deliver the corresponding decay branchings .
these branchings are compatible with what one expects from isospin coupling , if the intermediate state is a @xmath12 configuration .
this in turn agrees with the result from the dalitz plot in the golden channel @xmath33 , where the background situation is optimal .
we add that there is no sign of this resonance observed in isovector @xmath19 channels @xcite .
so the only possible hadronic decay channel , which missed so far a careful inspection , is the @xmath78 channel though we know of no simple mechanism , by which @xmath0 could decay into such an isoscalar configuration .
however , since such a scenario has not yet been examined experimentally , a dedicated experimental investigation appears to be in order .
the electromagnetic decay is expected to be tiny compared to the hadronic decay branchings . as we also pointed out
, the @xmath0 contribution in deuteron disintegration processes will be even small compared to the background from conventional processes . a way out
will possibly be the measurement of polarization observables .
in particular , the polarization of the outgoing proton or neutron appears to be very promising as discussed above .
forthcoming measurements at mami could possibly give a decisive answer on that .
we acknowledge valuable discussions on this matter with stanley j. brodsky , a. gal , j. haidenbauer , c. hanhart , f. hinterberger , t. kamae , e. oset , i. strakovsky , g.j .
wagner , c. wilkin , a. wirzba , r. workman and z. zhang .
this work has been supported by the forschungszentrum jlich ( cosy - ffe ) , dfg ( cl 214/3 - 1 ) and stfc .
m. bashkanov _
et al . _ ,
lett . * 102 * , 052301 ( 2009 ) .
p. adlarson _
et al . _ ,
lett . * 106*,242302 ( 2011 ) .
p. adlarson _
b * 721*,229 ( 2013 ) .
p. adlarson _
et al . _ ,
c * 88 * , 055208 ( 2013 ) .
p. adlarson _
et al . _ ,
lett : b * 743 * , 325 ( 2015 ) .
h. clement , m. baskanov and t. skorodko , proc .
stori 2014 , phys .
, in press ; arxiv:1506.00557 [ nucl - ex ] .
p. adlarson _
lett . * 112 * , 202301 ( 2014 ) .
p. adlarson _
et al . _ ,
c * 90 * , 035204 ( 2014 ) .
r. workman , epj web conf . * 81 * , 02023 ( 2014 ) .
f.j . dyson and n .- h .
xuong , phys .
rev . lett .
* 13 * , 815 ( 1964 ) ; ibid .
* 14 * , 339 ( 1965 ) ( errata ) .
t. kamae and t. fujita , phys .
lett * 38 * , 471 ( 1977 ) .
t. goldman _
et al . _ ,
c * 39 * , 1889 ( 1989 ) .
h. huang , j. ping and f. wang , phys .
c * 89 * , 034001 ( 2014 ) and references therein . q. b. li and p. n. shen , j. phys .
g * 26 * , 1207 ( 2000 ) .
f. huang , z. y. zhang , p. n. shen and w. l. wang , arxiv:1408.0458[nucl - th ] .
x. q. yuan , z. y. zhang , y. w. yu , p. n. shen , phys . rev .
c * 60*,045203 ( 1999 ) .
hua - xing chen _
_ , arxiv:1410.0394 [ hep - ph ] .
a. gal and h. garcilazo , phys .
111 * , 172301 ( 2013 ) .
a. gal and h. garcilazo , nucl .
phys . a * 928 * , 73 ( 2014 ) . m. bashkanov , stanley j. brodsky and h. clement , phys .
b * 727*,438 ( 2013 ) .
yubing dong , pengnian shen , fei huang and zongye zhang , arxiv:1503.02356 [ nucl - th ] .
f.huang , p. n. shen , y.- b. dong and z. y. zhang , arxiv:1505.05395 [ nucl - th ] .
( particle data group ) , chin .
c * 38 * , 090001 ( 2014 ) .
g. fldt and c. wilkin , phys .
b * 701 * , 619 ( 2011 ) .
m. albaladejo and e. oset , phys .
c * 88 * , 014006 ( 2013 ) .
g. agakishiev _
_ , arxiv:1503.04013 [ nucl - ex ] . m. platonova and v. kukulin , phys .
c * 87 * , 025202 ( 2013 ) .
m. bashkanov , h. clement and t. skorodko , arxiv:1502.07500 [ nucl - ex ] .
t. tsuboyama _
et al . _ , phys .
c * 62 * , 034001 ( 2000 ) .
d. c. brunt , m. j. calyton and b. a. westwood , phys . rev . *
187 * , 1856 ( 1969 ) .
v. v. sarantsev _ et al .
j. a * 43 * , 11 ( 2010 ) .
j. bystricky _
et al . _ , j. phys . *
48 * , 1901 ( 1987 ) .
chun wa wong , phys .
c * 61 * , 064011 ( 2000 ) . di qing , he - ming sun and fan wang , chin .
phys . lett .
* 18 * , 885 ( 2001 ) .
m. egorov and a. fix , nucl .
a * 933 * , 104 ( 2015 ) .
et al . _ ,
* 38 * , 468 ( 1977 ) .
et al . _ ,
nucl . phys .
b * 139 * , 394 ( 1978 ) .
et al . _ ,
lett . * 42*,1321 ( 1979 ) .
b * 172 * , 509 ( 1980 ) .
k. wijesooriya _
lett . * 86 * , 2975 ( 2001 ) .
t. skorodko _
et al . _ ,
b * 679 * , 30 ( 2009 ) .
t. skorodko _
et al . _ , phys .
b * 695 * , 115 ( 2011 ) . | based on measurements , the branching ratios for the decay of the recently discovered dibaryon resonance @xmath0 into two - pion production channels and into the @xmath1 channel are evaluated .
possibilities for a decay into the isoscalar single - pion channel are discussed .
finally , the electromagnetic decay of @xmath0 is considered . |
in recent years , the range of wavelengths where coherent radiation can be generated has grown dramatically into both high and low frequency domain . remarkably , most of the methods to obtain radiation at extreme frequencies use , in one or the other way , nonlinear processes in laser - induced plasma .
one prominent example is high harmonic generation ( hhg ) where frequencies thousand times larger than the frequency of the pump pulse are excited , exploiting the recollision dynamics of electrons ionized by the intense light pulses .
more recently it was demonstrated that a two - color fs beam allows generation of new frequencies just in the opposite part of the spectrum , namely in the thz range , hundreds times smaller than the optical pump frequency . to this end , a short two - color pulse of fundamental frequency and second harmonic is strongly focused into a plasma spot @xcite the observed thz emission generated in this scheme has been attributed to the laser - induced plasma current in the asymmetric two - color field @xcite . using such scheme generation of strong thz radiation was reported , with a spectrum which can be as broad as 70 thz @xcite .
such broad - band coherent radiation can allow new applications providing the possibility to probe complex molecules or as an analytical and imaging tool in a broad range of fields .
in this article we consider a modification of the above described focusing geometry in a bulk gas by using a metallic hollow waveguide with a cladding from aluminum to guide both thz and optical radiation .
the use of a waveguide prevents the diffraction of light for all wavelengths involved .
in addition , such setup allows to realize nearly single - mode operation for a wide range of frequencies .
high threshold intensities for ionization - induced processes in the range larger than 100 tw/@xmath2 have been already realized in the context of hhg schemes in hollow waveguides @xcite .
we will show that during propagation a dramatic spectral broadening of the low - frequency part of the spectrum occurs , caused by only modest changes in the fundamental and second harmonic fields .
in particular , ultrashort pulses of 10 fs duration lead to the formation of low - frequency spectrum extended from several hundreds of ghz to approximately 150 thz . upon further propagation
we report the generation of nearly single - cycle pulses with central wavelengths around 4.5 @xmath1 m and efficiency of the order of 0.25 % .
in the previous works , thz emission by two - color fields in gases has been interpreted by four - wave mixing rectification @xcite .
in contrast , later studies attributed thz emission to the laser - induced plasma - current in the asymmetric two - color field @xcite described by the local quasi - classical model of the electron current in the laser field without taking into account propagation effects . as an alternative approach , particle - in - cell simulations were used @xcite .
those methods are , however , computationally very expensive , and propagation effects have been taken into account over distances of a few micrometers only . in such models the microscopic current is described as a sum over contributions from all electrons , born at discrete times @xmath3 .
the continuous
analog of this approach is given by ( @xcite ) @xmath4 where @xmath5 is the electron charge .
the current density @xmath6 at the time @xmath7 is obtained by integrating over all impacts from different ionization times @xmath8 : @xmath9 is the velocity of electrons at time @xmath7 which were born at time @xmath8 , and @xmath10 is the density of electrons born at time @xmath8 .. ] this approach was utilized in the previous works of kim and coauthors @xcite .
the additional integration over the variable @xmath8 increases the numerical effort , but one can eliminate the microscopic velocity distribution @xmath9 and derive a differential equation for the macroscopic current density @xmath6 . assuming no interaction with other particles and zero velocity of new - born electrons , the equation for the electron velocity can be written as @xmath11 ( where @xmath12 is the electric field , @xmath13 is the electron mass ) . with the simple additive property of integrals we can rewrite it as @xmath14 . substituting this into eq .
( [ eq:1 ] ) and differentiating with respect to @xmath7 we obtain @xmath15 .
it is easy to see that by introducing new variables @xmath16 , @xmath17 ( where @xmath18 is the electron - ion collision rate ) and proceeding analogously we can include collisional effects as well , and finally obtain @xmath19 indeed , eq .
( [ eq:14 ] ) coincides with the well known ( and very commonly used ) equation for the current in plasmas @xcite . in the present article , the static model for the tunneling ionization @xcite
is used : @xmath20 [ \rho_{\rm at } - \rho(t ) ] , \quad \quad w_{st}[e(t ) ] = \alpha \frac{e_a}{e(t ) } \exp\left\{-\beta \frac{e_a}{e(t)}\right\}.\ ] ] here , @xmath21 is the neutral atomic density , and @xmath22 vm@xmath23 the atomic field ; the coefficients @xmath24 and @xmath25 are defined through the ratio of ionization potentials of argon and hydrogen atoms @xmath26 , @xmath27 ev , @xmath28 ev , and @xmath29 s@xmath23 @xcite .
to describe the light propagation , we use the so called forward maxwell equation ( fme ) for the fast oscillating optical field @xmath30 in the waveguide with the optical axis along @xmath31 direction @xcite : @xmath32 where @xmath33 , @xmath34 , @xmath35 is the refractive index of ar , @xmath36 .
@xmath37 and @xmath38 are the fourier transforms of @xmath39 , @xmath40 with respect to time , @xmath41 is the nonlinear polarization , which includes both the kerr ( @xmath42 ) nonlinearity and the plasma contribution : @xmath43 , where @xmath44 is a loss term accounting for photon absorption during ionization ; `` division by a vector '' means the component - wise division @xcite . decomposing the field into a series of linear eigenmodes of the waveguide @xmath45 with corresponding propagation constants @xmath46
, we obtain the following set of equations for the amplitudes of the eigenmodes @xmath47 , coupled through the nonlinear polarization term @xmath48 : @xmath49
in our simulations , we assume a dielectric waveguide with aluminum coating of the inner walls and diameter @xmath50 @xmath1 m . such waveguides can be routinely produced ( see , e.g. , @xcite ) .
the main purpose of the waveguide is not to modify the dispersion relation but to confine both pump and thz radiation ( especially the latter , because it quickly spreads out otherwise as a result of diffraction @xcite ) .
taking into account that thz as well as fundamental and second harmonic frequencies undergo strong absorption due to plasma generation a long waveguide is not useful .
therefore , we restrict our simulations to distances @xmath51 cm .
the linearly - polarized waveguide mode eh@xmath52 has no cut - off at low frequencies , and the fundamental and second - harmonic frequencies can be coupled into the waveguide with almost 100 % efficiency from the free - space propagating laser mode .
the dispersion and losses for the mode eh@xmath52 of the 100-@xmath1 m waveguide are shown in fig .
[ fig:1](b ) in low - frequency range .
both quantities are calculated using the direct solution of the corresponding boundary value problem @xcite , drude model for the dispersion in aluminum @xcite as well as sellmeier formula for the dispersion in argon @xcite .
one can see that the losses are relatively large only in the small frequency range corresponding to wavelengths of the order of the waveguide diameter .
the aluminum coating was chosen because then the modeling of linear dispersion and losses in the whole frequency range from sub - thz to sub - phz is straight forward using the drude model @xcite .
in contrast , for dielectric coatings ( such as fused silica ) such simple description for the whole spectral region of interest fails .
nevertheless , we expect even weaker losses than for the metallic waveguide .
( color online ) .
( a ) scheme for thz generation considered in this article . a short pulse with center frequency @xmath53 thz ( corresponding to wavelength @xmath54 nm ) and its second harmonic are focused into an aluminum coated hollow waveguide .
the thz emission is detected at the exit of the waveguide .
( b ) losses @xmath55 ( red solid line ) and refractive index change @xmath56 ( green dashed line ) induced by the al waveguide with the diameter @xmath50
@xmath1 m for the mode eh@xmath52 versus frequency @xmath57 . ]
results from simulations of ultrashort pulses with 10 fs duration and frequencies @xmath58 thz ( corresponding to the wavelength @xmath54 nm ) and @xmath59 , having intensities @xmath60 w/@xmath2 and @xmath61 w/@xmath2 correspondingly are shown in fig .
[ fig:2 ] .
we checked that the transverse field distribution consists predominantly of the fundamental and few higher order modes . at the initial stage of propagation
, broadband thz radiation is generated from approximately 0.5 thz to 80 thz ( see fig . [ fig:2](a ) and ( b ) ) . with further propagation ( up to 2 mm )
the low - frequency part of the spectrum grows in the direction to higher frequencies ( up to 150 thz ) . at the same time , the spectral intensity around 10 thz degrades because of the strong losses of the waveguide mode . from 2
mm up to @xmath62 mm the long - wavelength part of the spectrum shifts continuously towards higher frequencies . finally , starting from 7 mm the spectrum does not change considerably anymore , except for its decrease due to losses . at this propagation distance
, the long - wavelength part of the spectrum is localized around 60 - 80 thz , with the center around 66 thz ( corresponding to a wavelength @xmath63 @xmath1 m ) .
the observed efficiency of this frequency down - conversion reaches 0.25 percent . the electric field obtained by filtering the long - wavelength part of the spectrum for different propagation distances is shown in fig .
[ fig:2](c ) .
one can see that despite the strong reshaping of the spectrum , the envelope of the obtained thz pulses does not change significantly .
the amplitude of the pulse grows considerably during propagation , from approximately 1 % of the pump amplitude at @xmath64 mm ( which correspond to the conversion efficiency @xmath65 ) to about 5 % at @xmath66 mm ( which correspond to the conversion efficiency @xmath67 ) .
[ fig:2](a ) reveals that in contrast to the strong reshaping of the low frequency components , the spectrum in the range of the pump frequencies remains almost unchanged .
nevertheless , the spectrum of the pump broadens slightly upon propagation . due to plasma - induced refraction
index change the maximum frequency of the fundamental pulse is slightly shifted to the blue @xcite . in our case
the shift is small , only about 10 thz .
as elaborated in @xcite , even small changes in the pump fields can affect dramatically the spectral shape of the generated low - frequency radiation .
such changes affect significantly the temporal positions of the ionization events .
contributions from the electrons born in different ionization events are superimposed , and interference effects lead to a reshaping the thz spectra .
( color online ) .
spectra for different input pulse durations ( 25 , 50 and 1000 fs are shown by red solid , green dashed and blue dotted lines , respectively ) for 1 cm of propagation . in ( a ) and
( b ) the full spectra and their low - frequency parts are shown , respectively .
spectra are normalized to their maximum value , so that efficiencies can be compared .
] the situation is qualitatively similar for other pulse durations ( see fig . [
fig : dd ] ) . in general , the efficiency of the thz generation is larger for shorter pulse durations ( in fig .
[ fig : dd ] , spectra are normalized to unity at their maxima , so the efficiencies can be directly compared ) .
this observation can be explained by the additional asymmetry in the current which is introduced by the finite pulse duration and therefore more pronounced for short pulses .
in conclusion , we studied thz generation by plasma currents through noble gas ionization in strong asymmetric laser fields .
we have shown that the mechanism allows the relatively simple generation of extremely wide range of frequencies , starting from thz to near infrared . by using hollow waveguides
we overcome the strong diffraction of the generated low - frequency generation .
we report that depending on the length of the waveguide either thz supercontinuum or almost single - cycle near - infrared pulses can be produced at output with @xmath68 % efficiency .
we believe that our findings may open new intriguing possibilities to control the spectral and temporal shape of low - frequency radiation generated by the ionization dynamics in intense optical fields . | the generation of thz radiation from ionizing two - color femtosecond pulses propagating in metallic hollow waveguides filled with ar is numerically studied .
we observe a strong reshaping of the low - frequency part of the spectrum .
namely , after several millimeters of propagation the spectrum is extended from hundreds of ghz up to @xmath0 thz . for longer propagation distances ,
nearly single - cycle near - infrared pulses with wavelengths around 4.5 @xmath1 m are obtained by appropriate spectral filtering , with an efficiency of up to 0.25 %
. 10 d. j. cook and r. m. hochstrasser , `` intense terahertz pulses by four - wave rectification in air , '' opt
. lett . * 25 * , 1210 ( 2000 ) .
t. bartel _
et al .
_ , `` generation of single - cycle thz transients with high electric - field amplitudes , '' opt . lett . * 30 * , 2805 ( 2005 ) .
x. xie , j. dai , and x .- c .
zhang , `` coherent control of thz wave generation in ambient air , '' phys .
rev .
lett . * 96 * , 075005 ( 2006 ) .
j. dai , x. xie , and x .- c .
zhang , `` detection of broadband terahertz waves with a laser - induced plasma in gases , '' phys .
rev .
lett .
* 97 * , 103903 ( 2006 ) .
k. reimann , rep .
. phys .
`` table - top sources of ultrashort thz pulses , '' * 70 * , 1597 ( 2007 ) .
m. d. thomson , m. kre , t. lffler , h. g. roskos , `` broadband thz emission from gas plasmas induced by femtosecond optical pulses : from fundamentals to applications , '' laser & photon . rev . * 1 * , 349 ( 2007 ) .
k .- y .
kim , j. h. glownia , a. j. taylor , and g. rodriguez , `` terahertz emission from ultrafast ionizing air in symmetry - broken laser fields , '' opt .
express * 15 * , 4577 ( 2007 ) , http://www.opticsexpress.org/abstract.cfm?uri=oe-15-8-4577 .
k. y. kim , a. j. taylor , j. h. glownia , and g. rodriguez , `` coherent control of terahertz supercontinuum generation in ultrafast laser - gas interactions , '' , nat .
photon .
* 2 * , 605 ( 2008 ) .
t. d. wang , z .- m .
sheng , h .- c .
wu , m. chen , c. li , j. zhang , and k. mima , `` strong terahertz pulse generation by chirped laser pulses in tenuous gases , '' opt .
express * 16 * , 16999 ( 2008 ) , http://www.opticsexpress.org/abstract.cfm?uri=oe-16-21-16999 .
m. chen , a. pukhov , x .- y .
peng , and o. willi , `` theoretical analysis and simulations of strong terahertz radiation from the interaction of ultrashort laser pulses with gases , '' phys . rev .
e * 78 * , 046406 ( 2008 ) .
k .- w .
kim , `` generation of coherent terahertz radiation in ultrafast laser - gas interactions , '' phys .
plasma * 16 * , 056706 ( 2009 ) .
j. dai , n. karpowicz , and x .- c .
zhang , `` coherent polarization control of terahertz waves generated from two - color laser - induced gas plasma , '' phys . rev
. lett . * 103 * , 023001 ( 2009 ) . c. g. durfee , a. r. rundquist , s. backus , c. herne , m. h. murnane , and h. c. kapteyn , `` phase matching of high - order harmonics in hollow waveguides , '' phys .
rev .
lett . * 83 * , 2187 ( 1999 ) .
e. a. gibson , a. paul , n. wagner , r. tobey , d. gaudiosi , s. backus , i. p. christov , a. aquila , e. m. gullikson , d. t. attwood , m. m. murnane , and h. c. kapteyn , `` coherent soft x - ray generation in the water window with quasi - phase matching , '' science , * 302 * , 95 ( 2003 ) .
p. sprangle , j. r. peano , b. hafizi , and c. a. kapetanakos , `` ultrashort laser pulses and electromagnetic pulse generation in air and on dielectric surfaces , '' phys .
rev .
e * 69 * , 066415 ( 2004 ) .
l. berg , s. skupin , r. nuter , j. kasparian and j .- p . wolf , `` ultrashort filaments of light in weakly ionized , optically transparent media , '' rep .
. phys . * 70 * , 1633 ( 2007 ) .
l. d. landau , and e. m. lifshitz , _ quantum mechanics _
( pergamon , new york , 1965 ) , 2nd ed . , p. 276 .
a. v. husakou and j. herrmann , `` supercontinuum generation of higher - order solitons by fission in photonic crystal fibers , '' phys .
. lett . * 87 * , 203901 ( 2001 ) .
p. tzankov , o. steinkellner , j. zheng , m. mero , w. freyer , a. husakou , i. babushkin , j. herrmann and frank noack , `` high - power fifth - harmonic generation of femtosecond pulses in the vacuum ultraviolet using a ti : sapphire laser , '' opt . express * 15 * , 6389 ( 2007 ) , http://www.opticsexpress.org/abstract.cfm?uri=oe-15-10-6389 .
e. a. v. markatili and r. a. schmeltzer , `` hollow metallic and dielectric waveguides for long distance optical transmission and lasers , '' bell sys .
tech .
j. * 43 * 1784 ( 1964 ) .
_ handbook of optics _
, ed . by m. bass , mcgraw - hill , ( 1994 ) .
p. j. leonard , `` refractive indices , verdet constants and polarizabilities for inert gases , '' atom .
data nucl .
data tables * 14 * , 21 ( 1974 ) .
w. m. wood , c. w. siders , and m. c. downer , `` measurement of femtosecond ionization dynamics of atmospheric density gases by spectral blueshifting , '' phys .
rev .
lett . * 67 * , 3523 ( 1991 ) . s.c .
rae and k. burnett , `` detailed simulations of plasma - induced spectral blueshifting , '' phys .
rev .
a * 46 * , 1084 ( 1992 ) .
i. babushkin , w. kuehn , c. khler , s. skupin , l. berg , k. reimann , m. woerner , j. herrmann , and t. elsaesser , `` ultrafast spatio - temporal dynamics of terahertz generation by ionizing two - color femtosecond pulses in gases , '' in preparation . |
the time - dependent densities of the evolving initially - prepared bright soliton trains shown in fig . 1 reveal intricate `` interactions '' between the density humps . on the gross - pitaevskii ( gp ) or non - linear schrdinger equation level ( see in this respect refs .
[ s1,s2 ] ) , @xmath8-phase symmetric and asymmetric trains show attraction between the density humps , which manifests itself as `` bound '' periodic - like motion of the density humps one with respect to the other .
the @xmath9-phase anti - symmetric train reveals repulsive interaction between the density humps , which would eventually lead to their increasing separation from each other . on the many - body level , we have seen in fig . 1 that repulsive interaction between the density humps eventually prevails , even when the @xmath8-phase symmetric and asymmetric trains are used as the initially - prepared wave - packets of the attractive system .
in this section we would like to contract a simple mean - field model to qualitatively explain the forces between the density humps when soliton trains become fragmented and die .
let us commence with the development of fragmentation in the @xmath8-phase symmetric system .
the natural orbitals of the decaying soliton train ( see left lower panel of fig . 2 )
are composed of gerade ( even ) and ungerade ( odd ) delocalized functions , i.e. , they simultaneously exist in the right and left sub - clouds . oversimplifying the many - body picture of the decay of soliton trains and the development of fragmentation in the system
, we can imagine that at every point in time the system is described by a two - orbital mean - field state made of a single permanent ( fock state ) @xmath40 [ s3 ] with time - dependent occupation numbers and orbitals .
the orbitals comprising this permanent are the above described delocalized even and odd functions . in time , the initially - prepared condensed state @xmath41 ( built from the delocalized even orbital only , see left panels of fig .
2 ) becomes fully fragmented at about @xmath42 ( see left upper panel of fig . 2 ) .
then it is described by the mean - field state @xmath43 ( built from both the delocalized even and odd orbitals ) .
such a fragmetned state has been recently predicted to exist in one dimensional attractive becs and called fragmenton [ s4 ] .
what we would like now to compare within this simplified model is the inter - hump forces in the condensed state @xmath41 and fully fragmented state @xmath43 . for completeness , the inter - hump forces in the condensed state @xmath44 ( built from the delocalized odd orbital only , see right panels of fig .
2 ) will be registered as well .
the delocalized even and odd functions used in our model are conveniently taken as the @xmath8-phase symmetric and @xmath9-phase anti - symmetric linear combinations @xmath45 \pm \mathrm { sech } [ \gamma(x+x_0)]\right\}$ ] , which reproduce on the mean - field level the limiting cases of the bright soliton trains and the fragmenton .
the energy of the fock state @xmath40 is given as the expectation value of the hamiltonian @xmath46 and depends on the parameters @xmath10 ( inversely proportional to the width of the density humps ) and @xmath47 ( half the separation between the density humps ) , and on the relative occupation numbers @xmath48 and @xmath49 ( @xmath50 ) . by taking the derivative @xmath51
one immediately arrives at the inter - hump force @xmath52 .
the full expression is quite lengthy , but in the large @xmath53 limit , i.e. , when the humps are well separated from each other , we can prescribe an asymptotic analytical force in the lowest order : @xmath54 @xmath55 .
\eqno{\mathrm { ( s1)}}\ ] ] for pure condensed states , i.e. , when @xmath56 or @xmath57 with the optimal width @xmath58 , this general expression takes on a very simple form ( see in this respect ref .
[ s2 ] ) : @xmath59 where the `` - '' corresponds to the @xmath8-phase symmetric and the `` + '' to the @xmath9-phase anti - symmetric two - hump soliton .
the damping exponent originates from the residual overlap between the left and right sub - clouds . in fig .
s1 we plot the full inter - hump forces and their asymptotic expansions given in eqs .
( s1,s2 ) as a function of the inter - hump separation @xmath60 for the system of @xmath1 bosons and @xmath39 discussed in the main text . in the one limiting case , @xmath56 ,
the force between the density humps of the @xmath8-phase symmetric soliton train is negative , i.e. , attractive . in the other limit , @xmath61 ,
the force between the density humps of the @xmath9-phase anti - symmetric soliton train is positive , i.e. , repulsive . for the pure fragmenton @xmath43 this force vanishes asymptotically , because of the @xmath62 prefactor in eq .
this fact allows us to predict that the inter - hump interaction in the fragmented states is always weaker than in the respective two - hump coherent soliton trains .
interestingly , the residual interaction for the pure fragmenton at a large but not yet asymptotic separation is , nevertheless , repulsive . by going from one limit to another ,
i.e. , by transferring the bosons , say , from the @xmath8-phase symmetric to the @xmath9-phase anti - symmetric natural orbital , we see that the character of the interaction gradually changes from attractive to repulsive .
in the present section we would like to investigate the scaling of the lifetime of two - hump solition trains with the number of bosons @xmath38 in the system .
we concentrate on the @xmath8-phase symmetric soliton train . to make the comparison
, we keep the gp parameter @xmath63 and inter - hump separation @xmath64 fixed to the values used for @xmath1 throughout the main text .
it should be reminded that for a fixed value of @xmath28 the gp dynamics is independent of @xmath38 ( for @xmath65 , which is the situation here ) .
additionally , no fragmentation can be described by the gp equation , of course , and solitons lifetime is infinite .
hence , the many - body results presented in this section are to further demonstrate the substantial physical differences between the mean - field and many - body quantum dynamics of 1d attractive bose systems .
s2 collects the results for the lifetime as a function of @xmath38 .
we remind that the lifetime is defined as the time @xmath66 where the fragmentation ratio of @xmath67 is reached . translating from dimensionless to dimension - full quantities ,
the lifetime of , e.g. , the two - hump soliton made of @xmath1 @xmath31li atoms in a quasi-1d cigar - shaped trap with transverse confinement of @xmath35 hz is about @xmath68 milliseconds .
this lifetime is short compared to current experimental setups which are a couple of seconds long .
it is seen from fig .
s2 that the lifetime decreases roughly quadratically with the number of bosons @xmath38 . for @xmath69 @xmath31li atoms
it is below @xmath70 milliseconds .
all in all , we demonstrate that the death of soliton trains is an ultrafast process that takes place on a time scale which is an order of magnitude shorter than relevant time - scales for current experimental setups .
the density profile of bright soliton trains consists of localized structures . as we have seen in the main text this well - known gp ( mean - field ) characteristic
persists on the many - body level as well , despite the fact that the system becomes fragmented .
consequently , the spatial density of low - dimensional attractive systems is not a straightforward indicator for the development of fragmentation in the system , and hence is not likely to be utilized experimentally to identify it . in this section
we would like to augment the indicting information on the death of solitons by reporting the corresponding momentum distribution , see fig .
the time - dependent momentum distribution @xmath71 obtained by solving the time - dependent gp equation for the two - hump solitons of fig . 1
is plotted in fig .
s3 . in the left upper panel ,
s3a , we plot the momentum distribution of the @xmath8-phase symmetric soliton train . a sharp contrast and oscillatory pattern
is seen , which signifies the system being condensed ( and coherent , see right upper panel of fig .
as the two density humps approach each other in real space , the corresponding momentum distribution broadens . at the collision point ( @xmath14 ) the momentum distribution is the broadest ,
since the two colliding clouds momentarily merge into one cloud , yet its maximal value is significantly smaller . in the left middle panel ,
s3b , the momentum distribution of the @xmath9-phase anti - symmetric soliton is depicted . due to the larger separation between the density humps ( @xmath72 ,
see fig .
1b ) , more maxima and narrower peaks in @xmath71 are seen , in comparison to the soliton train of fig .
s3a ( for which @xmath73 , see fig
furthermore , the momentum distribution is essentially constant for all times shown , as the corresponding spatial density is , due to the initial larger separation between the density humps . finally , in fig .
s3c , the momentum distribution of the @xmath8-phase asymmetric soliton train is shown .
it also depicts a sharp contrast pattern , but asymmetric .
the broadest peaks in @xmath71 correspond to the near collision of the two density humps in real space occuring at about @xmath14 ( see fig .
we may summarize the following features of the momentum distribution of two - hump solitons computed by the gp theory : ( i ) appearance of sharp contrast pattern signifying condensation ( and coherence ) , and ( ii ) the number of peaks seen and their width depend on the initial separation of the density humps in real space ( in a reminiscence of a double - slit experiment with light ) . on the many - body level
, the time - dependent momentum distribution @xmath71 shows essential visual differences from the gp momentum distribution , compare the left and right panels of fig .
s3 . in the right upper panel ,
s3d , the many - body momentum - distribution dynamics of the initially - prepared @xmath8-phase symmetric soliton train is shown . for times
up to about @xmath74 , the sharp contrast pattern characterizing the initially - prepared condensed system survives .
looking at the evolution of the corresponding occupation numbers ( see left upper panel of fig .
2 ) , we see that up to this time the system remains at least @xmath75 condensed . for times longer than just about @xmath74
, the many - body momentum distribution almost suddenly becomes `` blared '' , in sharp contrast to the gp results .
this is yet another signature of the fragmentation of the system .
the `` blared '' momentum distribution pattern can be anticipated by `` interfering '' the momentum distributions carrying the two delocalized even and odd natural orbitals of the fragmented system ( see in this respect the left lower panel of fig . 2 ) .
similarly , the many - body momentum distribution of the initially - prepared @xmath9-phase anti - symmetric wave - packet , fig .
s3e , and the many - body momentum distribution of the initially - prepared @xmath8-phase asymmetric wave - packet , fig .
s3f , quickly become `` blared '' , when fragmentation sets it .
we stress that , despite the large separation between the density humps of the @xmath9-phase anti - symmetric wave - packet ( see fig .
1e ) , the changes in the physics are dramatic .
summarizing , the momentum distribution of low - dimensional attractive systems initially - prepared as ( condensed and coherent ) soliton trains quickly loose their contrast and become essentially single - peaked and `` blared '' . this consequence and fingerprint of fragmentation should be amenable to experimental detection , and is far better and straightforward indicator of fragmentation than the monitoring of localization in coordinate space . -phase symmetric two - hump soliton trains as a function of the total number of bosons @xmath38 ( the solid line connecting the open circles is to guide the eye only ) .
the gross - pitaevskii parameter @xmath76 and inter - hump separation @xmath64 are kept fixed ( corresponding to the system of @xmath1 bosons with @xmath39 discussed in the main text ) .
the lifetime decreases roughly quadratically with the number of bosons @xmath38 .
it is demonstrated that the death of soliton trains is an ultrafast process that can take place on a time scale which is an order of magnitude shorter than relevant time - scales for current experimental setups ( couple of seconds).,width=453 ] -2.0 truecm of the attractive system of @xmath1 bosons with @xmath39 computed by the popular gross - pitaevskii equation on the mean - field level in panels a)-c ) , and the corresponding dynamics computed on the many - body level in panels d)-f ) .
panels a ) and d ) show the dynamics of a @xmath8-phase symmetric train ; panels b ) and e ) show the dynamics of a @xmath9-phase anti - symmetric train ; panels c ) and f ) show the dynamics of a @xmath8-phase asymmetric train . on the mean - field level , panels a)-c ) ,
an interference - like sharp - contrast pattern emerges , indicating the two - hump system being condensed . on the many - body level , panels d)-f )
, the momentum distribution is much more blared , indicating the fragmentation of the system .
all quantities shown are dimensionless.,title="fig:",width=453 ] | experiments on ultra - cold attractive bose - einstein condensates ( becs ) have demonstrated that at low dimensions atomic clouds can form localized objects , propagating for long times without significant changes in their shapes and attributed to bright matter - wave solitons , which are coherent objects .
we consider the dynamics of bright soliton trains from the perspective of many - boson physics .
the fate of matter - wave soliton trains is actually to quickly loose their coherence and become macroscopically fragmented becs .
the death of the coherent matter - wave soliton trains gives birth to fragmented objects , whose quantum properties and experimental signatures differ substantially from what is currently assumed . solitons are common wave - packet phenomena in many areas of science and engineering @xcite .
such localized structures can form when a wave - packet s dispersion is compensated by self - focusing ( attractive ) forces .
in this context , the physics of low - dimensional attractive bose - einstein condensates ( becs ) has attracted much attention @xcite .
the similarity between the gross - pitaevskii ( gp ) equation @xcite commonly used to describe these ultra - cold quantum systems and the non - linear schrdinger equation used in optics @xcite to characterize self - focusing of light has encouraged the transfer of ideas , phenomena , and understanding from optics to ultra - cold atomic physics . particularly , bright matter - wave solitons have been predicted to occur in low - dimensional attractive bose gases @xcite , stimulating the recent experiments @xcite . very often , along with a single soliton , multi - hump matter - wave structures propagating for long times without significant changes in their shapes have been observed .
since the gp equation supports localized multi - hump solutions as excited states , the experimentally observed multi - hump structures have been attributed to soliton trains , which are coherent objects .
this is the current widely accepted interpretation in the literature . nowadays
the gp equation ( i.e. , the non - linear schrdinger equation ) is the cornerstone of studying soliton trains in attractive becs .
it is important to remember , however , that ultra - cold attractive becs are quantum many - particle systems , governed by the many - boson schrdinger equation .
hence , we consider in this work the dynamics of bright soliton trains from the perspective of many - boson physics .
as we shall show the fate of matter - wave soliton trains is to quickly loose their coherence and become macroscopically fragmented becs @xcite .
the death of the coherent matter - wave soliton trains gives birth to fragmented objects , whose quantum properties and experimental signatures differ substantially from what is currently assumed .
our starting point is the gp dynamics @xmath0 \psi(x , t ) = i\frac{\partial \psi(x , t)}{\partial t}$ ] of a soliton train with two humps ( briefly , two - hump soliton ) .
we consider in this work @xmath1 bosons in one dimension and attractive inter - particle interaction of strength @xmath2 . for convenience
, we use dimensionless quantities which are readily arrived at by dividing the dimension - full hamiltonian by @xmath3 , where @xmath4 is the mass of a boson and @xmath5 is a length scale .
the corresponding dimension - full quantities are within range of current experimental setups and discussed below . as an initial profile ( shape ) of the two - hump soliton
we take the commonly used linear combination of sech - shaped functions : @xmath6 \pm \mathrm { sech } [ \gamma(x+x_0 ) ] \right\},\ ] ] where @xmath7 is the normalization factor .
the `` + '' refers to a @xmath8-phase symmetric and the `` - '' to a @xmath9-phase anti - symmetric soliton train .
the parameter @xmath10 is inversely proportional to the width of the humps . by assuming a large separation between humps , one gets by minimizing the gp energy functional the optimal value of @xmath11 .
the time - dependent density @xmath12 obtained by solving the time - dependent gp equation for different two - hump solitons is plotted in fig . 1 .
in the left upper panel , fig . 1a , we plot the dynamics for the @xmath8-phase symmetric soliton train , initially located at @xmath13 .
the system reveals the well known mean - field oscillatory dynamics : the humps attract each other , collide at @xmath14 , then split and return to the initial separation and start a new cycle of oscillation . the left middle part , fig .
1b , shows the dynamics for the @xmath9-phase anti - symmetric soliton clouds placed initially at a larger separation of @xmath15 .
at such a separation the clouds very slightly repel each other , indicating that the inter - hump forces strongly depend on the initial separation .
in the left lower panel , fig . 1c , we depict the dynamics for a @xmath8-phase asymmetric soliton train , formed as the combination of the sech - shaped functions of slightly different widths @xmath16 ( left cloud ) and @xmath17 ( right cloud ) , placed at @xmath13 . a similar dynamics to fig .
1a is seen .
the gp theory describes the following characteristics of two - hump solitons : ( i ) the relative phase between humps defines whether the inter - hump interaction is attractive or repulsive , and ( ii ) the magnitude of the inter - hump interaction depends on the separation between the humps .
we remind the reader that gp theory is a mean - field approximation to the quantum many - boson problem which assumes all bosons to occupy one and the same quantum state throughout the system s evolution in time . what happens
if we relax the constraint of the attractive system to occupy a single quantum state ?
in other words , we allow the system to choose according to the variational principle its evolution in time by solving now the time - dependent many - boson schrdinger equation @xmath18 , with @xmath19 .
it is possible nowadays to go significantly beyond mean - field and compute the dynamics of becs on the many - body level , by solving the time - dependent many - boson schrdinger equation with the multiconfigurational time - dependent hartree method for bosons , see the literature for details @xcite and applications @xcite .
is the dynamics of soliton trains described above to change in a dramatic manner ?
we take as initial conditions the same @xmath8-phase symmetric and @xmath9-phase anti - symmetric soliton trains as in the above gp studies and propagate them now not on the mean - field but on the many - body level .
the time - dependent density @xmath12 obtained at the many - body level is plotted in the right panels of fig . 1 .
the first observation is that the initially - localized wave - packets remain localized at the many - body level of the description as well .
hence , localization is indeed a characteristic dynamical feature of attractive bosonic clouds in one dimension , as predicted at the gp level and confirmed here at the many - body level .
the many - body results , however , reveal different dynamics than the mean - field ones , compare the left and right panels of fig . 1 .
in the right upper panel , fig . 1d , we depict the many - body dynamics of the initially - prepared @xmath8-phase symmetric wave - packet .
the two separated humps start to attract each other , similarly to the mean - field dynamics , but do not come close enough to collide , as it is in the gp case , fig .
1a . instead of the full collision
the sub - clouds reach a minimal separation and then they start to depart from each other .
this observation reveals , thereby , a repulsive character of the inter - hump interaction .
so , during the many - body dynamics the interaction between the humps of the initially - prepared @xmath8-phase symmetric wave - packet changes from attractive to repulsive and stays repulsive afterwards .
the many - body dynamics of the initially - prepared @xmath9-phase anti - symmetric wave - packet is shown on the right middle panel , fig .
1e . due to the larger separation between the two humps ,
the spatial density profiles obtained at the many - body and gp levels are indeed very close to each other .
the weak repulsive interaction between the humps persists at the many - body level as well .
the right lower panel , fig .
1f , shows the density profile for the @xmath8-phase asymmetric wave - packet constructed , as above , from two functions of slightly different widths .
the dynamics obtained at the many - body level for the initially - prepared @xmath8-phase asymmetric wave - packet has a lot in common with the respective many - body dynamics obtained for the initially - prepared @xmath8-phase symmetric wave - packet , compare panel fig .
1d and fig .
1f .
the sub - clouds start to attract each other and approach some minimal distance , and then the inter - hump interaction becomes of repulsive character .
summarizing , the many - body dynamics shows that irrespective to the phase between initially - prepared humps , after some time the inter - hump interaction changes its nature and becomes repulsive .
this conclusion is general and has been checked for various initially - prepared symmetric and asymmetric wave - packets with different phases .
this behavior is in accord with the experimental observation that the inter - hump interactions in low - dimensional attractive bose gases are indeed repulsive . for more analysis of the inter - hump forces
see @xcite . the differences between the gp dynamics leading to bright soliton trains and the respective many - body dynamics collected in fig .
1 show that bright soliton trains in one - dimensional becs undergo changes in time .
we shall see below that these changes are fundamental and dramatic .
to get a deeper insight into the physics behind the intriguing many - body results , we examine further the quantum nature of the attractive bosons .
first , we perform the natural orbital analysis of the propagating wave - packets at every point of the propagation time by constructing and diagonalizing the reduced one - body density matrix .
the reduced one - body density matrix of the system is given by @xmath20 , where @xmath21 is the usual bosonic field operator creating a boson at position @xmath22 .
the eigenvalues obtained , called natural occupation numbers , are plotted in fig . 2 . all initial states , at @xmath23 , are fully condensed systems , i.e. , only one natural orbital is macroscopically occupied .
the gp theory implies that the system remains condensed all the time , i.e. , that only one natural orbital remains occupied during the evolution .
the many - body dynamics , in contrast , shows that both for the initially - prepared @xmath8-phase soliton train as well as for the @xmath9-phase soliton train the bosons quickly populate the second natural orbital , forming thereby macroscopically ( two - fold ) fragmented many - boson states .
the fate of bright soliton trains is thus death by fragmentation .
we note that fragmentation of becs is a general known and well - studied phenomena in the literature ( see @xcite and references therein ) . for the @xmath8-phase symmetric initial state
the fragmentation ratio of @xmath24 , where the condensed fraction has decreased by @xmath25 times , is reached already at @xmath26 ( see left upper panel of fig . 2 ) .
afterwards , the many - body wave - packet dynamics reveals oscillations around the two - fold fragmented state .
the @xmath9-phase soliton loses its coherence monotonously and lives a bit longer
it becomes fragmented at @xmath27 ( see right upper panel of fig . 2 ) .
we stress that the changes in the physics are dramatic despite the large separation between the density humps ( see fig .
1e ) . the lifetime of the @xmath8-phase asymmetric soliton ( not shown ) discussed above and in fig .
1 is also @xmath26 .
we have also found out that increasing the number of atoms in the system and keeping the gp parameter @xmath28 fixed , the lifetime of the soliton trains decreases , i.e. , the lifetime of the initially - prepared coherent objects becomes even shorter @xcite .
the conclusions are that initially coherent two - hump wave - packets become two - fold fragmented with time , and that the resulting coherence lifetime of the two - hump solitons is finite and ultrashort .
we have performed similar many - body propagations of initially coherent three - hump and four - hump soliton trains and found that the dynamics leads to the formation of three - fold and four - fold fragmented many - body states , respectively . summarizing , the death of multi - hump soliton trains is a generic rapid inevitable feature of the many - body dynamics of attractive bosons in one dimension , and gives birth to a potentially interesting object like the fragmenton @xcite , also see @xcite .
this result has been obtained for initially coherent states made of different numbers of atoms , different number of the constituting humps , placed at different inter - hump separations , with different inter - hump phases , and of different humps widths .
we have shown so far the death of soliton trains as an emerging many - body phenomenon in one dimensional attractive becs .
it is important to discuss how can it be resolved experimentally .
looking at fig .
1 we recall that on the many - body level the multi - boson attractive wave - packets remain localized objects in real space .
this observation , together with the fragmentation of the reduced one - body density matrix , suggest to look for signatures in the correlation functions , both in coordinate and momentum space , which is done in fig .
3 .
the first - order correlation function in coordinate space @xmath29 quantifies the degree of spatial coherence of the interacting system @xcite , and analogously @xmath30 in momentum space @xcite .
the upper two panels show the first - order correlation functions in coordinate ( left ) and momentum ( right ) space of the initially - prepared coherent @xmath8-phase symmetric soliton train at @xmath23 .
it should be stressed that , within gp theory , the wave - packet remains coherent at all times ; thus the upper two panels represent the first - order correlation functions of the two - hump attractive wave - packet at all times , assuming gp to be valid . in sharp contrast , the first - order correlation functions in coordinate and momentum space , when computed on the many - body level , show prominent structures , see middle and lower four panels in fig .
3 , which leave no place for confusion of the rise of fragmentation @xcite and the death of soliton trains .
another quantity of interest is the momentum distribution of the soliton train itself , shown in @xcite .
it is left to determine whether the death of soliton trains is on a time scale relevant for current experimental setups . for this
we consider explicitly @xmath1 @xmath31li atoms ( @xmath32 kg , @xmath33-wave scattering length is @xmath34 ) .
we emulate a quasi-1d cigar - shaped trap in which the transverse confinement is @xmath35 hz , which is amenable to current experimental setups . following @xcite ,
the transverse confinement renormalizes the interaction strength . combining all the above ,
the length scale is given by @xmath36 m , and the time scale by @xmath37 sec .
now , expressing the lifetime of the soliton trains ( see fig .
2 ) in real time , we get lifetimes of a couple of hundreds of milliseconds ( see @xcite for analysis of the lifetime as a fucntion of @xmath38 ) , much below current experimental run times of a couple of seconds @xcite .
our theoretical results thus call for exciting experimental demonstration of the ultrafast death of soliton trains in low - dimensional attractive becs , caused by fragmentation
. we are grateful to k. sakmann and h .- d .
meyer for fruitful discussions
. financial support by the dfg is greatly acknowledged .
64 s. trillo and w. torruellas ( eds . ) , _ spatial solitons _
( springer , berlin , 2001 ) . g. i. stegeman and m. segev , science * 286 * , 1518 ( 1999 ) .
p. a. ruprecht , m. j. holland , k. burnett , m. edwards , phys .
rev .
a * 51 * , 4704 ( 1995 ) .
v. m. prez - garca , h. michinel , and h. herrero , phys .
rev .
a * 57 * , 3837 ( 1998 ) .
k. e. strecker , g. b. partridge , a. g. truscott , and r. g. hulet , nature * 417 * , 150 ( 2002 ) .
l. khaykovich _
et al . _ ,
science * 296 * , 1290 ( 2002 ) .
u. al khawaja _
et al . _ ,
phys .
rev .
lett . * 89 * , 200404 ( 2002 ) .
l. salasnich , a. parola , and l. reatto , phys .
rev .
lett . * 91 * , 080405 ( 2003 ) . l. d. carr and j. brand , phys .
rev .
lett . * 92 * , 040401 ( 2004 ) .
a. d. martin , c. s. adams , and s. a. gardiner , phys .
rev .
lett . * 98 * , 020402 ( 2007 ) .
e. p. gross , nuovo cimento * 20 * , 454 ( 1961 ) .
l. p. pitaevskii , zh .
eksp .
. fiz . * 40 * , 646 ( 1961 ) [ sov . phys.jetp * 13 * , 451 ( 1961 ) ] .
p. nozires and d. saint james , j. phys .
france * 43 * , 1133 ( 1982 ) .
a. i. streltsov , o. e. alon , and l. s. cederbaum , phys .
rev .
a * 73 * , 063626 ( 2006 ) .
e. j. mueller , t .- l . ho , m. ueda , and g. baym , phys . rev . a * 74 * , 033612 ( 2006 ) . a. i. streltsov , o. e. alon , and l. s. cederbaum , phys . rev . lett . * 99 * , 030402 ( 2007 ) .
o. e. alon , a. i. streltsov , and l. s. cederbaum , phys .
rev .
a * 77 * , 033613 ( 2008 ) .
j. grond , j. schmiedmayer , and u. hohenester , phys .
rev .
a * 79 * , 021603(r ) ( 2009 ) .
k. sakmann , a. i. streltsov , o. e. alon , and l. s. cederbaum , phys .
rev .
lett .
* 103 * , 220601 ( 2009 ) .
see epaps document no .
xxxx for supplemental physical analysis of the death of soliton trains . for more information on epaps , see http://www.aip.org/pubservs/epaps.html .
a. i. streltsov , o. e. alon , and l. s. cederbaum , phys .
rev .
lett . * 100 * , 130401 ( 2008 ) .
m. naraschewski and r. j. glauber , phys .
rev .
a * 59 * , 4595 ( 1999 ) .
k. sakmann , a. i. streltsov , o. e. alon , and l. s. cederbaum , phys .
rev .
a * 78 * , 023615 ( 2008 ) .
m. olshanii , phys .
rev .
lett . * 81 * , 938 ( 1998 ) .
-1.5 truecm of the attractive system with @xmath1 bosons and @xmath39 computed by the popular gross - pitaevskii equation on the mean - field level in panels a)-c ) , and the corresponding dynamics computed on the many - body level in panels d)-f ) .
panels a ) and d ) show the dynamics of a @xmath8-phase symmetric train ; panels b ) and e ) show the dynamics of a @xmath9-phase anti - symmetric train ; panels c ) and f ) show the dynamics of a @xmath8-phase asymmetric train . on the many - body level the forces between the two density humps
are weaker , and there is essentially only repulsion and no attraction between them at all but short times .
all quantities shown are dimensionless.,title="fig:",width=453 ] -2.0 truecm -phase symmetric soliton train of fig . 1 in coordinate ( left ) and momentum ( right ) space .
it is completely flat signifying a coherent and condensed state . within gross - pitaevskii theory
the system remains condensed for all times ; thus the upper two panels represent the first - order correlation function in this case for all times . on the many - body level ,
the correlation functions are completely different and structured .
the pattern emerging in coordinate space indicates no coherence between the two humps ( left middle and lower panels ) , the respective pattern in momentum space represents a fragmented state ( right middle and lower panels ) .
thus , measuring the first - order correlation functions in either coordinate or momentum space provides a direct indication of the death of the soliton train .
all quantities shown are dimensionless.,title="fig:",width=453 ] |
a key parameter in the standard adiabatic cold - dark matter - based models of structure formation is the amplitude that fluctuations in the gravitational potential have when they enter the horizon .
this number , which we will denote @xmath3 , has been measured by the cobe satellite to be of order @xmath8 ( smoot 1992 ; bennett 1996 ) , and is assumed to be virtually independent of scale in the most popular models .
why @xmath8 ?
the answers proposed by theorists fall into two categories : 1 .
@xmath0 can be computed from first principles using some ( still undiscovered ) fundamental theory .
@xmath0 can not be computed from first principles , since the correct fundamental theory merely predicts an ensemble of superhorizon - sized spatial regions with a wide range of @xmath3 , forcing us to treat @xmath3 as random number subject to various anthropic selection effects .
the purpose of this paper is to consider such selection effects , by studying how the physical processes of structure formation depend on @xmath3 .
our motivation for this is threefold : 1 .
it affects which inflation / defect models should be considered natural as opposed to fine tuned .
2 . it is related to a crucial loophole in the recent arguments for an anthropic upper bound on @xmath4 .
it poses useful test problems for comparing cosmological simulations .
the structure of our universe is fixed by a rather small number of physical parameters .
the electron mass , the neutron mass and the low energy coupling constants of the four basic forces determine the physical properties of most objects on scales ranging from the atomic to the galactic ( see carr & rees 1979 ; davies 1982 ; barrow & tipler 1986 ) , and these parameters can in turn be computed from the roughly 20 free parameters of the standard model of particle physics . a number of additional parameters are often thought of as initial data laid down in the early universe : the baryon - to - photon ratio @xmath9 , the relative abundances of various dark - matter candidates , the vacuum density @xmath10 contributed by a cosmological constant @xmath4 , the spatial curvature ( related to @xmath5 ) and the amplitude @xmath3 of cosmological density fluctuations , although it is not implausible that abundances such as @xmath9 can ultimately be derived from other particle physics constants . together with the basic laws , these parameters determine when cosmic structures first emerge and how they evolve .
although the detailed outcome in any one locality , and what complex systems evolve there , depends on local accidents , these parameters nonetheless determine the statistical properties .
will it ever be possible to compute the values of all these parameters from first principles , within the framework of some yet to be discovered fundamental theory ?
the answer is a resounding no within some variants of inflationary cosmology ( , linde 1983 , 1987 , 1990 , 1995 ; linde & zelnikov 1988 ; coleman 1988 ; albrecht 1994 ; vilenkin 1995abcd ; vilenkin & winitzki 1997 ) , where the spatial region that we conventionally call `` our universe '' , itself perhaps extending far beyond the present observational horizon , is just one element in an ensemble whose members have widely disparate properties .
some physical parameters may take a range of different values throughout this ensemble of exponentially large and causally disconnected regions .
the predictions of such theories therefore take the form of probability distributions for the parameters in question , and these must be computed in bayesian fashion taking into account the selection effect that observers are not equally likely to inhabit all parts of the ensemble . for instance , just as we expect low surface brightness galaxies to be underrepresented in many surveys , we might expect o - stars to be underrepresented in solar systems containing planet - based extraterrestrial civilizations and , as we shall see , spacetime regions with @xmath11 to be underrepresented in the set of regions that contain observers .
the importance of such anthropic selection effects was stressed by carter ( 1974 ) , and is discussed in great detail in books by , , davies ( 1982 ) and barrow & tipler ( 1986 ) .
more recent reviews can be found in , , balashov ( 1991 ) and tegmark ( 1997 ) .
many inflationary models predict an ensemble of exponentially large space - time regions , each with a different value of @xmath3 ( see linde 1990 ; vilenkin 1995abcd and references therein ) .
although the cosmological literature abounds with remarks on the `` unnaturally '' flat potential required to produce @xmath0 in our own hubble volume , often as a motivation to study defect models , one can just as well argue that it is unnatural that the potential is not even flatter , since superflat potentials make inflation last longer and hence dominate the ensemble by volume ( vilenkin 1995a ) .
this dispute can not be resolved without taking the inevitable anthropic selection effects into account : if these turn out to place a firm _ upper _ limit on @xmath3 near the observed value , then inflation models predicting ensembles peaked at _ high _ @xmath3 clearly require no fine tuning to explain why we observe @xmath0 .
conversely , if these selection effects give a firm lower limit on @xmath3 near @xmath8 , then inflation models predicting ensembles peaked at low @xmath3 require no fine tuning .
another hotly debated parameter is @xmath4 , the cosmological constant .
although one might expect the most `` natural '' value of the vacuum density @xmath10 to be of order the planck density , the observational upper limits on @xmath12 are a striking factor of @xmath13 smaller .
this has led to fine tuning criticism of cosmological models with @xmath14 , the argument being that they were ruled out at high confidence , since such a small value of @xmath4 was extremely unlikely ( see dolgov 1997 for an up - to - date review ) . as was pointed out by barrow & tipler ( 1986 ) , weinberg ( 1987 , 1989 ) and efstathiou ( 1995 )
, there is a flaw in this argument , since it neglects a powerful anthropic selection effect .
if @xmath4 is too large , then the universe becomes vacuum dominated before the density fluctuations have grown enough to form non - linear structures . hence the fluctuations stop growing , and neither galaxies nor observers will ever form .
it is therefore no surprise that we find ourselves in a region where @xmath4 is small .
a calculation of the probability distribution for @xmath10 given our existence shows that values of order of the current limits are in fact rather typical ( efstathiou 1995 ) , and more accurate calculations ( weinberg 1996 ; martel 1997 ) have confirmed this conclusion .
unfortunately , there is a loophole in this argument ( rees 1997 ) . as described in more detail in section 5 ,
increasing @xmath4 by some factor @xmath15 can be completely offset by increasing @xmath3 by a factor @xmath16 as far as this argument is concerned . whether this is really
a loophole thus depends crucially on the topic of the present paper , specifically on whether observers could exist if @xmath17 . the analogous potential loophole exists for anthropic lower bounds on @xmath5 ( barrow 1982 ; vilenkin & winitzki 1997 ) .
a third and entirely different motivation for exploring counterfactual values of parameters such as @xmath3 is that it provides a challenging and bias - free test of cosmological simulation techniques .
state - of the art simulations including hydrodynamics ( which breaks the degeneracy between @xmath3 and @xmath18 in pure gravity simulations ) , gas chemistry and star formation often achieve a good fit to our actual universe ( see kang 1994 and references therein ) , but only after tweaking a number of parameters empirically . it is therefore unclear to what extent the agreement between different groups is due to realistic modeling as opposed to simply living in ( and parameter - fitting to ) the same universe
. it would be far more convincing if two groups could obtain indistinguishable results for hypothetical universes with other values of @xmath3 , where the answer would not be known beforehand .
in section 2 , we outline how @xmath3 affects structure formation in a universe with @xmath19 and @xmath20 .
we discuss the effects of lowering and raising @xmath3 in sections 3 and 4 , respectively , and the effects of changing @xmath5 and @xmath4 in section 5 .
we will find it convenient to work in planck units where @xmath21 , and the fundamental units of of length , time , mass and temperature are @xmath22 m , @xmath23s , @xmath24 kg and @xmath25k , respectively .
important dimensionless constants that will recur frequently are the electromagnetic coupling constant @xmath26 , the gravitational coupling constant @xmath27 , the electron - proton mass - ratio @xmath28 , the baryon - to - photon ratio @xmath29 , the baryon fraction @xmath30 of the nonrelativistic matter density ( which we take to equal the critical value that gives a spatially flat universe ) and the matter - to - photon ratio = _
g^1/2_b ~10 ^ -27 .
this constant @xmath31 is simply the amount of nonrelativistic matter per photon , @xmath32 , measured in planck masses .
as our goal is to highlight the main physical effects rather than to make detailed numerical calculations , we will frequently use the symbol @xmath33 , which we take to mean that numerical factors of other unity ( @xmath34 and the like ) have been omitted .
for instance , the hydrogen binding energy ( 1 rydberg ) , the bohr radius and the thomson cross section are given by @xmath35 , @xmath36 and @xmath37 , respectively . the reader may find it unfamiliar to see almost no reference below to familiar quantities such as the redshift @xmath38 , the current cmb temperature @xmath39 , the current hubble constant @xmath40 and the current density parameter @xmath41 .
this is because we strive to highlight how structure formation depends on fundamental parameters , and these quantities are not fundamental since they have meaning only once the epoch at which we happen to be living has been specified .
indeed , for the the open universe case , @xmath42 , @xmath40 and @xmath41 can be thought of as merely alternative time variables , since they all decrease monotonically with @xmath18 .
for instance , we are not interested in examining what @xmath3-values allow galaxies to form by the present epoch @xmath43 years , but what @xmath3-values allow them to form at all . the rising curves in show when different mass scales go nonlinear , defined as the time when linear perturbation theory predicts an overdensity of 1.69 in a top hat sphere containing the mass @xmath44 ( press & schechter 1974 ) .
the curves were computed for the cold dark matter ( cdm ) power spectrum fit of bond & efstathiou ( 1984 ) with @xmath45 , `` shape parameter '' @xmath46 , and an @xmath47 normalization @xmath48 .
we assume a standard spatially flat universe ( @xmath19 , @xmath20 ) everywhere in this paper , thus requiring more than a single number for their parametrization . ] except in .
since fluctuations can not grow before the matter - radiation equality epoch , the radiation energy per proton , @xmath49 , equals the dark matter energy per proton , @xmath50 , so @xmath51 . since the energy density is @xmath52 , the friedman equation gives the hubble expansion rate @xmath53 , and so the age of universe at this time is @xmath54 . ]
@xmath55s ( the vertical line in the figure ) , all scales below the horizon mass at this epoch , m_eq~_eq
t_eq^3~^-2 ~_g^-2~10 ^ 16 , have similar fluctuation levels , and are seen to virialize roughly simultaneously ( up to a logarithmic factor ) , at ~t_eq q^-3/2~^-2q^-3/2 .
( the origin of the `` @xmath56 '' is that , during the matter - dominated epoch , fluctuations grow as the scale factor @xmath57 and @xmath58 . )
since the figure shows that the actual curves approach vertical only for very small mass scales , we have included a factor @xmath59 in which depends weakly on mass .
@xmath60 for @xmath61 , with the value for typical galactic scales @xmath62 being @xmath63 .
far above this mass scale , @xmath64 ( we assume the standard spectral index @xmath65 ) , which means that @xmath66 , where the horizon mass is @xmath67 ( straight solid line ) .
thus the curves all have the same shape , and their left and right asymptotes lie about a factor @xmath68 to the right of the two heavy straight lines in the figure , giving the following broad - brush picture .
mass scales @xmath69 virialize roughly simultaneously , at @xmath70 .
as time progresses , ever larger scales keep virializing , the non - linear mass scale always being a fraction @xmath71 of the horizon mass scale ( a fraction @xmath72 in radius ) .
note that the the number @xmath73 occurring in this crucial mass @xmath74 is simply @xmath75 the well - known result that a stellar mass @xmath76 ( dyson 1971 ) was used in .
when an overdensity has collapsed , the resulting virial halo will have a typical density that exceeds the background density by a collapse factor @xmath77 , , _ vir~_eq(t_virt_eq)^-2~^4 ^-2q^3 . for a cdm halo of mass @xmath44
, this corresponds to a characteristic size @xmath78 , velocity @xmath79 and virial temperature t_vir~m_p v_vir^2 ~_g^1/2 ^4/3 ^-2/3 ^ 1/3 m^2/3 q. a number of isotherms are plotted in , and we see that as time progresses and ever larger halos form , the virial temperature stops increasing around the characteristic time @xmath80 and approaches a maximum value @xmath81 , corresponding to a maximum virial velocity @xmath82 .
thus for our @xmath0 universe , typical cluster temperatures are @xmath83 , about @xmath8 times the proton rest energy , and characteristic cluster velocity dispersions are @xmath84 , about @xmath85 times the speed of light .
this direct link between @xmath3 and halo temperatures immediately indicates why lowering @xmath3 can cause qualitatively different structure formation scenarios .
unless @xmath87 exceeds typical atomic energy scales @xmath88 , which corresponds to q ~ m_e c^2 ^ 2m_p c^2 = ^2~10 ^ -8 , it will be difficult for the gas in these halos to dissipate their energy to collapse and form stars .
hydrogen line cooling freezes out at about @xmath89 , for instance , corresponding to @xmath90 .
we will now discuss cooling constraints in more detail , and see that these cause qualitative changes even for much smaller departures from @xmath0 . the fate of the baryons in a virialized halo depends crucially on the ratio of the cooling timescale @xmath91 to the gravitational collapse timescale @xmath92 ( see binney 1977 , rees & ostriker 1977 ; silk 1977 ; white and rees 1978 ) . if @xmath44 and @xmath18 are such that @xmath93 ( the dark - shaded region in ) , the cloud can not promptly commence free - fall collapse and fragment into stars , but will remain pressure supported for at least a local hubble time . for the halo formation curve corresponding to @xmath0 ,
the part of the @xmath94 curve setting the upper limit on galaxy mass is seen to have a logarithmic slope around @xmath95 ( because bremsstrahlung , with @xmath96 , is the dominant cooling process ) , corresponding to @xmath97 and a constant radius @xmath98 ( carr & rees 1979 ) .
the corresponding mass scale is seen to be @xmath62 . for slightly lower
@xmath3 , the upper limit is dominated by line cooling in neutral hydrogen ( rightmost bump ) , helium ( second bump ) and any heavier elements released by early stars ( not included here ) . the lower mass limit is set by the @xmath99 isotherm , below which there are essentially no free electrons and both line cooling and bremsstrahlung become ineffective .
molecular cooling can potentially lower this mass limit slightly ( haiman 1996 ; abel 1997 ; gnedin & ostriker 1997 ; tegmark 1997 ) , but is ignored in the figure for the same reason as heavy elements : it is irrelevant to our @xmath3-constraints , which depend only how far the cooling region extents to the right , not on the vertical extent . what happens if we start lowering @xmath3 ?
the first change is that the upper limit becomes set not by bremsstrahlung but by line cooling .
indicates that as we keep lowering @xmath3 , the range of galactic masses narrows down and finally vanishes completely for @xmath100 .
let us express this critical value @xmath101 in fundamental constants .
the figure shows that it is determined by the `` hydrogen bump '' in the cooling function , which is caused by free electrons collisionally exciting neutral hydrogen atoms into their first excited state , which is immediately followed by emission of a ly@xmath102 photon .
this gives a cooling timescale ( , dalgarno & mccray 1972 ) t_cool~(m_e^2 c^2 ) ^-3/2 e^34x(1-x)n , where @xmath103 , @xmath104 is the total ( bound and free ) proton number density , and @xmath105 is the ionization fraction . in thermal equilibrium
, this is given by ( tegmark , silk & evrard 1993 ) x~[1+^3 ^ 7/6e^]^-1 .
substituting into gives t_cool~(m_e^2 c^2 ^ 2 n ) , where the dimensionless quantity in square brackets is minimized for @xmath106\sim 10 $ ] , corresponding to @xmath107 .
this minimum value is @xmath108^{-8/3}\sim 1/5,000 $ ] .
equating this minimal cooling timescale with @xmath109 using @xmath110 finally tells us that the latest time at which line cooling can be efficient is t_max ~^3/2 ^ 8/3 _
g^-3/2 ^ -2_b~10 ^ 19 . equating this with @xmath111 from thus
tells us that efficient cooling occurs when q^-1 ^ -16/9 _
g^4/3 ^ -4/3 ^ 2/3_b^-2/3 ~10 ^ -6 .
if @xmath112 , then what is the ultimate fate of the quasistatic pressure supported gas clouds ? it is plausible that they will become increasingly rarefied as their dark matter halos eventually merge into larger ( and less dense ) halos , thereby never entering a phase of runaway cooling , fragmentation and star formation .
however , even in the arguably contrived case where such a cloud escaped any further collisions , and eventually managed to cool after a ( perhaps exponentially ) long time , perhaps through some exotic mechanism such as 21 cm cooling , and developed a dense , self - gravitating core which fragmented into stars , there would still be reason to doubt whether it could produce intelligent observers .
since the binding energy of the halo is so low ( of order @xmath113 ) , the first supernova explosion might well eject all the gas from the halo , thereby precluding the production of population ii stars and planets containing heavy elements .
what happens if we start increasing @xmath3 instead ?
the allowed mass range for galaxies keeps broadening at a steady rate until compton cooling suddenly eliminates the upper mass limit altogether .
this is because the time scale on which cosmic microwave background ( cmb ) photons at temperature @xmath114 cool an ionized plasma , _
comp~^3 c^4 m_e_t(k t_)^4 x , is independent of both its density and temperature ( assuming that @xmath115 , so that @xmath116 ) .
since @xmath117 , this timescale @xmath118 equals the age of the universe @xmath18 at a characteristic time t_comp~^6/5_g^-13/10 ^ -9/5 ^ -4/5~10 ^ 16 . setting @xmath119
, we find that the upper limit to galaxy masses persists only for q^-4/5_g^3/5 ^ 6/5 ^ -4/5 ^ 2/3~10 ^ -4.5 . for larger @xmath3-values , all mass scales can cool efficiently , so the characteristic mass for the first generation of galaxies will simply be @xmath120 , given by .
this corresponds to a characteristic size @xmath121 for newly formed galaxies , which is constant in _ comoving _ coordinates ( rather than in absolute coordinates , as the above - mentioned cooling scale @xmath122 ) .
it would plainly need detailed simulations to determine the mix of discs and spheroids , and the effects of subsequent mergers .
however , the galaxies could well have a broader luminosity function than in our actual universe ( as well as a much higher characteristic mass ) ; and clustering would also extend up to a larger fraction of the hubble radius .
would this qualitative change affect the number of habitable planets produced ?
let us first consider the stability of planetary orbits .
lightman ( 1984 ) has shown that if the planetary surface temperature is to be compatible with organic life , the orbit around the central star should be fairly circular and have a radius of order r_au~^-5_g^-3/4 ^ -2~10 ^ 11 , roughly our terrestrial `` astronomical unit '' , precessing one radian in its orbit on a timescale
t_orb~^-15/2_g^-5/8 ^ -3~0.1 .
an encounter with another star with impact parameter @xmath123 has the potential to throw the planet into a highly eccentric orbit or even unbind it from its parent star .
this happens on a timescale @xmath124 , where @xmath125 and @xmath126 denote the typical stellar density and stellar velocity in a galaxy , respectively .
writing @xmath127 , where @xmath128 and @xmath129 is the additional factor by which the dissipating baryons collapse relative to the dark matter before fragmenting into stars , the milky way is empirically fit by @xmath130 . for earth , this gives @xmath131 , orders of magnitude above its present age .
moreover , the distant encounters that we have experienced in the past have had a completely negligible effect since they were adiabatic .
this means that the impact duration @xmath132 , so that the solar system returned to its unperturbed state once the encounter was over . for hypothetical galaxies forming before @xmath133 , on the other hand , @xmath61 ,
so the time between non - adiabatic encounters is _
adiab~1v^3 t_orb^2~^15_g^1/4 ^ 6 ^ 3 ^ 4 q^9/2 ^ 3/2 ~10 ^ 5 for @xmath134 and dropping as @xmath135 if we increase @xmath3 further . in other words ,
non - adiabatic encounters are frequent events for @xmath2 , occurring often during the geological timescales required for a planet to form , cool and ultimately evolve life . in the conservative approximation of ignoring gravitational focusing ( assuming that the flyby speed @xmath136 exceeds the orbital speed ) , the typical time interval between @xmath137 encounters is _
r_au^2 ~^10_g^1/2 ^ 4 ^ 7/3 ^ 4 ^ 7/6q^7/2 ~10 ^ 7 for @xmath138 . requiring this to exceed some geological or evolutionary timescale @xmath139 thus gives an upper limit @xmath140 .
although it is far from clear what is an appropriate @xmath139 to use , the smallness of the exponent @xmath141 implies that it makes only a minimal difference whether we choose @xmath142 or @xmath143 years . taking @xmath144 years @xmath145 , the lifetime of a bright star ( carr & rees 1979 ) , we obtain the limit q ^16/7_g^4/7 ^ 12/7 ^ -8/7 ^ 2/3 ^ -1/3 ^ -2/7 ~10 ^ -4 .
this upper limit appears more uncertain than the lower limit from cooling .
the momentum kick given to the planet scales as @xmath146 , so an impact with @xmath147 would not necessarily cause a catastrophic disturbance of the planetary orbit the event rate for this grows only as @xmath148 , or as @xmath149 if the galactic stars settle into a disk where @xmath136 is roughly independent of @xmath3 . on the other hand , a very close encounter ( especially with an o - star ) might cause disastrous heating of the planet . in view of this uncertainty , as well as the uncertainty regarding @xmath129 and @xmath150 , we now consider two additional effects of raising @xmath3 . for much greater @xmath3-values , of order unity , typical fluctuations
would be of black - hole magnitude already by the time they entered the horizon , converting some substantial fraction @xmath15 of the radiation energy into black holes already shortly after @xmath151 , the end of the inflationary era . at @xmath152 ,
the universe has expanded by a factor @xmath153 , and the energy densities in black holes and photons have dropped by factors of @xmath154 and @xmath155 , respectively .
the black hole density will therefore completely dwarf the density of cold dark matter and baryons if @xmath156 .
thus even if @xmath157 , extremely rare fluctuations that are @xmath158 standard deviations out in the gaussian tail can cause black hole domination if @xmath159 , where @xmath160du$ ] .
this gives the upper limit q^-1 [ ] ~^-1[10 ^ -27]~10 ^ -1 , where we have simply assumed that @xmath151 is within a few orders of magnitude of unity ( the planck time ) since this affects the result only logarithmically . as opposed to the previous constraints , this one depends strongly on whether the power spectrum is strictly scale invariant or not increasing the spectral index @xmath104 from its scale - invariant value @xmath65 to @xmath161 causes primordial black hole domination even if @xmath3 is as low as @xmath8 ( green 1997 ) .
even if @xmath3 were low enough to avoid black hole formation in the early radiation dominated phase ( say in the range @xmath162@xmath163 ) , rampant black - hole formation may still occur in the matter - dominated era . at times of order @xmath142 years , , shortly after recombination , clumps of order @xmath74 will collapse . if dissipation leads to enough reionization to make their thomson optical depth larger than @xmath164 ( itself of order @xmath165 ) , then they will trap the background radiation and collapse like supermassive stars , without being able to fragment .
the dominant structures in such a universe would then be supermassive black holes , and it is unclear whether any galaxies and stars would be able to form . even if they could , they would be hurtling around at speeds of order a tenth of the speed of light , and it is far from clear how anthropically favorable such an environment would be !
our discussion above applied to a flat frw universe with @xmath19 and @xmath20 .
as we will now describe , anthropic limits on these two parameters are intimately linked with @xmath3 . in planck units , the friedmann equation that governs the time evolution of the radius of curvature of the universe , @xmath57 ,
is conveniently written as ( aa)^2 = 83(_+_m+_c+ _ ) , where @xmath166 , @xmath167 and @xmath10 are the energy densities corresponding to radiation , nonrelativistic matter and vacuum energy ( a cosmological constant ) , respectively .
@xmath168 is the contribution from spatial curvature ( the sign is positive if @xmath169 and negative if @xmath170 for the flat case @xmath19 , the radius of curvature is infinite and @xmath57 must be redefined ) .
the first three of these densities evolve as _ & ~&_^-2a^-4 , + _ m&~&_^-3/2a^-3 , + _ c&~&a^-2 , and @xmath10 does not evolve at all .
the constant @xmath171 is defined as the curvature that the universe _ would _ have had at the planck time if there was no inflationary epoch , and can be evaluated at any time in the post - inflationary radiation - dominated epoch as @xmath172 , during which this quantity is time independent .
we have introduced @xmath171 simply because we need a constant that quantifies the curvature , and the more familiar @xmath5 is unusable since it changes with time .
the epochs of matter domination @xmath173 , curvature domination @xmath174 and vacuum domination @xmath175 are given by @xmath176 , @xmath177 and @xmath178 , respectively , , a_md&~&^-1_^-1/2 , + a_cd&~&_^-3/2 , + a_vd&~&^1/3_^-1/2_^-1/3 .
it is well - known that sub - horizon fluctuations can only grow during the matter - dominated epoch , where they grow at the same rate as the scale factor @xmath57 . as we saw in
, the first non - linear structures therefore form at @xmath179 providing that the universe remains matter - dominated until this epoch ( @xmath180 and @xmath181 ) otherwise no nonlinear structures will ever form .
we thus obtain the two anthropic constraints _ & & ^2 q~10 ^ -59 , + [ lambdalimiteq ] _ & & ^4 q^3~10 ^ -124 . although we tacitly assumed that @xmath169 here , the closed case gives essentially the same constraints indeed , if no non - linear structures have formed at the epoch @xmath174 in a closed universe , time is literally running out for not yet evolved life forms , since the big crunch is imminent ! in comparison , the current observational limits are ( very conservatively ) @xmath182 and @xmath183 , which corresponds to @xmath184 and @xmath185 .
the conclusion is that although the anthropic upper limits superficially appear quite strong on both curvature and vacuum density , these constraints are only strong if the two variables on the right - hand side ( @xmath31 and @xmath3 ) are independently constrained which was one of our motivations for studying upper limits on @xmath3 .
the parameter @xmath31 probably deserves more attention than it has received in this context so far ( rees 1979 ) , and the effects of varying the baryon / photon ratio and introducing a non - zero neutrino mass would also warrant further study .
we note in passing that we can obtain crude @xmath3-independent limits on @xmath31 by requiring that our lower limits on @xmath3 not exceed our upper limits .
for instance , the virialization epoch of will occur too late for cooling to be efficient ( after @xmath186 of ) unless @xmath187 .
thus the white region in the figure disappears completely if @xmath188 , and the conservative limit @xmath189 gives the ( rather weak ) constraint @xmath190 .
conversely , the planetary disruption constraint of gets stronger if we _ increase _
@xmath31 , and conflicts with the @xmath31-independent limit of unless @xmath191 .
in addition , there are of course separate limits on the baryon fraction @xmath192 , in that if there are too few baryons , the cooling becomes less efficient see . lowering @xmath192
may also impede galaxy and star formation , since a gas cloud must collapse by a larger factor before it becomes self - gravitating . for the reader
preferring to think in terms of @xmath41 and redshift @xmath38 , the above argument can be re - expressed as follows .
if the current matter density is @xmath167 , then vacuum domination occurs at the epoch @xmath193 . if @xmath194 , then the universe became curvature dominated at a redshift given by @xmath195 . since the first structures form at an epoch @xmath196 ,
the upper limits on @xmath4 and @xmath197 thus scale as @xmath198 and @xmath199 for the @xmath194 case .
for instance , maintaining spatial flatness but making @xmath4 a million times larger than the current observational limits could correspond to @xmath200 , with galaxy formation about ten expansion times after recombination .
when the universe had reached its current age of @xmath201 years , it would have expanded by a further factor @xmath202 , and ours would be the only galaxy in the local hubble volume alas , a drab and dreary place for extragalactic astronomers , but not ruled out by the above - mentioned @xmath4-arguments alone although perhaps by the @xmath3-arguments that we have presented .
we have explored counterfactual cosmological scenarios with @xmath3 shifted away from its observed value @xmath203 .
we found that qualitative changes occur if we either increase or decrease @xmath3 by about an order of magnitude .
if @xmath1 , efficient cooling becomes impossible for gas in virialized halos .
if @xmath2 , compton scattering against cmb photons enables efficient cooling in arbitrarily massive halos , and the higher stellar densities and velocities may lead to planetary orbits being disrupted before observers have had time to evolve .
needless to say , this does not preclude that some form of life might evolve in a universe with a more extreme @xmath3-value due to lucky circumstances , for instance around a field star that was ejected from its giant host galaxy in a @xmath204 scenario .
however , as stressed by vilenkin ( 1995a ) , the key feature of anthropic selection effects is not what the rock - solid extreme limits are on a parameter , but which is _ the most favorable value _ for producing observers .
this point is also emphasized by garca - bellido & linde ( 1995 ) . to predict a probability distribution for the observed value of @xmath3 from some inflationary model ( to potentially rule the model out ) , its a priori probability distribution for @xmath3 ( of quantum origin , say ) must be multiplied by some bayesian selection function such as the number of observers or civilizations corresponding to each @xmath3-value .
it seems plausible that much more stars with habitable planets are formed for @xmath0 ( where perhaps @xmath205 of all baryons are in stars ) than in a @xmath206 universe where 1000 times lower densities make cooling difficult .
likewise , it appears likely that @xmath207 gives fewer planets in favorable stable orbits than @xmath0 , where close encounters are completely negligible for most stars . in conclusion , it is possible that the anthropic selection function peaks at @xmath0 .
if this is the case , then what vilenkin terms `` the principle of mediocrity '' would imply that since we are most likely to be a typical civilization , this is what we should expect to observe .
the authors wish to thank tom abel , andreas albrecht , ted bunn , wayne hu , robert kirshner , andrei linde , avi loeb , scott tremaine , alexander vilenkin and ned wright for helpful suggestions .
this work was supported by nasa grant nag5 - 6034 and hubble fellowship # hf-01084.01 - 96a , awarded by the space telescope science institute , which is operated by aura , inc . under nasa contract nas5 - 26555 .
dyson f j;1971;sci .
am;225;50 efstathiou g;1995;mnras;274;l73 garca - bellido jlinde a d;1995;phys .
d;251;429 gnedin n yostriker j p;1997;apj;486;581 green a m , liddle a r riotto a;1997;astro - ph/9705166 haiman z , thoul a a loeb a;1996;apj;464;523 linde a d;1983;phys . lett.;129b;177
linde a d;1987;physics today;40;61 linde a d;1988;phys .
b;201;437 linde a dzelnikov m i;1988;phys .
b;215;59 linde a d;1990;particle physics and inflationary cosmology;harwood;chur , switzerland tegmark m , silk jevrard a;1993;apj;417;54 tegmark m , silk j , rees m j , blanchard a , abel tpalla f;1997;apj;474;1 tegmark m;1997;gr - qc/9704009 vilenkin a;1995a;phys .
lett.;74;846 vilenkin a;1995b;phys | we explore the qualitative changes that would occur if the amplitude @xmath0 of cosmological density fluctuations were different . if @xmath1 , the cosmological objects that form would have so low virial temperatures that they may be unable to cool and form stars , and would be so loosely bound that even if they could produce a supernova explosion , they might be unable to retain the heavy elements necessary for planetary life .
if @xmath2 , dense supermassive galaxies would form , and biological evolution could be marred by short disruption timescales for planetary orbits .
if @xmath3 were still larger , most bound systems would collapse directly to supermassive black holes .
these constraints on @xmath3 can be expressed in terms of fundamental constants alone , and depend only on the electromagnetic and gravitational coupling constants , the electron - proton mass ratio and the matter - to - photon ratio .
we discuss the implications for inflation and defect models , and note that the recent anthropic upper bounds on the cosmological constant @xmath4 would be invalid if both @xmath3 and @xmath4 could vary and there were no anthropic constraints on @xmath3 .
the same applies to anthropic bounds on the curvature parameter @xmath5 .
# 1 # 2#1 , # 2 .
# 1 # 2 # 3#1 , # 2 .
# 3 .
# 1 # 2 # 3 # 4#1 , # 2 .
# 3 .
# 4 .
# 1;#2;#3 # 1 # 2 , preprint # 3 # 1;#2;#3;#4;#5 # 1 # 2 , _ # 3 _ , * # 4 * , # 5 # 1;#2;#3;#4;#5 # 1 # 2 , _ # 3 _ ( # 4 : # 5 ) # 1;#2;#3;#4;#5;#6 # 1 # 2 , in _ # 3 _ , ed . # 4 ( # 5 : # 6 ) # 1#1 # 1@xmath6 # 1@xmath7 # 1equation ( [ # 1 ] ) # 1equation ( [ # 1 ] ) # 1 ( [ # 1 ] ) # 1figure 1 # 1figure 1 # 1section [ # 1 ] # 1section [ # 1 ] # 1to 0pt#1 # 110^#1 # 1#2#1 |
the cosmological constant ( cc ) problem @xcite-@xcite , the accelerated expansion of the late time universe@xcite , the cosmic coincidence @xcite are challenges for the foundations of modern physics ( see also reviews on dark energy@xcite-@xcite , dark matter @xcite and references therein ) .
numerous models have been proposed with the aim to find answer to these puzzles , for example : the quintessence@xcite , coupled quintessence@xcite , @xmath3-essence@xcite , variable mass particles@xcite , interacting quintessence@xcite , chaplygin gas@xcite , phantom field@xcite , tachyon matter cosmology@xcite , brane world scenarios@xcite , etc .. these puzzles have also motivated an interest in modifications and even radical revisions of the standard gravitational theory ( general relativity ( gr))@xcite,@xcite . although motivations for most of these models can be found in fundamental theories like for example in brane world@xcite , the questions concerning the einstein gr limit and relation to the regular particle physics , like standard model , still remain unclear .
it is very hard to imagine that it is possible to propose ideas which are able to solve the above mentioned problems keeping at the same time unchanged the basis of fundamental physics , i.e. gravity and particle field theory .
this paper may be regarded as _ an attempt to find a way for satisfactory answers at least to a part of the fundamental questions on the basis of first principles _ ,
i.e. without using semi - empirical models . in this paper
we explore a toy model including gravity and a single scalar field @xmath0 in the framework of the so called two measures field theory ( tmt)@xcite-@xcite . in tmt , gravity ( or more exactly , geometry ) and particle field theory intertwined in a very non trivial manner , but the einstein s gr is restored when the fermion matter energy density is much larger than the vacuum energy density@xcite,@xcite . here
we _ have no purpose _ of constructing a complete realistic cosmological scenario .
instead , we concentrate our attention on the possible role tmt may play in resolution of a number of the above mentioned problems . the field theory model we will study here
is invariant under global scale transformation of the metric accompanied with an appropriate shift of the dilaton field @xmath0 .
this scale symmetry is spontaneously broken due to intrinsic features of the tmt dynamics . except for a peculiar structure of the tmt action
, the latter does not contain any exotic terms .
the obtained dynamics represents an explicit example of @xmath3-essence resulting from first principles .
the parameter space of the model allows to find different regions where the following different effects can take place _ without fine tuning of the parameters and initial conditions _ :
\a ) possibility of a power law inflation driven by a scalar field @xmath0 which is followed by the late time evolution driven both by a small cosmological constant and a scalar field @xmath0 with a quintessence - like potential ; smallness of the cosmological constant can be achieved _ without fine tuning of dimensionfull parameters_. \b ) in another region of the parameters , there is a possibility of the power law inflation ended with damped oscillations of @xmath0 around the state with zero cosmological constant ( we want to emphasize again that this is realized without fine tuning of the parameters and initial conditions ) .
thus this scenario includes _ an exit from inflation together with resolution of the old cc problem_. this becomes possible because the basic assumptions of the weinberg s no - go theorem@xcite are generically violated in tmt models .
\c ) there is a region of the parameters where the model exhibits the possibility for a superacceleration phase of the universe : in the late time universe the dark energy density @xmath9 increases asymptotically ( as @xmath7 ) approaching from below to a constant value ; the equation - of - state @xmath10 and @xmath6 _ asymptotically approaches _ @xmath8 _ from below_. the organization of the paper is the following : in sec.ii we present a review of the basic ideas of tmt .
sec.iii starts from formulation of a simple scale invariant model containing all the terms respecting the symmetry of the model but without any exotic terms . in appendix
a we present equations of motion in the original frame . using results of appendix a , the complete set of equations of motion in the einstein frame is given in sec.iii as well .
it is shown there that if no fine tuning of the parameters is made , the effective action of our model in the einstein frame turns out to be a k - essence type action .
we start in sec.iv from a simple case with fine tuned parameters where the non - linear dependence of the kinetic term disappears
. then three different shapes of the effective potential are possible .
for the spatially flat frw universe we study some features of the cosmological dynamics for each of the shapes of the effective potential . for one of the shapes of the effective potential ,
the zero vacuum energy is achieved without any fine tuning . in sec.v
we study the cosmological dynamics without tuning parameters . it is shown that there is the possibility of a superacceleration phase of the universe , and some details of the dynamics are explored .
sec.vi is devoted to studying in more details the cc problems : a ) we discuss two mechanisms for resolution of the new cc problem ; b ) we analyze the difference between tmt and conventional field theories ( where only the measure of integration @xmath4 is used in the fundamental action ) which allows to understand why from the point of view of tmt the conventional field theories failed to solve the old cc problem . in appendix
b we shortly discuss what kind of model one would obtain when choosing fine tuned couplings to the measures of integration in the action . some additional remarks concerning the relation between the structure of tmt action and the cc problem are given in appendix c.
tmt is a generally coordinate invariant theory where _ all the difference from the standard field theory in curved space - time consists only of the following three additional assumptions _ :
the main supposition is that for describing the effective action for gravity @xmath11 matter at energies below the planck scale , the usual form of the action @xmath12 is not complete .
our positive hypothesis is that the effective action has to be of the form@xcite-@xcite @xmath13 including two lagrangians @xmath14 and @xmath15 and two measures of integration @xmath4 and @xmath16 or , equivalently , two volume elements @xmath17 and @xmath18 respectively .
one is the usual measure of integration @xmath4 in the 4-dimensional space - time manifold equipped with the metric @xmath19 .
another is the new measure of integration @xmath16 in the same 4-dimensional space - time manifold .
the measure @xmath16 being a scalar density and a total derivative in string and brane theories see ref.@xcite . ] may be defined * either by means of four scalar fields @xmath20 ( @xmath21 ) , ( compare with the approach by wilczek ) , @xmath22 * or by means of a totally antisymmetric three index field @xmath23 @xmath24 to provide parity conservation in the case given by eq.([phi ] ) one can choose for example one of @xmath20 s to be a pseudoscalar ; in the case given by eq.([aabg ] ) we must choose @xmath23 to have negative parity .
some ideas concerning the nature of the measure fields @xmath20 are discussed in ref.@xcite .
a special case of the structure ( [ s ] ) with definition ( [ aabg ] ) has been recently discussed in ref.@xcite in applications to supersymmetric theory and the cc problem .
it is assumed that the lagrangian densities @xmath14 and @xmath15 are functions of all matter fields , the dilaton field , the metric , the connection but not of the `` measure fields '' ( @xmath20 or @xmath23 ) .
in such a case , i.e. when the measure fields enter in the theory only via the measure @xmath16 , the action ( [ s ] ) possesses an infinite dimensional symmetry . in the case given by eq.([phi ] ) these symmetry transformations have the form @xmath25 , where @xmath26 are arbitrary functions of @xmath27 ( see details in ref.@xcite ) ; in the case given by eq.([aabg ] ) they read @xmath28 where @xmath29 are four arbitrary functions of @xmath27 and @xmath30 is numerically the same as @xmath31
. one can hope that this symmetry should prevent emergence of a measure fields dependence in @xmath14 and @xmath15 after quantum effects are taken into account .
important feature of tmt that is responsible for many interesting and desirable results of the field theory models studied so far@xcite-@xcite consists of the assumption that all fields , including also metric , connection and the _ measure fields _ ( @xmath20 or @xmath23 ) are independent dynamical variables .
all the relations between them are results of equations of motion .
in particular , the independence of the metric and the connection means that we proceed in the first order formalism and the relation between connection and metric is not necessarily according to riemannian geometry .
we want to stress again that except for the listed three assumptions we do not make any changes as compared with principles of the standard field theory in curved space - time . in other words , all the freedom in constructing different models in the framework of tmt consists of the choice of the concrete matter content and the lagrangians @xmath14 and @xmath15 that is quite similar to the standard field theory .
since @xmath16 is a total derivative , a shift of @xmath27 by a constant , @xmath32 , has no effect on the equations of motion .
similar shift of @xmath15 would lead to the change of the constant part of the lagrangian coupled to the volume element @xmath33 . in the standard
gr , this constant term is the cosmological constant . however in tmt the relation between the constant term of @xmath15 and the physical cosmological constant is very non trivial ( see @xcite-@xcite,@xcite-@xcite ) . in the case of the definition of @xmath16 by means of eq.([phi ] ) , varying the measure fields @xmath20 , we get @xmath34 since @xmath35 it follows that if @xmath36 , @xmath37 where @xmath38 and @xmath39 is a constant of integration with the dimension of mass . in what follows we make the choice @xmath40 . in the case of the definition ( [ aabg ] ) , variation of @xmath23 yields @xmath41 that implies eq.([varphi ] ) without the condition @xmath36 needed in the model with four scalar fields @xmath20 .
one should notice _ the very important differences of tmt from scalar - tensor theories with nonminimal coupling _ : + a ) in general , the lagrangian density @xmath27 ( coupled to the measure @xmath16 ) may contain not only the scalar curvature term ( or more general gravity term ) but also all possible matter fields terms .
this means that tmt modifies in general both the gravitational sector and the matter sector ; b ) if the field @xmath16 were the fundamental ( non composite ) one then instead of ( [ varphi ] ) , the variation of @xmath16 would result in the equation @xmath42 and therefore the dimensionfull integration constant @xmath43 would not appear in the theory .
applying the palatini formalism in tmt one can show ( see for example @xcite and appendix a of the present paper ) that in addition to the christoffel s connection coefficients , the resulting relation between connection and metric includes also the gradient of the ratio of the two measures @xmath44 which is a scalar field .
consequently geometry of the space - time with the metric @xmath19 is non - riemannian .
the gravity and matter field equations obtained by means of the first order formalism contain both @xmath45 and its gradient as well .
it turns out that at least at the classical level , the measure fields affect the theory only through the scalar field @xmath45 .
the consistency condition of equations of motion has the form of a constraint which determines @xmath46 as a function of matter fields .
the surprising feature of the theory is that neither newton constant nor curvature appear in this constraint which means that the _ geometrical scalar field _
@xmath46 _ is determined by the matter fields configuration _ locally and straightforward ( that is without gravitational interaction ) . by an appropriate change of the dynamical variables which includes a conformal transformation of the metric
, one can formulate the theory in a riemannian space - time .
the corresponding conformal frame we call `` the einstein frame '' .
the big advantage of tmt is that in the very wide class of models , _ the gravity and all matter fields equations of motion take canonical gr form in the einstein frame_. all the novelty of tmt in the einstein frame as compared with the standard gr is revealed only in an unusual structure of the scalar fields effective potential ( produced in the einstein frame ) , masses of fermions and their interactions with scalar fields as well as in the unusual structure of fermion contributions to the energy - momentum tensor : all these quantities appear to be @xmath45 dependent@xcite-@xcite .
this is why the scalar field @xmath46 determined by the constraint as a function of matter fields , has a key role in dynamics of tmt models .
the tmt models possessing a global scale invariance@xcite-@xcite,@xcite-@xcite are of significant interest because they demonstrate the possibility of spontaneous breakdown of the scale symmetry .
in fact , if the action ( [ s ] ) is scale invariant then this classical field theory effect results from eq.([varphi ] ) , namely from solving the equation of motion ( [ varphib ] ) or ( [ adl1 ] ) . one of the interesting applications of the scale invariant tmt models@xcite is a possibility to generate the exponential potential for the scalar field @xmath0 by means of the mentioned spontaneous symmetry breaking even without introducing any potentials for @xmath0 in the lagrangians @xmath14 and @xmath15 of the action ( [ s ] ) . some cosmological applications of this effect have been also studied in ref.@xcite .
a dilaton field @xmath0 allows to realize a spontaneously broken global scale invariance@xcite and together with this it governs the evolution of the universe on different stages : in the early universe @xmath0 plays the role of inflaton and in the late time universe it is transformed into a part of the dark energy .
according to the general prescriptions of tmt , we have to start from studying the self - consistent system of gravity ( metric @xmath19 and connection @xmath47 ) , the measure @xmath16 degrees of freedom ( i.e. measure fields @xmath48 or @xmath23 ) and the dilaton field @xmath0 proceeding in the first order formalism .
we postulate that the theory is invariant under the global scale transformations : @xmath49 if the definition ( [ aabg ] ) is used for the measure @xmath16 then the transformations of @xmath20 in ( [ st ] ) should be changed by @xmath50 .
this global scale invariance includes the shift symmetry of the dilaton @xmath0 and this is the main factor why it is important for cosmological applications of the theory@xcite-@xcite,@xcite-@xcite .
we choose an action which , except for the modification of the general structure caused by the basic assumptions of tmt , _ does not contain any exotic terms and fields _ as like in the conventional formulation of the minimally coupled scalar - gravity system . keeping the general structure ( [ s ] ) , it is convenient to represent the action in the following form : @xmath51 \label{totaction}\end{aligned}\ ] ] in the action ( [ totaction ] ) there are two types of the gravitational terms and of the `` kinetic - like terms '' which respect the scale invariance : the terms of the one type coupled to the measure @xmath16 and those of the other type coupled to the measure @xmath4 .
using the freedom in normalization of the measure fields ( @xmath20 in the case of using eq.([phi ] ) or @xmath23 when using eq.([aabg ] ) ) , we set the coupling constant of the scalar curvature to the measure @xmath16 to be @xmath52 .
normalizing all the fields such that their couplings to the measure @xmath16 have no additional factors , we are not able in general to provide the same in terms describing the appropriate couplings to the measure @xmath4 . this fact explains the need to introduce the dimensionless real parameters @xmath53 and @xmath54 and we will only assume that _ they have close orders of magnitudes_. note that in the case of the choice @xmath55 we would proceed with _ the fine tuned model_. the real positive parameter @xmath56 is assumed to be of the order of unity ; in all solutions presented in this paper we set @xmath57 . for newton
constant we use @xmath58 , @xmath59 .
one should also point out the possibility of introducing two different pre - potentials which are exponential functions of the dilaton @xmath0 coupled to the measures @xmath16 and @xmath4 with factors @xmath60 and @xmath61 .
such @xmath0-dependence provides the scale symmetry ( [ st ] ) .
however @xmath60 and @xmath61 might be higgs - dependent and then they play the role of the higgs pre - potentials , but this will be done in the future publication . in appendix
a we present the equations of motion resulting from the action ( [ totaction ] ) when using the original set of variables .
the common feature of all the equations in the original frame is that the measure @xmath16 degrees of freedom appear only through dependence upon the scalar field @xmath45 , eq.([zeta ] ) . in particular , all the equations of motion and the solution for the connection coefficients include terms proportional to @xmath62 , that makes space - time non riemannian and all equations of motion - noncanonical .
it turns out that when working with the new metric ( @xmath0 remains the same ) @xmath63 which we call the einstein frame , the connection becomes riemannian .
since @xmath64 is invariant under the scale transformations ( [ st ] ) , spontaneous breaking of the scale symmetry ( by means of eq.([varphi ] ) which for our model ( [ totaction ] ) takes the form ( [ app1 ] ) ) is reduced in the einstein frame to the _ spontaneous breakdown of the shift symmetry _ @xmath65 notice that the goldstone theorem generically is not applicable in this model@xcite .
the reason is the following .
in fact , the scale symmetry ( [ st ] ) leads to a conserved dilatation current @xmath66 .
however , for example in the spatially flat frw universe the spatial components of the current @xmath67 behave as @xmath68 as @xmath69 . due to this anomalous behavior at infinity
, there is a flux of the current leaking to infinity , which causes the non conservation of the dilatation charge .
the absence of the latter implies that one of the conditions necessary for the goldstone theorem is missing .
the non conservation of the dilatation charge is similar to the well known effect of instantons in qcd where singular behavior in the spatial infinity leads to the absence of the goldstone boson associated to the @xmath70 symmetry .
after the change of variables to the einstein frame ( [ ct ] ) and some simple algebra , eq.([app4 ] ) takes the standard gr form @xmath71 where @xmath72 is the einstein tensor in the riemannian space - time with the metric @xmath64 ; the energy - momentum tensor @xmath73 is now @xmath74 where the function @xmath75 is defined as following : @xmath76 -v_{2}}{(\zeta + b_{g})^{2}}. \label{veff1}\ ] ] putting @xmath39 in the arguments of @xmath77 we indicate explicitly that @xmath77 incorporates our choice for the integration constant @xmath39 that appears as a result of the spontaneous breakdown of the scale symmetry .
we will see in the next sections that @xmath45-dependence of @xmath78 in the form of _ square _ of @xmath79 has a key role in the resolution of the old cc problem in tmt .
the reason that only such @xmath45-dependence emerges in @xmath78 , and a @xmath45-dependence is absent for example in the numerator of @xmath78 , is a direct result of our basic assumption that @xmath80 and @xmath81 are independent of the measure fields ( see item 2 in sec.ii ) .
the dilaton @xmath0 field equation ( [ phi - orig ] ) in the einstein frame reads @xmath82 \nonumber\\ & -&\frac{\alpha}{m_{p}(\zeta + b_{g})^{2 } } \left[(\zeta + b_{g})m^{4}e^{-2\alpha\phi / m_{p}}-(\zeta -b_{g})v_{1 } -2v_{2}-\delta b_{g}(\zeta + b_{g})\frac{1}{2}\tilde{g}^{\alpha\beta}\phi_{,\alpha}\phi_{,\beta}\right ] = 0 \label{phief}\end{aligned}\ ] ] the scalar field @xmath45 in eqs.([tmn])-([phief ] ) is determined by means of the constraint ( [ app3 ] ) which in the einstein frame ( [ ct ] ) takes the form @xmath83 - 2v_{2}-\delta\cdot b_{g}(\zeta + b_{g})x = 0\label{constraint2}\end{aligned}\ ] ] where @xmath84 applying the constraint ( [ constraint2 ] ) to eq.([phief ] ) one can reduce the latter to the form
@xmath85-\frac{2\alpha\zeta}{(\zeta + b_{g})^{2}m_{p}}m^{4}e^{-2\alpha\phi / m_{p } } = 0 , \label{phi - after - con}\ ] ] where @xmath45 is a solution of the constraint ( [ constraint2 ] ) .
the effective energy - momentum tensor ( [ tmn ] ) can be represented in a form of that of a perfect fluid @xmath86 with the following energy and pressure densities resulting from eqs.([tmn ] ) and ( [ veff1 ] ) after inserting the solution @xmath87 of eq.([constraint2 ] ) : @xmath88 } , \label{rho1}\ ] ] @xmath89}. \label{p1}\ ] ] in a spatially flat frw universe with the metric @xmath90 filled with the homogeneous scalar sector field @xmath0 , the @xmath0 field equation of motion takes the form @xmath91 where @xmath92 is the hubble parameter and we have used the following notations @xmath93 @xmath94\rho_{,x } = ( b_{g}+b_{\phi})(m^{4}e^{-2\alpha\phi / m_{p}}+v_{1})- 2v_{2}-3\delta^{2}b_{g}^{2}x \label{q1}\ ] ] @xmath95p_{,x}= ( b_{g}+b_{\phi})(m^{4}e^{-2\alpha\phi / m_{p}}+v_{1})- 2v_{2}-\delta^{2}b_{g}^{2}x\label{q2}\ ] ] @xmath96\rho_{,\phi } = \frac{1}{[b_{g}(m^{4}e^{-2\alpha\phi / m_{p}}+v_{1})-v_{2 } ] } \nonumber \\ & \times & \left[(m^{4}e^{-2\alpha\phi / m_{p}}+v_{1 } ) [ b_{g}(m^{4}e^{-2\alpha\phi / m_{p}}+v_{1})-2v_{2 } ] + 2\delta b_{g}v_{2}x+3\delta^{2}b_{g}^{3}x^{2}\right ] .
\label{q3}\end{aligned}\ ] ] it is interesting that the non - linear @xmath2-dependence appears here in the framework of the fundamental theory without exotic terms in the lagrangians @xmath80 and @xmath81 , see eqs.([s ] ) and ( [ totaction ] ) .
this effect follows just from the fact that there are no reasons to choose the parameters @xmath97 and @xmath54 in the action ( [ totaction ] ) to be equal in general ; on the contrary , the choice @xmath98 would be a fine tuning . besides one should stress that the @xmath0 dependence in @xmath9 , @xmath99 and in equations of motion emerges only in the form @xmath100 where @xmath39 is the integration constant ( see eq.(a1 ) ) , i.e. due to the spontaneous breakdown of the scale symmetry ( [ st ] ) ( or the shift symmetry ( [ phiconst ] ) in the einstein frame ) .
thus the above equations represent an _ explicit example of @xmath3-essence_@xcite _ resulting from first principles_. the system of equations ( [ gef ] ) , ( [ rho1])-([phi1 ] ) accompanied with the functions ( [ q1])-([q3 ] ) and written in the metric @xmath90 can be obtained from the k - essence type effective action @xmath101 \label{k - eff},\ ] ] where @xmath102 is given by eq.([p1 ] ) .
in contrast to the simplified models studied in literature@xcite , it is impossible here to represent @xmath103 in a factorizable form like @xmath104 .
the qualitative analysis of equations is significantly simplified if @xmath105 .
this is what we will assume in this section .
although it looks like a fine tuning of the parameters ( i.e. @xmath98 ) , it allows us to understand qualitatively the basic features of the model .
in fact , only in the case @xmath105 the effective action ( [ k - eff ] ) takes the form of that of the scalar field without higher powers of derivatives .
role of @xmath106 in a possibility to produce an effect of a super - accelerated expansion of the late time universe will be studied in sec.v . in the models with @xmath107 : fig.(a ) @xmath108 ( in the graph @xmath109 and @xmath110 ) ; fig.(b ) @xmath111 ( in the graph @xmath109 and @xmath112 ) .
the value of @xmath113 in the minimum @xmath114 is larger than zero ; fig.(c ) @xmath115 , @xmath116 ( in the graph @xmath117 and @xmath118 ) .
@xmath119 in the minimum @xmath120 . in all the cases here as well as in all solutions presented in this paper
we choose @xmath121.,width=680,height=453 ] so let us study spatially flat frw cosmological models governed by the system of equations @xmath122 and ( [ rho1])-([phi1 ] ) where one should set @xmath105 . in the fine tuned case under consideration ,
the constraint ( [ constraint2 ] ) yields @xmath123 the energy density and pressure take then the canonical form , @xmath124 where the effective potential of the scalar field @xmath0 results from eq.([veff1 ] ) @xmath125^{2 } } { 4[b_{g}\left(v_{1}+m^{4}e^{-2\alpha\phi / m_{p}}\right)-v_{2 } ] } \label{veffvac - delta=0}\ ] ] and the @xmath0-equation ( [ phi1 ] ) is reduced to @xmath126 notice that @xmath127 is non - negative for any @xmath0 provided @xmath128 that we will assume in this paper .
we assume also that @xmath129 . in the following three subsections we consider three different dilaton - gravity cosmological models determined by different choice of the parameters @xmath60 and @xmath61 : one model with @xmath115 and two models with @xmath130 .
the appropriate three possible shapes of the effective potential @xmath127 are presented in fig.1 .
a special case with the fine tuned condition @xmath131 is discussed in appendix b where we show that equality of the couplings to measures @xmath16 and @xmath4 in the action ( equality @xmath131 is one of the conditions for this to happen ) gives rise to a symmetric form of the effective potential .
the most remarkable feature of the effective potential ( [ veffvac - delta=0 ] ) is that it is proportional to the square of @xmath132 which is a straightforward consequence of our basic assumption that @xmath80 and @xmath81 are independent of the measure fields ( see item 2 in sec.ii , eq.([veff1 ] ) and the discussion after it ) . due to this , as @xmath115 and @xmath116 , _ the effective potential has a minimum where it equals zero automatically _
, without any further tuning of the parameters @xmath60 and @xmath61 ( see also fig.1c ) .
this occurs in the process of evolution of the field @xmath0 at the value of @xmath133 where @xmath134 this means that the universe evolves into the state with zero cosmological constant without any additional tuning of the parameters and initial conditions . versus @xmath0 ) for the model with @xmath115 and @xmath116 .
all phase curves started with @xmath135 quickly approach the attractor long before entering the oscillatory regime .
the region of the oscillatory regime is marked by point @xmath136 .
the oscillation spiral is not visible here because of the chosen scales along the axes.,width=566,height=302 ] ( fig .
( a ) ) and the energy density @xmath9 ( fig .
( b ) ) upon @xmath137 . here and in all the graphs of this paper describing scale factor @xmath138 dependences , @xmath139 is normalized such that at the end point of the described process @xmath140 . in the model with @xmath115
the power law inflation ends with damped oscillations of @xmath0 around @xmath141 determined by eq.([veff=0 ] ) . for the choice @xmath117 eq.([veff=0 ] )
gives @xmath142 .
( b ) the exit from the early inflation is accompanied with approaching zero of the energy density @xmath9 .
the graphs correspond to the evolution which starts from the initial values @xmath143 , @xmath144 .
, width=604,height=264 ] as function of the scale factor for the parameters and initial conditions as in fig.3 .
most of the time the expansion of the universe is a power law inflation with almost constant @xmath145 ; @xmath6 oscillates between @xmath8 and @xmath146 at the exit from inflation stage , i.e. as @xmath147 and @xmath148.,title="fig:",width=321,height=302 ] as function of the scale factor for the parameters and initial conditions as in fig.3 .
most of the time the expansion of the universe is a power law inflation with almost constant @xmath145 ; @xmath6 oscillates between @xmath8 and @xmath146 at the exit from inflation stage , i.e. as @xmath147 and @xmath148.,title="fig:",width=321,height=302 ] to provide the global scale invariance ( [ st ] ) , the prepotentials @xmath149 and @xmath150 enter in the action ( [ totaction ] ) with factor @xmath151 .
however , if quantum effects ( considered in the original frame ) break the scale invariance of the action via modification of existing prepotentials or by means of generation of other prepotentials with arbitrary @xmath0 dependence ( and in particular a `` normal '' cosmological constant term @xmath152 ) , this can not change the result of tmt that the effective potential generated in the einstein frame is proportional to a perfect square .
note that the assumption of scale invariance is not necessary for the effect of appearance of the perfect square in the effective potential in the einstein frame and therefore for the described mechanism of disappearance of the cosmological constant , see refs.@xcite-@xcite .
if such type of the structure for the scalar field potential in a conventional ( non tmt ) model would be chosen `` by hand '' it would be a sort of fine tuning .
but in our tmt model it is not the starting point , _ it is rather a result _ obtained in the einstein frame of tmt models with spontaneously broken global scale symmetry including the shift symmetry @xmath153 .
on the first glance this effect contradicts the weinberg s no - go theorem@xcite which states that there can not exist a field theory model where the cosmological constant is zero without fine tuning . in sec.vib we will study in detail the manner our tmt model avoids this theorem .
as @xmath154 .
so , as @xmath155 the model describes the well studied power law inflation of the early universe@xcite if @xmath156 : @xmath157 further behavior of the solutions in the model of sec.iva with @xmath115 and @xmath116 , i.e. for the potential of fig.1c , is qualitatively evident enough . with the choice @xmath121 , @xmath117 and @xmath118 , the results of numerical solutions are presented in figs
the phase curves in fig.2 demonstrate the well known attractor behavior of the solutions ) may be nonapplicable for a relatively short period of time from the very beginning where @xmath158 may be negative that we observe in fig.2 .
] describing the power law inflation@xcite .
exit from the inflation starts as @xmath0 becomes close to @xmath159 determined by eq.([veff=0 ] ) .
then the energy density starts to tend to zero very fast , fig.3b .
the exit from the inflation occurs when all the phase trajectories corresponding to different initial conditions with @xmath160 practically coincide .
the process ends with oscillatory regime , fig.4a , where @xmath0 performs damped oscillations around the minimum of the effective potential ( see also fig.1c ) . and @xmath61 are as in fig.1a .
the graphs correspond to the initial conditions @xmath164 , @xmath165 .
the early universe evolution is governed by the almost exponential potential ( see fig.1a ) providing the power low inflation ( @xmath145 interval in fig.(c ) ) .
after transition to the late time universe the scalar @xmath0 increases with the rate typical for a quintessence scenario . later on the cosmological constant @xmath166 becomes a dominated component of the dark energy that is displayed by the infinite region where @xmath167 in fig.(c).,width=680,height=226 ] versus @xmath0 ) for the model with @xmath168 .
all trajectories approach the attractor which in its turn asymptotically ( as @xmath169 ) takes the form of the straight line @xmath170.,width=529,height=302 ] in this model the effective potential ( [ veffvac - delta=0 ] ) is a monotonically decreasing function of @xmath0 ( see fig.[fig1]a ) . as @xmath171 the model describes the power law inflation ( [ p - l - solution ] ) ,
similar to what we discussed in the model of previous subsection . applying this model to the cosmology of the late time universe and assuming that the scalar field @xmath169 as @xmath7 , it is convenient to represent the effective potential ( [ veffvac - delta=0 ] ) in the form @xmath172 with the definition @xmath173 here @xmath163 is the positive cosmological constant ( see ( [ bv1>v2 ] ) ) and @xmath174 } , \label{v - quint - without - ferm - delta=0}\ ] ] that is the evolution of the late time universe is governed both by the cosmological constant @xmath175 and by the quintessence - like potential @xmath176 .
thus the effective potential ( [ veffvac - delta=0 ] ) provides a possibility for a cosmological scenario which starts with a power law inflation and ends with a cosmological constant @xmath175 .
the smallness of @xmath175 may be achieved without fine tuning of dimensionfull parameters , that will be discussed in sec.vi .
such scenario may be treated as a generalized quintessential inflation type of scenario . recall that the @xmath0-dependence of the effective potential ( [ veffvac - delta=0 ] ) appears here only as the result of the spontaneous breakdown of the global scale symmetry and @xmath116 was studied in ref.@xcite .
the application of the tmt model with explicitly broken global scale symmetry to the quintessential inflation scenario was discussed in ref@xcite . ] .
results of numerical solutions for such type of scenario are presented in figs.5 and 6 ( @xmath109 , @xmath110 ) the early universe evolution is governed by the almost exponential potential ( see fig.1a ) providing the power low inflation ( @xmath145 interval in fig.(c ) ) with the attractor behavior of the solutions , see ref.@xcite .
after transition to the late time universe the scalar @xmath0 increases with the rate typical for a quintessence scenario .
later on the cosmological constant @xmath166 becomes a dominated component of the dark energy that is displayed by the infinite region where @xmath177 in fig.(c ) .
the phase portrait in fig.3 shows that all the trajectories started with @xmath135 quickly approach the attractor which asymptotically ( as @xmath169 ) takes the form of the straight line @xmath170 .
qualitatively similar results are obtained also when @xmath60 is positive but @xmath61 is negative . in this case
the effective potential ( [ veffvac - delta=0 ] ) has the minimum ( see fig.1b ) @xmath179 for the choice of the parameters as in fig.1b , i.e. @xmath109 and @xmath112 , the minimum is located at @xmath114 .
the character of the phase portrait one can see in fig.[fig4 ] . versus @xmath0 ) for the model with @xmath111 and @xmath130 (
the parameters are chosen here as in fig.1b ) .
trajectories started anywhere in the phase plane in a finite time end up at the same point @xmath180 which is a node sink .
there exist two attractors ending up at @xmath181 , one from the left and other from the right .
all phase curves starting with @xmath135 quickly approach these attractors.,width=604,height=302 ] and @xmath130 : typical dependence of @xmath0 ( fig.(a ) ) , the energy density @xmath9 ( fig.(b ) ) and the equation - of - state @xmath6 ( fig.(c ) ) upon @xmath137 where the scale factor @xmath139 normalized as in fig.2 .
the graphs correspond to the initial conditions @xmath182 , @xmath183 .
the early universe evolution is governed by an almost exponential potential ( see fig.1b ) providing the power low inflation ( @xmath145 interval in fig.(c ) ) . after arriving the minimum of the potential at @xmath184 ( see fig.1b and the point @xmath185 of the phase plane in fig.4 ) the scalar @xmath0 remains constant . at this stage the dynamics of the universe is governed by the constant energy density @xmath186 ( see the appropriate intervals @xmath187 in fig.(b ) and @xmath188 in fig.(c)).,width=680,height=226 ] for the early universe as @xmath171 , similar to what we have seen in the models of the previous two subsections , the model implies the power low inflation .
however , the phase portrait fig.[fig4 ] shows that now all solutions end up without oscillations at the minimum @xmath114 with @xmath189 . in this final state of the scalar field @xmath0 ,
the evolution of the universe is governed by the cosmological constant @xmath190 determined by eq.([minveff ] ) .
for some details of the cosmological dynamics see fig.[fig5 ] .
the desirable smallness of @xmath190 can be provided again without fine tuning of the dimensionfull parameters that will be discussed in sec.vi .
the absence of appreciable oscillations in the minimum is explained by the following two reasons : a ) the non - zero friction at the minimum determined by the cosmological constant @xmath190 ; b ) the shape of the potential near to minimum is too flat .
the described properties of the model are evident enough .
nevertheless we have presented them here because this model is a particular ( fine tuned ) case of an appropriate model with @xmath106 studied in the next subsection where we will demonstrate a possibility of states with @xmath191 without any exotic contributions , like a phantom term , in the original action .
we return now to the general case of our model ( see sec.iii ) with no fine tuning of the parameters @xmath53 and @xmath54 , i.e. the parameter @xmath192 , defined by eq.([delta ] ) , is non zero .
we have checked that the models of secs.ivb and ivc do not suffer qualitative changes at all and quantitative modifications are not strong , at least in the range of @xmath193 . however the model of sec.ivd changes drastically as @xmath106 : the model enables k - essence type solutions for the late time universe with equation - of - state @xmath191 . , @xmath194 , @xmath109 and @xmath112 .
the region with @xmath195 is divided into two dynamically disconnected regions by the line @xmath196 . to the left of this line @xmath197 ( the appropriate zone we call zone 1 ) and to the right - @xmath198 .
the phase portrait in zone 1 corresponds to processes similar to those of sec.ivd .
the @xmath195 region to the right of the line @xmath196 is divided into two zones ( zone 2 and zone 3 ) by the line @xmath199 ( the latter coincides with the line where @xmath200 ) . in zone 2 @xmath201 but @xmath202 . in zone 3 @xmath191 and @xmath203 .
phase curves can start in zone 2 in points very close to the line @xmath204 .
after they cross the line @xmath200 , i.e. in zone 3 , they exhibit processes with super - accelerating expansion of the universe .
the phase curves in zone 3 demonstrate dynamical attractor behavior .
as @xmath169 the phase curves approach the straight line @xmath205.,width=642,height=453 ] ( fig.(a ) ) and of the energy density @xmath9 , defined by eq.([rho1 ] ) , ( fig(b ) ) in the regime corresponding to the phase curves started in zone 2 .
both graphs correspond to the initial conditions @xmath206 , @xmath207 ; @xmath9 increases approaching asymptotically @xmath208 , where @xmath209 is the value of @xmath163 determined by eq.([lambda ] ) as @xmath109 and @xmath112 ; see also fig.1b.,width=604,height=188 ] ( fig.(a ) ) and effective speed of sound for perturbations @xmath210 ( fig.(b ) ) for the initial conditions @xmath206 , @xmath207.,width=604,height=188 ]
so , we consider now the dynamics of the frw cosmology described by eqs.([cosm - phi])-([q3 ] ) . before choosing the appropriate parameters for numerical studies ,
let us start from the analysis of eq.([phi1 ] ) .
the interesting feature of this equation is that for certain range of the parameters , each of the factors @xmath211 ( @xmath212 ) can get to zero .
equation @xmath213 determines a line in the phase plane @xmath214 . in terms of a mechanical interpretation of eq.([phi1 ] ) , the change of the sign of @xmath215 can be treated as the change of the mass of `` the particle '' .
therefore one can think of situation where `` the particle '' climbs up in the potential with acceleration .
it turns out that when the scalar field is behaving in this way , the flat frw universe may undergo a super - acceleration . for @xmath216 , eqs .
( [ phi1 ] ) , ( [ cosm - phi ] ) result in the well known equation@xcite @xmath217 where @xmath218 is the effective sound speed of perturbations@xcite @xmath219 @xmath220 @xmath9 and @xmath99 are defined by eqs.([rho1 ] ) , ( [ p1 ] ) and @xmath221 @xmath222 - by eqs.([q1])-([q3 ] ) .
with simple algebra one can see that the following `` sign rule '' is fulfilled for the equation - of - state @xmath223 : @xmath224 and the line @xmath225 divides the phase plane @xmath226 into two regions : one with @xmath201 and other with @xmath191 .
there are a lot of sets of parameters providing the @xmath191 phase in the late universe . for example
we are demonstrating here this effect with the following set of the parameters of the original action([totaction ] ) : @xmath121 , @xmath109 and @xmath112 used in sec.ivd but now we choose @xmath227 .
the results of the numerical solution are presented in figs.9 - 11 .
the phase plane , fig.9 , is divided into two regions by the line @xmath228 .
the region @xmath195 is divided into two dynamically disconnected regions by the line @xmath229 . to the left of the line @xmath229 - zone 1 where @xmath197 .
comparing carefully the phase portrait in the zone 1 with that in fig.[fig4 ] of the previous subsection , one can see an effect of @xmath106 on the shape of phase trajectories .
however the general structure of these two phase portraits is very similar . in particular
, they have the same node sink @xmath230 . at this point
`` the force '' equals zero since @xmath231 .
the value @xmath232 coincides with the position of the minimum of @xmath127 because in the limit @xmath233 the role of the terms proportional to @xmath192 is negligible . among trajectories converging to node @xmath181 there are also trajectories corresponding to a power low inflation of the early universe , which is just a generalization to the case @xmath106 of the similar result discussed in the previous subsection . in the region to the right of the line @xmath229 , all phase curves approach the attractor which in its turn asymptotically ( as @xmath169 ) takes the form of the straight line @xmath170 .
this region is divided into two zones by the line @xmath225 . in all points of this line @xmath200 . in zone 2 ,
i.e. between the lines @xmath204 and @xmath200 , the equation - of - state @xmath201 and the sound speed @xmath202 . in zone 3 ,
i.e. between the line @xmath200 and the line @xmath228 , the equation - of - state @xmath191 and the sound speed @xmath203 . for a particular choice of the initial data @xmath206 , @xmath234 ,
the features of the solution of the equations of motion are presented in figs.10 and 11 .
the main features of the solution as we observe from the figures are the following : 1 ) @xmath0 slowly increases in time ; 2 ) the energy density @xmath9 slowly increases approaching the constant @xmath235 defined by the same formula as in eq.([lambda ] ) , see also fig.1b ; for the chosen parameters @xmath236 ; 3 ) @xmath237 is less than @xmath8 and asymptotically approaches @xmath8 from below . using the classification of ref.@xcite of conditions for the dark energy to evolve from the state with @xmath201 to the phantom state
, we see that transition of the phase curves from zone 2 ( where @xmath201 ) to the phantom zone 3 ( where @xmath191 ) occurs under the conditions @xmath238 , @xmath239 , @xmath240 .
qualitatively the same behavior one observes for all initial conditions @xmath241 disposed in the zone 2 .
the question of constructing a realistic scenario where the dark energy can evolve from the power low inflation state disposed in zone 1 to the phantom zone 3 is beyond of the goal of this paper .
the smallness of the observable cosmological constant @xmath163 is known as the new cosmological constant problem@xcite . in tmt
, there are two ways to provide the observable order of magnitude of @xmath242 by an appropriate choice of the parameters of the theory ( see eqs.([lambda ] ) and ( [ bv1>v2 ] ) ) but _ without fine tuning of the dimensionfull parameters_. if @xmath116 then there is no need for @xmath60 and @xmath61 to be small : it is enough that @xmath243 and latexmath:[$v_{1}/ mechanism@xcite,@xcite ) .
for instance , if @xmath60 is determined by the energy scale of electroweak symmetry breaking @xmath245 and @xmath61 is determined by the planck scale @xmath246 then @xmath247 .
the range of the possible scale of the dimensionless parameter @xmath97 remains very broad .
let us start from the notion that if @xmath248 or alternatively @xmath249 and @xmath250 then @xmath251 .
hence the second possibility to ensure the needed smallness of @xmath163 is to choose the _ dimensionless _
parameter @xmath129 to be a huge number . in this case
the order of magnitudes of @xmath60 and @xmath61 could be either as in the above case of the seesaw mechanism or to be not too much different from each other ( or even of the same order ) .
for example , if @xmath252 then for getting @xmath253 one should assume that @xmath254 . below we will use this value just for illustrative purposes .
note that @xmath97 is the ratio of the coupling constants of the scalar curvature to the measures @xmath4 and @xmath16 respectively in the fundamental action ( [ totaction ] ) .
the lagrangians @xmath80 and @xmath81 have the same structure : both of them contain the scalar curvature , kinetic and pre - potential terms .
it is natural to assume that the ratio of couplings of all the appropriate terms in @xmath80 and @xmath81 to the measures @xmath4 and @xmath16 have the same or close orders of magnitude .
this is why in sec.iii we have made an assumption that the dimensionless parameters @xmath53 and @xmath54 have close orders of magnitude .
for the same reason we will also assume that @xmath255 . if this is the case then the choice @xmath256 means that in this case the second way of resolution of the new cc problem is a particular case of the seesaw mechanism .
however the second way is applicable also if @xmath257 then the huge value of @xmath53 can be treated as an indication that tmt implies a certain sort of _ the correspondence principle _ between tmt and conventional field theories ( i.e theories with only the measure of integration @xmath4 in the action ) .
in fact , using the notations of the general form of the tmt action ( [ s ] ) in the case of the action ( [ totaction ] ) , one can conclude that the relation between the `` usual '' ( i.e. entering in the action with the usual measure @xmath4 ) lagrangian density @xmath81 and the new one @xmath80 ( entering in the action with the new measure @xmath16 ) is roughly speaking @xmath258 . in the case
@xmath259 becomes negligible , the remaining term of the action @xmath260 would describe gr instead of tmt .
it seems to be very interesting that such _ a correspondence principle for the tmt action ( [ s ] ) may have a certain relation to the extreme smallness of the cosmological constant_. appearance of a big dimensionless constant in particle field theory is usually associated with hierarchy of masses and/or interactions describing by _ different _ terms in the lagrangian .
the way big numbers can appear in the tmt action is absolutely different .
it is easier to see this difference in the case of a fine tuned model , where @xmath55 and @xmath261 , see appendix b. in such a case the lagrangians @xmath80 and @xmath81 not only have the same type of terms but they are just proportional : @xmath262
. therefore the nature of the huge value of @xmath53 differs here very much from the conventional hierarchy issue and @xmath54 one can define the fundamental tmt action such that the appropriate factors in @xmath81 equal unity . then in front of the appropriate terms in @xmath80 one should add factors like @xmath263 and @xmath264 .
with such a definition of the lagrangians @xmath80 and @xmath81 , instead of a huge factor in @xmath81 , a very small factor will appear in @xmath80 . ] .
if such the ratio between @xmath80 and @xmath81 is actually realized , then taking into account the fact that @xmath80 and @xmath81 describe the same matter and gravity degrees of freedom in a very similar manner , the question arises why @xmath80 is not dynamically negligible in comparison with @xmath81 . to answer this question we have to turn to the fundamental action ( [ totaction ] ) that it is convenient to rewrite in the following form @xmath265 \label{totaction1}\end{aligned}\ ] ] where one can see that the ratio @xmath266 has an important dynamical role . analyzing the constraint ( [ constraint2 ] ) and cosmological dynamics studied in secs.iv and v it is easy to see that the order of magnitude of the scalar field @xmath267 is generically close to that of @xmath53 . in other words ,
when analyzing the action on the mass shell , it turns out that the ratio of the measures @xmath268 generically compensates the smallness of @xmath269 .
so , similar terms ( @xmath270-terms , kinetic terms and pre - potential terms ) appearing in the action ( [ totaction ] ) with measures @xmath16 and @xmath4 respectively , are both dynamically important in general . in the light of this understanding of the general picture
it is interesting to check the tmt dynamics in situations where @xmath266 becomes very small or very large .
let us start from the fine tuned model where @xmath271 and @xmath55 ( recall that @xmath161 and @xmath272 in all our models ) . in this case
it follows from the constraint ( [ zeta - without - ferm - delta=0 ] ) that @xmath273 then the effective potential ( [ veffvac - delta=0 ] ) ( see also eqs.([rho - without - ferm])-([v - quint - without - ferm - delta=0 ] ) ) reads @xmath274 in such a fine tuned model , @xmath266 approaches zero asymptotically as @xmath169 where the effective potential becomes flat .
however when looking into the tmt action written in the form ( [ totaction1 ] ) we see that the asymptotic disappearance of @xmath266 means that we deal with an asymptotic transition from tmt to a conventional field theory model with only measure of integration @xmath4 and only one lagrangian density . in the limit
@xmath275 , the conformal transformation to the einstein frame ( [ ct ] ) takes the form @xmath276 .
therefore in the einstein frame , the limit of the action ( [ totaction1 ] ) as @xmath275 is reduced to the following model : @xmath277 .
\label{totaction - zeta0}\ ] ] it is easy to see that for example in the frw universe , the asymptotic ( as the scale factor @xmath278 ) behavior of the universe in the model ( [ totaction - zeta0 ] ) coincides with the appropriate asymptotic result of tmt model under consideration : both of them asymptotically describe the universe governed by the cosmological constant @xmath279 .
similar conclusion is obtained in a model where @xmath280 , i.e. with no fine tuning of the prepotentials , which have been studied in sec.ivd .
the only difference is that now @xmath275 and @xmath281 as @xmath282 , see eq.([minveff ] ) ; in a small neighborhood of @xmath283 the tmt action presented in the einstein frame looks like ( [ totaction - zeta0 ] ) .
note however that in the context of the k - essence model studied in sec.v where @xmath106 and @xmath280 , the asymptotic value of @xmath45 in the late time universe ( i.e. as @xmath169 and @xmath284 ) is @xmath285 but nevertheless the energy density tends to @xmath286 as well ( see fig.1b ) .
similar situation takes place in the model where @xmath287 , studied in sec.ivc .
the asymptotic value of @xmath45 in the late time universe is again @xmath288 and the energy density tends to @xmath289 as well ( see fig.1a ) .
thus in all the models , the huge value of @xmath53 can ensure the needed smallness of the dark energy density in the late time universe but it is not always realized due to the limit @xmath275 .
as we have seen in sec.ivb , in the model with @xmath115 and @xmath116 , the old cosmological constant problem is resolved without fine tuning : the effective potential ( [ veffvac - delta=0 ] ) is proportional to the square of @xmath290 , and @xmath133 where @xmath291 , is the minimum of the effective potential without any further tuning of the parameters and initial conditions .
now we want to analyze some of the essential differences we have in tmt as compared with the conditions of the weinberg s no - go theorem and show what are the reasons providing solution of the old cc problem in tmt .
this has to be done when tmt is considered in the original frame since in the einstein frame we observe only the results in the effective picture after some of the symmetries are broken . * the basic assumption of the weinberg s theorem is that in the vacuum all the fields ( metric tensor @xmath19 and matter fields @xmath292 ) are constant . as it was pointed out by s.weinberg in the review@xcite ,
the euler - lagrange equations for such constant fields ( with the action @xmath293 ) have the form @xmath294 @xmath295 and these equations constitute the basis for further weinberg s arguments .
in particular , if @xmath296 symmetry @xmath297 survives as a vestige of general covariance when all the fields are constrained to be constant , the lagrangian @xmath298 transforms as a density : @xmath299 weinberg concludes that when eq.([w1 ] ) is satisfied then the unique form of @xmath298 is @xmath300 where @xmath301 is independent of @xmath19 .
as a matter of fact this means that for example in the case of a scalar matter field @xmath0 model considered by weinberg in sec.vi of the review @xcite , @xmath301 is determined by the value of the scalar field @xmath0 potential as @xmath0 is a constant determined by eq.([w2 ] ) .
+ however , if @xmath298 contains a term linear in space - time derivatives then eq.([w2 ] ) may be not valid even for constant field .
this is what happens in tmt where the first term in the action ( [ s ] ) is linear in space - time derivatives of @xmath23 ( when using the definition ( [ aabg]))(see also ) .
then instead of eq.([w2 ] ) which appears to be an identity in this case , the euler - lagrange equations for @xmath23 look @xmath302 which are nontrivial even for constant @xmath23 and resulting in eq.([adl1 ] ) .
note that @xmath303 is a scalar density and transforms exactly according to eq.([w4 ] ) . therefore generically
( i.e. if @xmath23 are not constant while other fields are constant ) , the lagrangian @xmath304 satisfying ( [ w4 ] ) can have the following form @xmath305 where @xmath306 and @xmath307 are independent of @xmath19 and @xmath16 .
this is why the equation @xmath308 where @xmath309 is the trace of the energy - momentum tensor , obtained by weinberg@xcite for all constant @xmath19 and matter fields , is generically no longer valid .
* let us now note that @xmath310 , eq.([zeta - without - ferm - delta=0 ] ) , becomes singular @xmath311 in this limit the effective potential ( [ veff1 ] ) ( see also eq.([veffvac - delta=0 ] ) ) behaves as @xmath312 + thus , disappearance of the cosmological constant occurs in the regime where @xmath313 . in this limit ,
the dynamical role of the terms of the lagrangian @xmath81 ( coupled with the measure @xmath4 ) in the action ( [ totaction ] ) becomes negligible in comparison with the terms of the lagrangian @xmath80 ( see also the general form of the action ( [ s ] ) ) .
a particular realization of this we observe in the behavior of @xmath77 , eq.([v - singular ] ) .
it is evident that the limit of the tmt action ( [ s ] ) as @xmath313 is opposite to the conventional field theory ( with only measure @xmath4 ) limit of the tmt action discussed in the previous subsection .
* from the point of view of tmt , this is the answer to the question why the old cosmological constant problem can not be solved ( without fine tuning ) in theories with only the measure of integration * @xmath4 * in the action*. * recall that one of the basic assumptions of the weinberg s no - go theorem is that all fields in the vacuum must be constant .
this is also assumed for the metric tensor components of which in the vacuum must be * nonzero * constants . however , this is not the case in the fundamental tmt action ( [ totaction ] ) defined in the original ( non einstein ) frame if we ask what is the metric tensor @xmath19 in the @xmath314 vacuum . to see this
let us note that in the einstein frame all the terms in the cosmological equations are regular .
this means that the metric tensor in the einstein frame @xmath64 is always well defined , including the @xmath314 vacuum state @xmath133 where @xmath45 is infinite .
taking this into account and using the transformation to the einstein frame ( [ ct ] ) we see that * all components of the metric in the original frame @xmath19 go to zero overall in space - time as @xmath0 approaches the @xmath315 vacuum state * : @xmath316 this result shows that the weinberg s analysis based on the study of the trace of the energy - momentum tensor misses any sense in the case @xmath317 .
+ the metric is an attribute of the space - time term .
hence disappearance of the metric @xmath19 in the limit @xmath147 means that the strict formulation of the tmt model ( [ totaction ] ) with @xmath318 and @xmath249 requires a new mathematical basis .
a manifold which is not equipped with the metric ( corresponding to the @xmath314 vacuum state ) emerges as a certain limit of a sequence of space - times .
thus the model under consideration is in fact formulated not in a space - time manifold but rather by means of a set of space - time manifolds .
a limiting point of a sequence of space - times is a `` vacuum space - time manifold '' ( vstm ) one of the differences of which from a regular space - time is the absence of the metric @xmath19 .
+ it follows immediately from ( [ g - degen ] ) that @xmath4 tends to zero like @xmath319 then the definition @xmath267 implies that the integration measure @xmath16 also tends to zero but rather like @xmath320 thus both the measure @xmath16 and
the measure @xmath4 become degenerate in the @xmath314 vacuum state @xmath133 .
however @xmath4 tends to zero more rapidly than @xmath16 .
* as we have discussed in detail ( see secs.ii , iiib and refs.@xcite,@xcite ) , with the original set of variables used in the fundamental tmt action it is very hard or may be even impossible to display the physical meaning of tmt models .
one of the reasons is that in the framework of the postulated need to use the palatini formalism , the original metric @xmath19 and connection @xmath321 appearing in the fundamental tmt action describe a non - riemannian space - time .
the transformation to the einstein frame ( [ ct ] ) enables to see the physical meaning of tmt because the space - time becomes riemannian in the einstein frame .
now we see that the transformation to the einstein frame ( [ ct ] ) plays also the role of a * regularization of the space - time metric * : the singular behavior of the transformation ( [ ct ] ) as @xmath322 compensates the disappearance of the original metric @xmath19 in the vacuum @xmath133 . as a result of this the metric in the einstein frame @xmath64 turns out to be well defined in all physical states including the @xmath314 vacuum state .
the main idea of tmt is that the general form of the action @xmath323 is not enough in order to account for some of the fundamental problems of particle physics and cosmology .
the key difference of tmt from the conventional field theory in curved space - time consists in the hypothesis@xcite-@xcite that in addition to the term in the action with the volume element @xmath18 there should be one more term where the volume element is metric independent but rather it is determined either by four ( in the 4-dimensional space - time ) scalar fields @xmath20 or by a three index potential @xmath23 , see eqs.([s])-([aabg ] ) .
we would like to emphasize that including in the action of tmt the coupling of the lagrangian density @xmath27 with the measure @xmath16 , we modify in general both the gravitational and matter sectors as compared with the standard field theory in curved space - time . besides we made two more assumptions :
the measure fields ( @xmath20 or @xmath23 ) appear only in the volume element ; one should proceed in the first order formalism . _
these assumptions constitute all the modifications of the general structure of the theory we have made as compared with the conventional field theory where only the measure of integration @xmath4 is used in the action principle_. in fact , the lagrangian densities @xmath27 and @xmath15 studied in the present paper , contain only such terms which should be present in a conventional model with minimally coupled to gravity scalar field .
in particular there is no need for the non - linear in kinetic energy terms as well as in the phantom type terms in the fundamental lagrangian densities @xmath27 and @xmath15 in order to obtain a super - acceleration phase at the late time universe . after making use of the variational principle and formulating the resulting equations in the einstein frame ,
we have seen that the effective action ( [ k - eff ] ) represents a concrete realization of the @xmath3-essence@xcite obtained from first principles of tmt without any exotic terms in the lagrangian densities .
as @xmath105 , the dynamics of @xmath0 can be analyzed by means of its effective potential ( [ veffvac - delta=0 ] ) .
as @xmath155 the effective @xmath0 potential has the exponential form and it is proportional to the integration constant @xmath43 . in other words , the effective potential governing the dynamics of the early universe results from the spontaneous breakdown of the global scale invariance ( [ st ] ) caused by the intrinsic feature of tmt ( see eqs .
( [ varphi ] ) and ( [ app1 ] ) ) .
we have seen that independently of the values of the parameters @xmath60 , @xmath61 and under very general initial conditions , solutions rapidly approach a regime characterized by a power law inflation . if @xmath324 , we deal with the intrinsically @xmath3-essence dynamics .
the numerical solutions in this case have showed that there is no qualitative difference from the power law inflation obtained in the case with @xmath105 .
in our toy model there are three regions of the parameters @xmath60 and @xmath61 and appropriate three shapes of the effective potentials , fig.1
. therefore three different types of scenarios for exit from inflation can be realized : \a ) @xmath115 and @xmath116 , sec.ivb . in this case
the power law inflation ends with damped oscillations of @xmath0 approaching the point of the phase plane ( @xmath325 ) where the vacuum energy @xmath326 .
this occurs without fine tuning of the parameters @xmath60 , @xmath61 and the initial conditions .
\b ) @xmath130 and @xmath168 , sec.ivc . in this case
the power law inflation monotonically transforms to the late time inflation asymptotically governed by the cosmological constant @xmath175 .
qualitatively the same results are also obtained in the cases a ) and b ) if @xmath106 .
\c ) @xmath178 , sec.ivd . in this case
the power law inflation ends without oscillations at the final value @xmath283 , corresponding to the ( non zero ) minimum of the effective potential .
the model we have studied in this paper may be extended by including the higgs field , as well as gauge fields and fermions .
it turns out that the scalar sector of such an extended model enables a scenario which resembles a hybrid inflation@xcite .
these results will be presented in a future publication .
\1 ) _ the old cosmological constant problem_. in sec.iva we have seen in details that if @xmath318 then , for a broad range of other parameters , the vacuum energy turns out to be zero without fine tuning . this effect is a direct consequence of the tmt structure which yields the following results : a ) the effective scalar sector potential generated in the einstein frame is proportional to _ a perfect square _ of two terms ; b ) one of those terms is proportional to the integration constant @xmath327 the appearance of which is also the intrinsic feature of tmt . note that the spontaneously broken global scale invariance is not necessary to achieve this effect@xcite .
if such type of the structure for the scalar field potential in a conventional ( non tmt ) model would be chosen `` by hand '' it would be a sort of fine tuning . in sec.vi
we have explained in details how this result avoids the well known no - go theorem by weinberg@xcite stating that generically in field theory one can not achieve zero value of the potential in the minimum without fine tuning .
it is interesting that the resolution of the old cc problem in the context of tmt happens in the regime where @xmath328 . from the point of view of tmt ,
the latter is the answer to the question why the old cosmological constant problem can not be solved ( without fine tuning ) in theories with only the measure of integration @xmath4 in the action .
\2 ) _ the new cosmological constant problem_. interesting result following from the general structure of the scale invariant tmt model with @xmath161 is that the cosmological constant @xmath163 , eq.([lambda ] ) , is a ratio of quantities constructed from pre - potentials @xmath60 , @xmath61 and the dimensionless parameter @xmath97 .
such structure of @xmath163 allows to propose two ways ( see sec.vi ) for resolution of the problem of the smallness of @xmath163 that should be @xmath242 : \a ) the first way is a kind of a _ seesaw _ mechanism@xcite .
for instance , if @xmath245 and @xmath246 then @xmath253 .
\b ) the second way is realized if the dimensionless parameters @xmath97 , @xmath329 and @xmath330 of the action ( [ totaction ] ) are huge numbers of the close orders of magnitude . for example , if @xmath252 then for getting @xmath242 one should assume that @xmath331 . possibility of this idea means that the resolution of the new cosmological constant problem may have a certain relation to _ the correspondence principle _ between tmt and conventional field theories ( see details in sec.via2 ) .
if no fine tuning of the parameters is made in the fundamental action , namely if @xmath332 , then our tmt model has big enough regions in the parameter space where the super - acceleration phase in the late time universe becomes possible .
the appropriate phantom dark energy asymptotically approaches a cosmological constant .
however it is impossible to obtain _ a pure classical solution _ which connects the early universe power law inflation with the late time super - acceleration .
this problem is apparently related with the toy character of the scenario where the role of the matter creation has been ignored : in tmt the fermionic matter generically contributes to the constraint equation for the scalar field @xmath45 and so can effect the field @xmath0 dynamics as well . in this paper we have studied only classical tmt and its possible effects in the context of cosmology
. however quantization of tmt as well as influence of quantum effects on the processes explored in this paper may have a crucial role .
we summarize here some ideas and speculations which gives us a hope that quantum effects can keep the main results of this paper .
recall first two fundamental facts of tmt as a classical field theory : ( a ) the measure degrees of freedom appear in the equations of motion only via the scalar @xmath45 , eq.([zeta ] ) ; ( b ) the scalar @xmath45 is determined ( as a function of matter fields , in our toy model - as a function of @xmath0 ) by the constraint which is nothing but a consistency condition of the equations of motion ( see eqs.([app1])-([app3 ] ) in appendix a and eq.([constraint2 ] ) ) .
therefore the constraint plays a key role in tmt .
note however that if we were ignore the gravity from the very beginning in the action ( [ totaction ] ) then instead of the constraint ( [ app3 ] ) we would obtain eq.([app1 ] ) ( where one has to put zero the scalar curvature ) .
in such a case we would deal with a different theory .
this notion shows that the gravity and matter intertwined in tmt in a much more complicated manner than in gr . hence introducing the new measure of integration @xmath16 we have to expect that the quantization of tmt may be a complicated enough problem .
nevertheless we would like here to point out that in the light of the recently proposed idea of ref.@xcite , the incorporation of four scalar fields @xmath48 together with the scalar density @xmath16 , eq.([phi ] ) , ( which in our case are the measure fields and the new measure of integration respectively ) , is a possible way to define local observables in the local quantum field theory approach to quantum gravity .
we regard this result as an indication that the effective gravity @xmath11 matter field theory has to contain the new measure of integration @xmath16 as it is in tmt .
the assumption formulated in item 2 in sec.iia , that the measure fields @xmath48 ( or @xmath23 ) appear in the action ( [ s ] ) _ only _ via the measure of integration @xmath16 , has a key role in the tmt results and in particular for the resolution of the old cosmological constant problem . in principle
one can think of breakdown of such a structure by quantum corrections .
however , tmt possesses an infinite dimensional symmetry mentioned in item 2 of sec.ii which , as we hope , is able to protect the postulated structure of the action from a deformation caused by quantum corrections .
another effect of quantum corrections is the possible appearance of a nonminimal coupling of the dilaton field @xmath0 to gravity in the form like for example @xmath333 .
proceeding in the first order formalism of tmt one can show that the nonminimal coupling can affect the k - essence dynamics but the mechanism for resolution of the old cc problem exhibited in this paper remains unchanged .
this conclusion together with expected effect of quantum corrections on the scale invariance ( see our discussion in the paragraph after eq.([veff=0 ] ) ) allows us to hope that the exhibited resolution of the old cc problem holds in the quantized tmt as well .
quantization of tmt being a constrained system requires developing the hamiltonian formulation of tmt .
preliminary consideration shows that the einstein frame appears in the canonical formalism in a very natural manner .
a systematic exploration of tmt in the canonical formalism will be a subject of forthcoming research .
we thank l. amendola , s. ansoldi , r. barbieri , j. bekenstein , a. buchel , a. dolgov , s.b .
giddings , v. goldstein , p. gondolo , b - l .
. hung , d. kazanas , d. marolf , d.g .
mckeon , j. morris , v. miransky , h. nielsen , y.jack ng , h. nishino , e. nissimov , s. pacheva , l. parker , r. peccei , m. pietroni , s. rajpoot , r. rosenfeld , v. rubakov , e. spallucci , a. vikman , a. vilenkin , s. wetterich and a. zee for helpful conversations on different stages of this research .
our special gratitude to l. prigozhin for his help in implementation of numerical solutions .
variation of the measure fields @xmath20 with the condition @xmath36 leads , as we have already seen in sec.ii , to the equation @xmath334 where @xmath27 is now defined , according to eq .
( [ s ] ) , as the part of the integrand of the action ( [ totaction ] ) coupled to the measure @xmath16 .
equation ( [ varphi ] ) in the context of the model ( [ totaction ] ) reads ( with the choice @xmath335 ) : @xmath336e^{\alpha\phi /m_{p } } -v_{1}e^{2\alpha\phi /m_{p } } = m^{4 } , \label{app1}\ ] ] it can be noticed that the appearance of a nonzero integration constant @xmath337 spontaneously breaks the scale invariance ( [ st ] ) .
variation of the action ( [ totaction ] ) with respect to @xmath338 yields @xmath339 = 0.\label{app2}\ ] ] we see that in contrast to field theory models with only the measure @xmath4 , in tmt there are two independent equations containing curvature .
contracting eq.([app2 ] ) with @xmath338 and solving eq.([app1 ] ) for @xmath340 we obtain the following _ consistency condition _ of these two equations : @xmath341 that we will call _ the constraint in the original frame_. it follows from eqs.([app1 ] ) and ( [ app2 ] ) that @xmath342 \label{app4}\end{aligned}\ ] ] the scalar field @xmath0 equation of motion in the original frame can be written in the form @xmath343 \nonumber\\ & -&\frac{\alpha}{m_{p}}e^{\alpha\phi /m_{p}}\left[(\zeta
+ b_g)m^{4}e^{-\alpha\phi /m_{p}}+ [ ( b_g-\zeta)v_{1}-2v_{2}]e^{\alpha\phi /m_{p } } -(b_g - b_{\phi})\frac{1}{2}g^{\alpha\beta}\phi_{,\alpha}\phi_{,\beta}\right ] = 0 \label{phi - orig}\end{aligned}\ ] ] where eq.([app1 ] ) has been used .
variation of the action ( [ totaction ] ) with respect to the connection degrees of freedom leads to the equations we have solved earlier@xcite .
the result is @xmath344 where @xmath345 are the christoffel s connection coefficients of the metric @xmath19 and @xmath346 .
the results obtained in secs.iv and v depend very much on the choice of the parameters @xmath60 , @xmath61 and @xmath192 in the action ( [ totaction ] ) .
let us recall that the curvature term in the action ( [ totaction ] ) couples to the measure @xmath347 while the @xmath0 kinetic term couples to the measure @xmath348 .
this is the reason of @xmath106 .
if we were choose the fine tuned condition @xmath349 then both the curvature term and the @xmath0 kinetic term would be coupled to the same measure @xmath350 .
one can also pay attention that depending on the choice of one of the alternative conditions @xmath168 or @xmath111 we realize different shapes of the effective potential if @xmath107 ( see fig.1 ) . and again , if instead we were choose the fine tuned condition @xmath131 then the action would contain only one prepotential coupled to the measure @xmath347 .
so , in order to avoid fine tunings we have introduced asymmetries in the couplings of the different terms in the lagrangian densities @xmath27 and @xmath15 to measures @xmath16 and @xmath4 . in order to display the role of these asymmetries
it is useful to consider what happens if such asymmetries are absent in the action at all . in other words
we want to explore here the gravity+dilaton model where both @xmath105 and @xmath131 .
in such a case the action contains only one lagrangian density coupled to the measure @xmath350 : @xmath351 where @xmath352 .
an equivalent statement is that @xmath353 ; it is an example of the very special class of the tmt models where @xmath27 is proportional to @xmath15 . to see the cosmological dynamics in this model one can use the results of sec.iiib .
if we assume in addition @xmath129 and @xmath130 , then after the shift @xmath354 where @xmath355 ( which is not a shift symmetry in this case ) , the effective potential ( [ veffvac - delta=0 ] ) takes the form @xmath356 in contrast to general cases ( @xmath357 ) this potential has no flat regions and it is symmetric around a certain point in the @xmath0-axis .
this form of the potential ( with an additional constant ) has been used in a model of the early inflation@xcite .
although we have assumed in the main text that @xmath27 and @xmath15 are @xmath20 independent , a contribution equivalent to the term @xmath358 can be effectively reproduced in the action ( [ s ] ) if a nondynamical field ( lagrange multipliers ) is allowed in the action . for this purpose
let us consider the contribution to the action of the form @xmath359d^{4}x \label{l - m - action}\ ] ] where @xmath360 is an auxiliary nondynamical field and @xmath361 is an analytic function .
varying @xmath360 we obtain @xmath362 that can be solved for @xmath360 : @xmath363 where @xmath364 is the inverse function of @xmath365 . inserting this solution for @xmath360 back into the action ( [ l - m - action ] ) we obtain @xmath366 where the auxiliary field has disappeared and @xmath367 to see the difference between effect of this type of auxiliary fields as compared with a model where the @xmath360 field is equipped with a kinetic term ,
let us consider two toy models including gravity and @xmath360 field : one - without kinetic term @xmath368d^{4}x \label{toy-1}\ ] ] and the other - with a kinetic term @xmath369d^{4}x \label{toy-2}\ ] ] where @xmath370 is a real constant .
for both of them it is assumed the use of the first order formalism .
the first model is invariant under local transformations @xmath371 , @xmath372 , @xmath373 where @xmath374 is an arbitrary space - time function while in the second model the same symmetry transformations hold only if @xmath374 is constant .
variation of the measure fields @xmath20 in the model ( [ toy-2 ] ) leads ( if @xmath36 ) to @xmath375 where @xmath337 is the integration constant . on the other hand varying the action ( [ toy-2 ] ) with respect to @xmath338 gives @xmath376 where @xmath377 .
the corresponding equations in the model ( [ toy-1 ] ) are obtained from ( [ toy-1-phi ] ) and ( [ toy-1-gmunu ] ) by omitting the terms with gradients of @xmath360 .
it follows from eqs.([toy-1-phi ] ) and ( [ toy-1-gmunu ] ) that @xmath378 this result holds in both models . in the model ( [ toy-1 ] ) , variation of @xmath360 results in @xmath379 which is consistent with eq.([toy - constr ] ) only if the integration constant @xmath380 .
this means that through the classical mechanism displayed in tmt , it is impossible to achieve spontaneous breakdown of the _ local _ scale invariance the first model possesses .
this appears consistent with arguments by elitzur@xcite concerning impossibility of a spontaneous breaking of a local symmetry without gauge fixing .
transition to the einstein frame where the space - time becomes riemannian is implemented by means of the conformal transformation @xmath381 . for the model ( [ toy-1 ] ) the gravitational equations in the einstein frame read @xmath382 this means that the model ( [ toy-1 ] ) with auxiliary ( nondynamical ) field @xmath360 intrinsically contains a constant vacuum energy . in the model ( [ toy-2 ] ) , where @xmath360 appears as a dynamical field , the gravitational equations in the einstein frame results from eq.([toy-1-gmunu ] ) @xmath383 it is convenient to rewrite this equation in terms of the scalar field @xmath384 : @xmath385 the @xmath0-equation reads @xmath386 .
similar to the general discussion in the main text we see that if the @xmath360 field is dynamical then tmt provides the vacuum with zero energy without fine tuning .
however it is very unlikely that a nondynamical scalar field will not acquire a kinetic term after quantum corrections@xcite .
then it becomes dynamical which restores the above results for the model ( [ toy-2 ] ) .
this is why we have ignored the rather formal possibility of introducing the nondynamical scalars into the fundamental action of the models studied in this paper .
a. g. riess et al .
, _ astron . j. _ * 116 * , 1009 ( 1998 ) ; s. perlmutter _
_ , _ astrophys
. j. _ * 517 * , 565 ( 1999 ) ; n. bahcall , j.p .
ostriker , s.j .
perlmutter and p.j .
steinhardt , _ science _ * 284 * , 1481 ( 1999 ) ; d.n .
spergel , _ et al . _
[ wmap collaboration ] , _ astrophys .
148 _ , * 175 * ( 2003 ) ; a. c. s. readhead , _ et al . _
_ astrophys .
j. _ * 609 * , 498 ( 2004 ) ; j. h. goldstein et al . , _
_ 599 , 773 - 785 ( 2003 ) ; r. rebolo , _ et al .
_ , astro - ph/0402466 ; m. tegmark , _ et al .
_ _ phys.rev .
_ * d69 * , 103501 ( 2004 ) ; e. hawkins _ et al . , _ _ mon.not.roy.astron.soc . _ * 346 * , 78 ( 2003 ) ; w. l. freedman , _ et al . , astrophys.j . _
* 553 * , 47 ( 2001 ) ; r. daly , s.g .
djorgovski , _
_ * 597 * , 9 , ( 2003 ) .
v. sahni , a.a .
starobinsky int.j.mod.phys.d9:373-444,2000 ; p.j.e .
peebles and b. ratra , rev .
phys . * 75 * , 559 ( 2003 ) ; t. padmanabhan , phys . rep . * 380 * , 235 ( 2003 ) ; n. straumann , hep - ph/0604231 ; t. padmanabhan , astro - ph/0603114 .
c. wetterich , nucl .
b302 * , 668 ( 1988 ) .
b. ratra and p.j.e .
peebles , phys . rev . *
d37 * , 3406 ( 1988 ) ; p.j.e .
peebles and b. ratra , astrophys .
j. * 325 * , l17 ( 1988 ) .
r. caldwell , r. dave and p. steinhardt , phys .
lett . * 80 * , 1582 ( 1998 ) ; n. weiss , phys .
b197 * , 42 ( 1987 ) ; y. fujii and t. nishioka , phys .
* d42 * , 361 ( 1990 ) ; m.s .
turner and m. white , phys . rev .
* d56 * , r4439 ( 1997 ) ; p. ferreira and m.joyce , phys .
* 79 * , 4740 ( 1997 ) ; phys . rev . *
d58 * , 023503 ( 1998 ) .
e. copeland , a. liddle and d. wands , phys .
d57 , 4686 ( 1998 ) ; p. steinhardt , l wang and i. zlatev , phys . rev . *
d59 * , 123504 ( 1999 ) .
t. chiba , t.okabe and m yamaguchi , phys.rev.d*62 * 023511 ( 2000 ) ; c. armendariz - picon , v. mukhanov and p.j .
steinhardt , phys.rev.lett . * 85 * 4438 , 2000 ; phys.rev.d*63 * 103510 ( 2001 ) ; j.k .
erickson , r.r .
caldwell , p.j .
steinhardt , c. armendariz - picon , v. mukhanov , phys.rev.lett . * 88 * 121301 ( 2002 ) ; t. chiba , phys.rev.d*66 * 063514 ( 2002 ) .
j.a.casas , j. garcia - bellido and m. quiros , _ class . quant .
grav . _ * 9 * , 1371 ( 1992 ) ; g.w .
anderson and s.m .
carroll , astro - ph/9711288 ; d. comelli , m. pietroni and a. riotto , _ phys .
_ * b571 * , 115 ( 2003 ) .
chimento , a.s .
jakubi and d. pavon , phys .
rev . d*62 * , 063508 ( 2000 ) ; w. zimdahl , d. schwarz , a. balakin and d. pavon phys .
rev . d*64 * , 063501 ( 2001 ) ; l.p .
chimento , a.s .
jakubi , d. pavon and w. zimdahl , phys .
d*67*,083513 ( 2003 ) ; a.a .
sen and s. sen , mod .
a * 16 * , 1303 ( 2001 ) .
holden and d. wands , _ phys .
rev . _ * d61 * , 043506 ( 2000 ) ; a.p .
billyard and a.a .
coley , _ phys .
rev . _ * d61 * , 083503 ( 2000 ) ; n. bartolo and m. pietroni , _ phys .
rev . _ * d61 * , 023518 ( 2000 ) ; m. gasperini , _ phys .
_ * d64 * , 043510 ( 2001 ) ; a. albrecht , c.p .
burges , f. ravndal and c. skordis , astro - ph/0107573 ; l.p .
chimento , a.s .
jakubi and d. pavon , _ phys .
rev . _ * d62 * , 063508 ( 2000 ) ; w. zimdahl , d. schwarz , a. balakin and d. pavon _ phys .
rev . _ * d64 * , 063501 ( 2001 ) ; l.p .
chimento , a.s .
jakubi , d. pavon and w. zimdahl , _ phys .
_ * d67*,083513 ( 2003 ) ; a.a . sen and s.
sen , _ mod .
_ * a16 * , 1303 ( 2001 ) ; g.r .
farrar and p.j.e .
peebles , astro - ph/0307316 m ; axenides , k. dimopoulos , _ jcap _ * 0407 * , 010,2004 ; m. nishiyama , m. morita , m. morikawa , astro - ph/0403571 ; r. catena , m. pietroni and l. scarabello , _ phys.rev . _ * d70 * , 103526 ( 2004 ) .
kamenshchik , u. moschella and v. pasquier , phys .
b*511 * , 265 ( 2001 ) ; n. bilic , g.b .
tupper and r.d .
viollier , phys .
b*535 * , 17 ( 2002 ) ; m.c .
bento , o. bertolami and a.a .
sen , phys .
d*66 * ( 2002 ) ; r.r .
caldwell , phys.lett . b*545 * , 23 ( 2002 ) ; g.w .
gibbons , hep - th/0302199 ; p. singh , m. sami , n. dadhich , phys.rev . d*68 * 023522 ( 2003 ) ; m. sami , a. toporensky , mod.phys.lett . a*19 * , 1509 ( 2004 ) ; p.f .
gonzalez - diaz , phys.lett . b*586 * 1 ( 2004 ) ; a. vikman , phys.rev . d*71 * , 023515 ( 2005 ) ; p.f .
gonzalez - diaz , c.l .
siguenza nucl.phys .
b*697 * 363 2004 ; wayne hu , phys.rev . d*71 * 047301 ( 2005 ) ; r .- g .
cai , a. wang , jcap 0503 , 002 ( 2005 ) ; s. nojiri , s. d. odintsov , s. tsujikawa , phys.rev . d*71
* 063004 ( 2005 ) ; e. elizalde , s. nojiri , s.d .
odintsov , p. wang , phys.rev . d*71 * 103504 ( 2005 ) ; h. stefancic , phys.rev . d*71 * 124036 ( 2005 ) ; l. amendola , s. tsujikawa , m. sami , phys.lett . b*632 * , 155 ( 2006 ) ; i.ya .
arefeva , a.s .
koshelev , s.yu .
vernov phys.rev .
d*72 * , 064017 ( 2005 ) .
gibbons , phys .
b*537 * , 1 ( 2002 ) ; m. fairbairn and m.h.g .
tytgat , phys.lett.b*546 * , 1 ( 2002 ) ; s. mukohyama , phys . rev . d*66 * , 024009 ( 2002 ) ; a.v .
frolov , l. kofman , a.a .
starobinsky , phys.lett . b*545 * , 8 ( 2002 ) ; y .- s .
piao , r .-
cai , x. zhang and t .- z .
zhang , phys .
rev . d*66 * , 121301 ( 2002 ) ; t. padmanabhan and t.r .
chounhury , phys .
d*66 * , 081301 ( 2002 ) ; a. feinstein , phys.rev . d*66 * , 063511 ( 2002 ) .
l. randall and r. sundrum , phys .
lett . * 83 * , 4690 ( 1999 ) ; m. gogberashvili , int.j.mod.phys . d*11 * , 635 ( 2002 ) ; t. shiromizu , k. maeda and m. sasaki , phys . rev . d*62 * , 024012 ( 2000 ) ; r. maartens , d. wands , b. bassett and i. heard , phys .
rev . d*62 * 041301 ( 2000 ) ; h. collins and b. holdom , phys .
d*62 * , 105009 ( 2000 ) ; c. deffayet , g. dvali and g. gabadadze , phys .
rev . d*65 * , 044023 ( 2002 ) ; v. sahni and y. shtanov , jcap * 0311 * , 013 ( 2003 ) . s. capozziello , v.f .
cardone , s. carloni and a. troisi , int.j.mod.phys . * d12 * 1969 ( 2003 ) ; a. d. dolgov and m. kawasaki , phys .
lett * b573 * , 1 ( 2003 ) ; s.m.carroll , a. de felice , v. duvvuri , d.a .
easson , m. trodden and m.s .
turner , phys.rev .
* d71 * 063513 ( 2005 ) .
guendelman and a.b .
kaganovich , _ phys .
rev . _ * d53 * , 7020 ( 1996 ) ; _ mod .
lett . _ * a12 * , 2421 ( 1997 ) ; _ phys .
_ * d55 * , 5970 ( 1997 ) ; _ mod .
* a12 * , 2421 ( 1997 ) ; _ phys .
_ * d56 * , 3548 ( 1997 ) ; _ mod .
phys . lett . _ * a13 * , 1583 ( 1998 ) .
guendelman , _ mod .
lett . _ * a14 * , 1043 ( 1999 ) ; _ class .
* 17 * , 361 ( 2000 ) ; gr - qc/9906025 ; _ mod .
lett . _ * a14 * , 1397 ( 1999 ) ; gr - qc/9901067 ; hep - th/0106085 ; _ found .
_ * 31 * , 1019 ( 2001 ) ; e.i .
guendelman , _
class.quant.grav_. * 17 * , 3673 ( 2000 ) ; e.i .
guendelman , _ phys .
rev_. d*63 * , 046006 ( 2002 ) ; e.i .
guendelman , a.b .
kaganovich , e. nissimov , s. pacheva , _ phys .
_ d*66 * , 046003 ( 2002 ) ; hep - th/0210062 ; hep - th/0304269 ; e.i .
guendelman , a. kaganovich , e. nissimov , s. pacheva , _ phys.rev .
_ d*72 * , 086011 ( 2005 ) .
abbott , m.b .
wise , nucl.phys .
b*244 * , 541 ( 1984 ) ; f. lucchin , s. matarrese , phys.rev . d*32 * , 1316 ( 1985 ) ; j.d .
barrow , phys.lett . b*187 * , 12 ( 1987 ) ; j. yokoyama , kei - ichi maeda , phys.lett .
b*207 * , 31 ( 1988 ) ; a.r .
liddle , phys.lett . b*220 * , 502 ( 1989 ) ; v. muller , h.j .
schmidt , a.a .
starobinsky , class.quant.grav .
* 7 * , 1163 ( 1990 ) ; b. ratra , phys.rev . d*45 * , 1913 ( 1992 ) . | the dilaton - gravity sector of the two measures field theory ( tmt ) is explored in detail in the context of cosmology .
the model possesses scale invariance which is spontaneously broken due to the intrinsic features of the tmt dynamics .
the dilaton @xmath0 dependence of the effective lagrangian appears only as a result of the spontaneous breakdown of the scale invariance .
if no fine tuning is made , the effective @xmath0-lagrangian @xmath1 depends quadratically upon the kinetic energy @xmath2 .
hence tmt may represent an explicit example of _ the effective _ @xmath3-_essence resulting from first principles without any exotic term _ in the fundamental action intended for obtaining this result .
depending of the choice of regions in the parameter space , tmt exhibits different possible outputs for cosmological dynamics : a ) possibility of a power law inflation driven by the field @xmath0 which is followed by the late time evolution driven both by a small cosmological constant and the field @xmath0 with a quintessence - like potential .
tmt enables two ways for achieving small cosmological constant without fine tuning of dimensionfull parameters : either by a _
seesaw _ type mechanism or due to _ a correspondence principle _ between tmt and conventional field theories ( i.e theories with only the measure of integration @xmath4 in the action ) .
b ) possibility of _ resolution of the old cosmological constant problem_. from the point of view of tmt , it becomes clear why the old cosmological constant problem can not be solved ( without fine tuning ) in conventional field theories .
c ) the power law inflation without any fine tuning can end with damped oscillations of @xmath0 around the state with zero cosmological constant .
d ) there is a broad range of the parameters such that : the equation - of - state in the late time universe @xmath5 ; @xmath6 _ asymptotically _ ( as @xmath7 ) _ approaches _ @xmath8 _ from below _ ; @xmath9 approaches a cosmological constant , the smallness of which does not require fine tuning of dimensionfull parameters . |
let us assume in the following that the underlying state variables whose values determine the value of an option at exercise dates follow a known stochastic process .
generally , a stochastic process is specified by defining the state space , an index parameter ( usually the time ) and the dependence relation between the random variables . the latter is usually given in terms of a stochastic integro - differential equation containing some deterministic drift terms and random changes in addition , notably diffusion or jump processes .
memory or so - called non - markov effects may readily be included into the transition probabilities .
non - markov models have found recently considerable interest in the finance community , because the heath , jarrow and morton ( 1992 ) approach to interest rate evolution can be shown to possess non - markovian features , in general .
we assume ` smoothness ' for the transition probability density function which determines the probabilities based on a path s evolution history that the path will comes close to any element of the state space at some time in the future .
these are supposed to be continuous functions of the path variables . a discrete set of sequentially selected random points in time
is often called a chain " .
a path is the corresponding generalization for a continuous time evolution . at first
, we will restrict ourselves to options and stochastic processes which result in the following simple property of the yes - no boundary : * the _ boundary _ between yes and no regions at any given decision time @xmath20 is a simply connected hypersurface in the state space with dimension of one unit less than the dimension of the state space . " as a consequence of the assumption , there are no disconnected pieces of yes or no regions at a given time either . with the exception of some pathological cases , assumption a1
will be compatible only with option payoffs contingent on the values of the state variables at exercise date . instead of specifying these pathological stochastic processes
we simply require here : * the payoff from exercising an option is only a function of the state variables at the exercise date but not at earlier times , i.e. @xmath21 . " in general , the topology of exercise and hold regions of american options may be more complicated than allowed by our assumptions .
two examples are some types of barrier options and path - dependent options .
suppose the owner of a barrier option receives only some payoff if a security stays within some specified range of values at exercise date .
therefore the option will not be exercised for ` too small ' and ` too large ' values of the underlying which leads to two disconnected no regions .
another perhaps more interesting example of options not covered by the assumption stated above are american path - dependent options whose valuation we will return to later in this section . here the payoff depends on the values which the underlying state variables have taken in the past , e.g. on the arithmetic mean within some time interval . projected into the space of state variables at any given time yes and no region are completely scrambled .
it may be optimal to exercise or hold an option for the _ same _ present value of the state variables @xmath2 depending on the different histories of each path .
lateron , we will show that a redefinition of the state space enables us to cover the case of path - dependent options by the same algorithm .
we should also mention that the assumption a2 restricts the memory properties of the underlying stochastic process .
they should not spoil the assumed property of the yes - no boundary in the state space .
this simply means that although the future evolution of each path may depend on the past , the ` average outcome ' determining the decision whether to exercise the option at present time is not influenced by the past but of markovian nature .
assumption a1 allows us to define coordinates in the @xmath22 dimensional state space such that any element @xmath2 of the state space can be uniquely characterized by some point of the yes - no boundary @xmath23 and an additional coordinate @xmath24 @xmath25 in a peculiar way . for any path passing through @xmath2 at present time
the following inequalities hold between the payoff from exercising the option @xmath26 and the option value @xmath27 if the option is being held at least until the next exercise date : @xmath28 the content of eq .
( [ eq_coord ] ) is to provide convenient coordinates from which one can easily read off whether point @xmath2 is a boundary point ( @xmath29 ) , in the yes ( @xmath30 ) or in the no ( @xmath31 ) region .
the knowledge of all @xmath24 or just their signs is tantamount to finding the yes - no boundary itself .
unfortunately , a straightforward application of the inequalities as expressed in eq .
( [ eq_coord ] ) to determine the boundary is prohibitive . estimating the expected payoff from holding the option and comparing it to the payoff from exercising would require a ` new simulation within the simulation ' .
such a procedure becomes numerically awkward and would severely restrict the complexity of the problem to be tackled , concerning e.g. the nature of the stochastic process and the number of exercise dates .
we will proceed now to estimate the location of the yes - no boundary based on a set of sampled paths . for this purpose
it will be necessary to evolve the state variables backward in time .
this feature of the novel monte carlo approach which it shares with all other numerical approaches to option pricing like the binomial tree method is caused by the boundary conditions of the problem .
the option price is a specified function of the state variables only on the expiration date .
of course , the location of the yes - no boundary is also trivially determined at expiration time . for a simple call or put option on a single asset it is the point @xmath32 .
the continuous time process will be presented as a sequence of time steps @xmath33 which add up to the lifetime of the contract @xmath34 .
the time steps are supposed to match the distance between decision times exactly if a finite number of exercise opportunities has been specified in the option contract . for american options which give the owner the right of exercise at any time this is an approximation . however , choosing ` small ' time steps @xmath35 the results will be close to the continuum limit .
the current time @xmath3 for which the location of the yes - no boundary is being sought can be represented by the integer index @xmath36 ( with @xmath37 for american options ) .
since the location of the yes - no boundary will be determined step by step moving backward in time , it is implicitly assumed from now on that the yes - no boundary has already been determined for later times @xmath36 + 1,@xmath36 + 2, ,@xmath38 ( within the usual uncertainties due to the numerical imprecision ) .
therefore we are able to evaluate for each path its payoff from optimal exercise at _ later _ times : @xmath39 the path - dependent payoff has been discounted to time @xmath40 in eq .
( [ calg_i+1 ] ) .
@xmath41 represents the earliest time ( under the constraint @xmath42 ) at which path @xmath18 crosses the boundary from the no region into the yes region or equals @xmath38 if the path stays within the no region for all times between @xmath3 and @xmath11 .
( for the latter case @xmath43=0 , of course . )
we have used the subscript @xmath44 for the argument @xmath18 of @xmath45 to emphasize that the payoff defined above depends only on the values of the path variables for times @xmath46 but not on the past evolution of the state .
so far , we are not able to calculate the discounted pathwise payoff @xmath47 , i.e. for time @xmath3 .
this would require the knowledge of the yes - no boundary at time @xmath3 . however , the path - dependent payoffs given in eq .
( [ calg_i+1 ] ) enter into the expression for the value of the option at time @xmath36 subject to the condition that the option is _ not _ immediately exercised ( and , of course , that the state variables have evolved along path @xmath18 so far ) : @xmath48 the integration over all future path segments in eq . ( [ eq_opt_price_ih ] )
@xmath49 is an ordinary multiple integral ( but a path integral in the continuous time limit ) .
@xmath50 denotes the transition probability density function and describes the probability ( density ) to evolve along the path segment @xmath51 in the future up to time @xmath11 depending on the present and previous ( for non - markov processes ) states of path @xmath52 .
next we define a continuous path @xmath53 in the state space by letting @xmath54 be a continuous parameter within some range such that all @xmath55 are elements of the state space and vice versa .
let us assume for a moment that we know already that this line intersects with the yes - no boundary but we do not know which point it is .
( somebody else has constructed the path @xmath53 . )
can we estimate the value of the yes - no boundary point using the information provided by the path sample and obtain its correct value in the limit @xmath56 ?
the answer is positive .
we are able to construct a function of the states @xmath57 whose only maximum is at @xmath58 , i.e. the boundary point , for @xmath56 .
it is clear that our ability to define such a function has important repercussions for the question how to _ find _ the yes - no boundary .
the virtue of path @xmath53 is to provide us with the knowledge that it has exactly one intersection point with the yes - no boundary .
we do not necessarily need this information .
instead , we may search for the global maximum on continuous paths connecting points in the yes region with points in the no region . according to statement a1 they will cut through the yes - no boundary at least once .
( it will not hurt the search procedure , if we accidentally hit the boundary several times reflected by several local maxima . )
it is also rather straightforward to determine selected points which are in either the yes or the no region for sure .
for instance , points lie in the no region if the payoff from exercising is zero . or the point with zero value of the underlying is certainly an element of the yes region for a plain vanilla put , because the payoff can not get any better . in order to proceed with the proof of our conjecture we define a generally suboptimal exercise policy for all paths intersecting with @xmath53 at time @xmath3 which depends on an arbitrary point @xmath59 .
we define a preliminary `` yes '' region ( with the quotation marks reminding us that the true yes boundary may be different ) as to cover all paths with state variable @xmath60 which fulfill @xmath61 . for instance , in case of a plain vanilla put ( call ) option on a single underlying asset we may declare that @xmath62 lies in the `` yes '' region if it fulfills the condition @xmath63(@xmath64 @xmath65 .
the pathwise payoff from the @xmath65-dependent policy is @xmath66 it is clear that the pathwise payoffs will sometimes be larger for ` wrong ' boundary points ( @xmath67 ) than for the correct one .
this reflects the additional information about the future which is encoded in the path segment @xmath68 .
the essential point is now to sum over all paths which removes this bias : @xmath69 the function @xmath70 depends on time , the sampled paths intersecting with @xmath53 ( of number @xmath71 ) and on @xmath65 .
our central proposition can be stated now as follows : * taking the limit of infinitely many paths , @xmath72 has one extremum which is a maximum within the one - dimensional domain @xmath53 .
its position is the intersection point of @xmath53 with the yes - no boundary @xmath65=@xmath73 . " in order to prove proposition p1 we are going to rearrange the sum in eq .
( [ eq_optim_mc ] ) .
we put all paths which have essentially the same transition probability density function into the same ( albeit infinitesimally small ) ` bin ' . here
we need the assumption stated above that the transition probability density function is a continuous function of the path variables . after ` binning ' and taking the limit @xmath74 eq . ( [ eq_optim_mc ] )
turns into @xmath75 the transition probability density function @xmath50 appears in eq .
( [ eq_optim_cont ] ) , because relative frequencies approach the probabilities which are governing the random selection of paths according to bernoulli s law of large numbers .
the probability density @xmath76 to have a path segment close to @xmath52 needs to be defined in accordance with the integration measure @xmath77 and is normalized to one . the quantity in brackets @xmath78 on the right hand side of eq .
( [ eq_optim_cont ] )
@xmath79 can be readily evaluated using eqs .
( [ eq_opt_price_ih ] ) and ( [ eq_pisss ] ) to be @xmath80 @xmath81 coincides with the option price @xmath82 , except for @xmath54 values in a certain range .
furthermore , on virtue of eq .
( [ eq_coord ] ) @xmath81 is always _ smaller _ than @xmath82 for such chosen @xmath54 values : @xmath83 the conditions of eq .
( [ eq_di ] ) together with the positiveness of the weight factors @xmath84 in eq .
( [ eq_optim_cont ] ) guarantee that @xmath85 this completes the proof of proposition p1 .
as stated above we may turn things around and construct the yes - no boundary from a set of lines in state space which connect elements of the yes and no region .
the global maximum of function @xmath86 along each line is the supposed intersection point with the yes - no boundary .
this process may be iterated until the desired precision is reached .
this standard optimization procedure becomes particularly simple in case of plain vanilla options .
since the yes - no boundary is just a point at any exercise date , its location is already completely specified with one iteration step .
the determination of the yes - no boundary at time @xmath3 ( index @xmath36 ) enables us to calculate the pathwise payoff at this time .
going step - by - step backward in time we may finally calculate the pathwise payoffs at initial time whose average determines the option value at contract time .
the presented optimization technique may be used in the problem of valuation of american _ path - dependent _ options as well . here the payoff from exercising the option depends on some function of the state variables in the past which we will denote as @xmath87 in the following .
for instance , @xmath87 may be the arithmetic or geometric average of any state variable over some time window or the value at some specified time(s ) of the past ( look - back options ) . for simplicity
, we restrict ourselves here to stochastic processes of markovian nature whose parameters have been specified . in this case
the exercise decision at any time before expiry is a function of just two each possibly vector - valued variables , the values of the state variables and of the path - dependent variable @xmath87 , both taken at the decision time .
the difference between standard path - dependent and plain vanilla options is just that the role of the current value of the underlying variable in plain - vanilla contracts is played by @xmath87 in path - dependent contracts , e.g. max@xmath88 for a put .
the topology of the exercise region for this kind of path - dependent options is very simple but only in the space of state variables _ enlarged by the additional dimension(s ) of @xmath87 _ .
the yes - no boundary in the @xmath2-@xmath87 space is again simply connected .
only its projection into the @xmath2 space leads to a scrambling of yes and no regions .
this suggests immediately how the suggested optimization technique may be applied for the case of these path - dependent options .
we equip the space of states with additional dimensions by declaring @xmath87 to be one of the stochastic variables .
of course , the stochastic character of this variable is rather funny .
its change with time is completely deterministic and of non - markovian type .
however , these features are in accordance with the basic assumptions entering into the proof of proposition p1 . the complications of such path - dependent option contracts just result in a higher dimension of the space in which we have to track the yes - no boundary .
we are now in a position to compare our novel strategy for an extraction of american - style option prices from monte carlo simulation to previously suggested ones .
boyle , broadie , and glassermann discuss mainly three strategies in their review paper to overcome the inherent limitations of the naive monte carlo procedure : the bundling algorithm invented by tilley ( 1993 ) , the stratified state aggregation algorithm suggested by barraquand and martineau ( 1995 ) and broadie and glassermann s ( 1995 ) algorithm based on simulated trees . in tilley s
approach bundles of paths which are ` close ' in state space at given time are considered to emanate from one single point for the purpose of option valuation .
this introduces some error whose magnitude can not be easily controled .
the bias of the exercise decision will be very strong if a bundle contains only few paths .
such a situation is likely to emerge in situations in which there are many exercise dates or several state variables .
barraquand and martineau introduce another type of ` bundling ' , however , in the payoff space .
our approach shares similarities with tilley s algorithm in the sense that it contains an implicit ` bundling ' of paths ( called ` binning ' ) in the proof of the proposition p1 .
however , we have actually never to carry out this bundling in a numerical calculation , because we stick to an optimization of the ` inclusive ' function @xmath89 which sums up all bins .
broadie and glassermann proposed an algorithm based on a ` bushy tree ' structure in state space which avoids partitioning of the state or payoff space . in their model
many branches emanate from each node .
this process is replicated as often as there are decision times @xmath90 in the problem .
the replication process effectively limits @xmath90 to small numbers ( on the order of 4 ) in practical computations .
they also discuss the problem that the suggested estimators of the option prices tend to be biased in ` up ' or ` down ' direction . and converge only to the correct value in the limit @xmath56 .
this problem is of great relevance for practical calculations ( which are always at finite @xmath1 ) .
they use the idea to calculate two estimators one biased low and the other high to obtain confidence limits from the simulations .
lateron , we are going to apply this idea in the framework of our approach .
in this paper we will only consider stochastic processes of rather simple nature , so - called ideal wiener or diffusion processes ( for the @xmath91 of any state variables ) . in physics these processes
are usually called brownian motion .
such processes seem to be most commonly used in valuation problems .
furthermore , numerical techniques have been developed to calculate prices for plain vanilla and other simple types of american options .
in particular , the binomial tree method may provide very accurate values and will be used as a benchmark here against which the results from the monte carlo simulations are being tested . in this section
we will restrict ourselves to _ puts _ on the current value and on the geometric or arithmetic average of one single underlying security .
an ideal wiener process can be characterized by the differential transition law @xmath92 governing the change of the state variable @xmath93 within a time interval @xmath94 .
@xmath95 is a random number drawn from a normal distribution with mean 0 and variance 1 .
@xmath96 is in general the drift velocity which may be replaced by the risk - free instantaneous interest rate @xmath13 using arbitrage arguments .
sometimes one needs to follow an ideal wiener process only _ backward _ in time , e.g. in case of monte carlo estimation of plain vanilla options . using the _ bridge _ construction start @xmath97 and final value @xmath98 are known a random value at intermediate time may be chosen according to @xmath99 after having fixed final and start values , iterative use of eq .
( [ eq_wiener_bw ] ) may be employed to generate a time reverted copy of the standard diffusive wiener - type motion . for the simulation results presented below
we have used a simple importance sampling scheme .
the exercise value of strongly out - of - the money options at expiry will be zero on most paths .
therefore the weight of the rare paths resulting in nonzero option value has been artificially enhanced in the path generation process . of course , the enhancement factors are taken out again in the calculation of the path - averaged prices . in its most simple form
the monte carlo algorithm which is suitable for valuation of american - style options consists of the following steps : ( 1 ) : : generate path sample .
( 2 ) : : go step - by - step backward in time starting at expiry . ( 3 ) : : track the yes - no boundary by employing the optimization of function @xmath86 along arbitrary paths connecting yes and no region ( cf .
( [ eq_optim_mc ] ) ) .
( 4 ) : : evaluate the pathwise payoffs at the earlier time analogous to eq .
( [ calg_i+1 ] ) .
( 5 ) : : repeat the process until the contract time is reached . ( 6 ) : : average the pathwise payoffs discounted to initial time over the whole path sample ( cf .
( [ eq_amopt_mc ] ) ) .
let us discuss now some aspects related to the _ finite _ size of the path sample in practical calculations .
it is clear that step * ( 6 ) * does not provide an unbiased estimator of the option price at any finite @xmath1 but is biased upward .
the reason is that an error is made extracting the yes - no boundary from the finite size path sample .
a deviation of the extracted from the true boundary has its root in the information about the mismatch between frequency distribution of the @xmath1 paths and underlying probabilities which enters into the optimization process .
the estimator becomes only asymptotically unbiased .
since we are not able to construct an unbiased estimator , we may get an estimate of the error introduced by the bias in step * ( 6 ) * by constructing an estimator which is downward biased at finite @xmath1 but also asymptotically unbiased .
it is very simple to construct such an estimator : ( 6 ) : : generate a path sample independent from the first one and evaluate the american option prices using the previously calculated yes - no boundary .
the reason that the prices calculated from this path sample are biased downward mirrors the one for the upward bias in the former case . here
any deviation from the true boundary implies a suboptimal exercise policy .
we hasten to add that the algorithm presented so far has one shortcoming related to the combined effects of finite sample size @xmath1 and well - determined values of the underlying variables at start time . only with paths crossing the yes - no boundary
its optimal location can be determined , of course .
formally , the proof of proposition p1 requires that the probability densities appearing in eq .
( [ eq_optim_cont ] ) be positive - definite in a region covering the yes - no boundary .
however , this condition will be violated in practice at early times , since all sampled paths emanate from a common starting point . on the other side , in such a situation
the knowledge of the precise location of the boundary becomes irrelevant for the valuation problem .
it suffices to know at this stage on which side of the boundary the vast majority of paths can be found .
two variants with differing degree of sophistication have been developed to determine the earliest time for which the algorithm above is to be used : ( 7a ) : : if the yes - no boundary gets closest to either one of the endpoints of @xmath53 , e.g. the paths with smallest or largest @xmath2 value for plain vanilla options , it is assumed that the boundary can be found outside the range of the path sample for all earlier times .
( 7b ) : : `` flashlight mechanism '' : when the yes - no boundary moves towards a region only poorly covered by the original path sample , additional path segments are created which evolve in the region of the yes - no boundary at this time .
as far as the proof of proposition p1 is concerned the probability densities in eq .
( [ eq_optim_cont ] ) do _ not _ need to be consistent with the initial value specification for the pricing problem . here
we make use of this freedom to generate a set of paths whose initial values are tuned to ` shed light ' on the location of the yes - no boundary .
concerning accuracy simulation mode * ( 7b ) * is only barely noticably superior over mode * ( 7a ) * for the tested sample of plain vanilla options .
all simulation results for more complicated options reported on in this paper have been achieved employing mode * ( 7b ) * . at
finite @xmath1 the values and therefore also the maximum of the payoff function @xmath86 do not change between values of its argument which are taken by two neighboring paths . furthermore
, the precision with which the true maximum of function @xmath100 can be estimated from a finite size sample is a multiple ( @xmath101 ) of the typical distance between paths in the region along the yes - no boundary .
[ fig_1 ] shows the monte carlo estimation of the optimization function @xmath70 for a plain vanilla put option on a security of ( time - dependent ) value @xmath102 in order to illustrate the finite @xmath1 effect .
selected @xmath70 values are displayed at a fixed time but varying the size of the path sample .
the true minimum is quite shallow for the selected time as can be seen from the figure .
the shallowness is related to the strong time dependence of the location of the yes - no boundary , because we consider a time shortly before expiry .
the location of the maximum can be reliably extracted from the simulations only for rather large samples on the order of @xmath103 paths .
one may ask whether in view of the imprecisions associated with any finite sample size it is worthwile to search for the exact maximum of @xmath70 .
in fact , a well - chosen ` smoothing ' procedure which subtracts the white random noise may give more accurate results than taking the real maximum of @xmath70 as a boundary point .
for this paper we compared two variants how to determine the yes - no boundary points : ( 3a ) : : determine the yes - no boundary point as the element of the set of all intersection points @xmath62 of the sampled paths with path @xmath53 for which @xmath104 restricted to these points attains a maximum value @xmath105 .
( 3b ) : : determine the yes - no boundary point as the element of the set of all grid points @xmath106 for which @xmath104 restricted to these points attains a maximum value @xmath107 .
the distance between sites on the grid is lowered stepwise until some pre - defined precision is reached or more than one local extremum appears for the set of values of @xmath104 restricted to the grid sites .
ultimately , in the limit of infinitely many paths both variants may give the same correct result for the optimization of @xmath104 .
however , computation employing variant * ( 3b ) * is much quicker .
it requires a number of operations growing only linearly with @xmath1 while variant * ( 3a ) * needs at least on the order of @xmath108 operations .
we have generated monte carlo estimates for prices of randomly sampled standard american - style put options ( see table for details ) .
we have employed the two simulation modes * ( 3a ) * and * ( 3b ) * to compare their effectiveness .
the frequency distribution @xmath109 of the relative errors @xmath110 in the simulation results based on monte carlo mode * ( 3a ) * are shown in fig .
[ fig_2 ] , for mode * ( 3b ) * in fig .
[ fig_3 ] .
these errors are determined from the normalized difference to the prices extracted from binomial - tree calculations : @xmath111 the monte carlo estimates have been calculated for the path sample which has been used to determine the yes - no boundary and for an independently generated path sample .
the comparison between the results in the two modes reveals that both lead to similar accuracy .
in addition , the figures display the distribution of the differences between monte carlo values based on the original and the independent path sample .
upward and downward bias are clearly revealed for the simulation results in mode * ( 3a ) * ( lower left panel of fig .
[ fig_2 ] ) . on the other side ,
the fact that mode * ( 3b ) * employs only an approximate optimization of the payoff function @xmath104 tends to wash out most of the up- and downward biases .
[ fig_3 ] shows also the error distribution of the corresponding european option estimations .
we note that the typical size of the errors is comparable to the results for american puts .
the error in both cases is thus dominated by the finite size of the path sample ( @xmath112 in figs .
[ fig_2 ] and [ fig_3 ] ) . the error induced by the bias in the construction of the yes - no boundary at finite @xmath1 seems to be of less importance .
let us turn now to the case of put options contingent on the geometric or arithmetic mean value of the underlying variable . as discussed in the preceeding section exercise and hold regions and therefore the yes - no boundary are simply connected in the @xmath113 space .
the boundary is therefore one - dimensional in this space .
we have chosen a very simple procedure in order to track the boundary .
these options are not directly contingent on the current @xmath102 value at any of the decision times which makes @xmath87 the more relevant variable
. therefore we sort the sampled paths into bins covering the relevant @xmath2 region . ignoring the differences between @xmath2 values _ inside _ each bin
we have reduced the optimization problem again to finding the @xmath87 value at which the payoff function @xmath104 attains maximum value .
this optimization proceeds as described already for plain vanilla options . for the calculations described here we have taken the number of bins in @xmath2
direction to be 20 .
monte carlo simulation results for a randomly generated sample of put options contingent on geometric averages are presented in fig .
[ fig_4 ] . as before ,
errors in the simulation have been calculated in case of geometric averaging by comparing the monte carlo estimations for the option prices to binomial tree results .
here we consider only averages over the whole lifetime of the option .
this is an @xmath114 problem for the binomial - tree method , @xmath115 being the number of time steps . allowing the time window for the averaging to slide would increase the storage and cpu time requirements by one power of @xmath115 .
in contrast , the monte carlo approach is not impacted severely by such a generalization .
a comparison with the corresponding errors of the simulation results for european options under the same conditions ( not shown ) reveals again that the errors are dominated by the finite size of the path sample and not by the uncertainty in determining the yes - no boundary . the effect of upward and downward bias in the estimations due to choosing either the original or the independently generated path sample are very weak as can be seen from the lower left panel of fig . [ fig_4 ] .
as explained before this is a direct result of the approximations in the optimization procedure * ( 3b ) * .
it is noteworthy that the bias from parametrizing the yes - no boundary on a grid ( in @xmath2 direction ) is stronger than either bias from taking one of the two path samples .
the representation of the yes - no boundary on a grid tends to _ underestimate _ the option prices because of the coarse graining procedure involved .
the small nonstatistical downward net bias can be read off from the average deviation between monte carlo estimation and binomial tree results for the options contingent on geometric averages .
it amounts to - 0.24 % for the results based on the original path sample and - 0.1 % for the independent path sample .
results for prices of options contingent on arithmetic averages over the whole lifetime of the option are also included in fig . [ fig_4 ] ( lower right panel ) . since no other accurate method is available , we present the distribution of relative deviations between the values based on an approximate formula suggested by ritchken , sankarasubramanian , and vijh ( 1989 ) and the simulation results .
the approximate formula relates the price for a claim contingent on the arithmetic average to the corresponding option price using geometric averaging via @xmath116 here @xmath117 ( @xmath118 ) denotes `` value of american ( european ) option '' .
the subscript @xmath119 ( @xmath120 ) refers to arithmetic ( geometric ) averaging .
the superscripts @xmath121 and @xmath122 characterize the method of calculation , binomial tree and monte carlo respectively .
[ fig_4 ] shows that the approximation works reasonably well .
we have shown in this paper how to construct monte carlo algorithms for american option valuation .
we are in the lucky situation that the suggested algorithm turns out to be conceptually simpler than the other algorithms suggested so far .
it depends only linearly on the number of sampled paths and exercise dates which makes it computationally feasable to get close to the continuum limit for these two variables .
thus no price is to be paid for the accuracy of this novel monte carlo algorithm in terms of more cpu time or exceedingly large storage requirements .
the crucial element , the optimization of a certain payoff function of the path sample , turns out to be numerically rather stable .
the tracking of the yes - no boundary is achieved most precisely in those regions of the state space covered densely by sampled paths , i.e. where it is needed for the option price estimation .
there is a further reason for the algorithm s stability . in case
a maximum turns out to be shallow and is therefore not easily found by the algorithm , the option price does not depend much on the exact location of the boundary between early - exercise and hold region .
indeed , we have demonstrated that the errors in the simulation results are typically dominated by the statistical errors if the size of the path sample is on the order of @xmath103 .
this points to one important direction of future research .
paskov and traub ( 1995 ) have shown that the replacement of pseudo - random numbers in the traditional monte carlo approach by a series of so - called quasi - random numbers may lead to a spectacular gain in the precision of the achievable results , at least for some problems .
the reason is that the quasi - random numbers are distributed more evenly . by using appropriately chosen sequences of quasi - random numbers sample sizes for estimation of european options
could be reduced by orders of magnitude , without loss of accuracy .
it would open new dimensions for the applications of the novel monte carlo algorithm if this would also hold true for american options .
moreover , we would like to add that the extraction of sensitivity coefficients usually called the `` greeks '' employing the novel monte carlo algorithm is as straightforward as for previously considered situations , e.g. by fu and hu ( 1995 ) .
we have presented some simulations for option types for which other means of calculation are known .
however , the monte carlo algorithm presented here provides unique opportunities to evaluate path - dependent option prices for which no other methods are available or computationally feasible .
one example , the case of options contingent on arithmetic averages has been discussed already in this paper .
another candidate are `` look - back '' options of american style .
usually , look - back options allow the owner to buy or sell an asset for a price dependent on the values of the asset during the whole lifetime until expiry , in the simplest case the minimum or maximum .
restricting the time window for the look back but allowing for early exercise adds american features to these rather common options .
another virtue of the presented monte carlo algorithm is to allow option valuation for more complicated stochastic processes than wiener - type diffusion .
p. boyle , m. broadie , and p. glassermann , 1997 , monte carlo methods for security pricing , in risk publication group , ed .
: pre - course material for risk@xmath123 training course ( january 1997 ) ` monte carlo techniques for effective risk management and option pricing ' , ( risk publication group , london ) .
caption : mean values of riskless interest rate @xmath13 , volatility @xmath124 , start price @xmath125 , strike price @xmath126 and time until expiry @xmath11 are given in first row , their standard deviations respectively in second row .
the random values are chosen according to independent normal distributions . in a second step
, some parameter values may be rejected and replaced by other random values if they are nonpositive .
furthermore , strike price @xmath126 is required to be within a 2 ( 1 ) @xmath127 range around the mean @xmath128 value at maturity @xmath129 for options contingent on the current ( average ) value .
this constraint could be relaxed using a more sophisticated importance sampling scheme .
the number of evenly spaced time steps @xmath130 for the monte carlo simulation and the binomial - tree calculations has been fixed arbitrarily to be equal 100 .
monte carlo estimation of the optimization function @xmath70 for a plain vanilla put option . option and
stochastic process parameters are chosen as in hull s book ( 1993 ) , example 14.1 . the monte carlo simulation with number of time steps equal
50 has been repeated varying the total number of sampled paths : @xmath131 , @xmath132 , @xmath133 and @xmath103 . only the function values at time step 45 are displayed for 100 values of the argument @xmath65 .
the location of the yes - no boundary according to a binomial - tree calculation is pointed at by an arrow .
frequency distribution of the relative errors in simulation estimates of a sample of plain put option prices .
the error is determined from the normalized difference to the prices @xmath134 extracted from binomial - tree calculations .
the monte carlo estimates are calculated for the path sample which has been used to determine the yes - no boundary ( biased up ) as well as for an independently generated path sample ( biased down ) and the average of the two samples ( mc average ) .
in addition , the figure displays the distribution of the difference between upward and downward biased monte carlo values ( lower left panel ) .
each path sample encompasses @xmath103 generated paths per option .
the monte carlo simulation employed the mode in which the ` global maximum ' of the optimization function @xmath70 was searched .
the content of this figure is the same as in the previous figure , with the exception of two features .
the mode of the monte carlo simulation has been to look up the maximum of the optimization function on a grid ( see text ) .
furthermore , the distribution of errors for the averaged monte carlo values which looks very similar to the corresponding distributions of upward and downward biased estimates is not displayed . instead , the distribution of the errors for the corresponding _
european _ option values is displayed ( in the lower right panel ) .
frequency distribution of deviations between simulation estimates of put option prices on geometric ( left side ) and arithmetic average values ( right side ) and benchmark calculations .
the benchmark for the prices on geometric averages is provided by the rather accurate binomial - tree method , and for the arithmetic averages by the approximate formula eq .
( [ eq_approx ] ) . | it is shown how to obtain accurate values for american options using monte carlo simulation .
the main feature of the novel algorithm consists of tracking the boundary between exercise and hold regions via optimization of a certain payoff function .
we compare estimates from simulation for some types of claims with results from binomial tree calculations and find very good agreement .
the novel method allows to calculate so far untractable path - dependent option values .
21.5 cm -1.5 cm * valuation of path - dependent american options + using a monte carlo approach + * + h. sorge + department of physics and astronomy , + state university of new york at stony brook , ny 11794 - 3800 pricing of options is an important area of research in the finance community .
the field has been pioneered by black and scholes ( 1973 ) .
they made a major breakthrough by deriving a formula for the price of any contingent claim which depends on a non - dividend - paying stock . using the assumption of no arbitrage
, they were able to show that the price of a derivative security can be expressed as the expected value of its discounted payoffs .
the expectation is taken under the assumption of a _ risk - neutral _ evolution of the value of the underlying security .
merton ( 1973 ) generalized these ideas to situations in which the interest rates themselves fluctuate in time .
due to the complexity of the underlying dynamics , numerical methods have become increasingly popular in modern finance .
they are used for a variety of purposes , for instance valuation of securities and stress testing of portfolios .
analytical solutions for problems in finance have been found only for rather special cases .
for instance , in order to gain the solution for prices of european options black and scholes had to assume that the evolution of the underlying asset price ( more precisely its natural logarithm ) follows a so - called wiener process with time - independent volatility .
european options can be exercised only at expiration date .
in contrast , american - style options which can be exercised during the whole lifetime of the option can be valued only numerically .
furthermore , numerical methods have to be employed if the dimension of the problem increases ( more than one state variable ) or in case of more realistic approaches to the stochastic process like for some no - arbitrage models of interest rate evolution .
the monte carlo approach lends itself very naturally to the evaluation of security prices and interest rates .
schematically , it consists of the following steps . first , sample paths of the state variables ( asset prices and so on ) over the relevant time horizon are generated .
the cash flows of the securities on each sample path are evaluated , as determined by the structure of the security in question .
the discounted cash flows are averaged over the sample paths .
the monte carlo method is very flexible , since it does not depend much on the specific nature of the underlying stochastic process .
its accuracy is also independent on the dimensionality of the problem which is its dominant advantage over more traditional numerical integration methods .
it is outside the scope of this paper to discuss the various facets of recent research on the use of monte carlo in the finance area .
the reader may consult the concise recent reviews of boyle , broadie , and glassermann ( 1997 ) on these topics .
in particular , they report on recent progress in developing more _ efficient _ monte carlo algorithms .
standard monte carlo methods converge notoriously slow ( with 1/@xmath0 , @xmath1 being the number of sampled paths ) .
a central goal of recent research activity has therefore been the refinement of variance reduction techniques ( antithetic and control variates ) , importance sampling and low - discrepancy random number sequences .
if the development in sciences like physics is any guide , monte carlo applications in finance will become even more important . with computational prowess further increasing the monte carlo method may even shed its ` brute force ' image from the past which was founded on its rather slow convergence property .
boyle ( 1977 ) was the first to propose the use of monte carlo methods for option pricing in the literature .
since the exercise date of european contingent claims is fixed , the mechanics of a price evaluation of european options employing the monte carlo method is rather straightforward .
let us begin by collecting the notation for the most important variables which we are going to use throughout the rest of the paper .
in general , we follow the notation in hull s book ( 1993 ) , except that the present time for which the option price is calculated is designated to be `` 0 '' ( without loss of generality ) .
@xmath2 = the value of the state ( asset price , ) at time @xmath3 + underlying the derivative security + ( which may be a single variable or a vector ) , + @xmath4 = the initial value of the state variables , + + x = strike price of the option , + p = path on which the underlying state evolves in time , + @xmath5 = the value of the option _ if _ exercised at time @xmath3 + ( @xmath6 for a simple call and + @xmath7 for a put ) , + @xmath8 = instantaneous risk - less interest rate at time @xmath3 , + @xmath9 = volatility ( standard deviation ) of state variable + or their components respectively , + t = the lifetime of the option contract ( expiration date ) , + @xmath10 = the option value at time @xmath3 .
we may easily estimate the european option value from a monte carlo path sample for the variables related to the derivative security .
these paths need to be generated with probabilities determined by the underlying stochastic process .
the option value at the expiration date @xmath11 has been specified in the option contract .
in general , it may depend on the path of the relevant state variables between closing the contract and expiration date .
each pathwise final option value from the simulation is discounted backward in time to determine its value at time 0 .
since we do not know at time @xmath3=0 yet on which path the underlying variables will evolve , the average over all paths is taken to estimate the present option value : @xmath12 the discount factor includes the risk - free interest rate @xmath13 , possibly averaged over the lifetime @xmath11 of the option .
for simplicity , it is assumed in eq .
( [ eq_europt_mc ] ) that all sampled paths have equal probability .
generalization to a situation in which paths are characterized by different probabilities which may arise e.g. in the context of monte carlo optimization by importance sampling is straightforward . the pricing of options which may be exercised prior to the expiration date by monte carlo simulation is more involved .
here , the owner holds the right to exercise the option at several ( bermudan ) or possibly infinitely many ( american ) ` decision dates ' .
many types of american contingent claims trade on exchanges and in the over - the - counter market .
examples include options , swaptions , binary options and asian options .
it has been shown by roll ( 1977 ) , geske ( 1979 ) , and whaley ( 1981 ) that exercise of _ call _ options which give the owner the right to purchase some underlying asset at the agreed - on strike price is usually unfavorable before expiration date , except close to ex - dividend dates .
the situation is much less clear - cut in case of _ put _ options which give the owner the right to sell the asset at the strike price . in these situations
the ` naive ' monte carlo is encountering unsurmountable difficulties . in order to decide whether it is more favorable at some intermediate time to hold or exercise the option the owner needs to compare the expected payoff in the two cases .
the maximum of the two forms the option value at that time which after discounting and taking the expectation value leads to the following expression for the option at closing date : @xmath14 the angular brackets denote the expectation value with respect to the ( risk - free ) probability measure . the path - dependent times @xmath15 are the so - called optimum stopping times which may be any of the decision dates .
a single path provides clearly insufficient information to evaluate the option value in case of non - exercise at any of the decision dates .
the insufficiency of the naive monte carlo method to deal with the optimum stopping problem has lead some authors like hull ( 1993 ) to the claim in the literature that monte carlo simulation can only be used for european - style options " . on the other side , evaluation of american - style options via
monte carlo simulation has found already some consideration in the literature as reviewed boyle , broadie , and glassermann ( 1997 ) .
their common denominator is a ` clever ' estimate of the option prices on the decision dates by bundling subsets of paths .
these authors noted that the suggested algorithms can not be considered satisfactory yet .
either some approximations may lead to uncontroled errors affecting the simulation results or the required large computational effort limits their applicability , e.g. to a small number of exercise dates . in this paper
we are suggesting a completely different strategy to calculate american - style put and call option prices via monte carlo simulation .
during the sampling of the paths we do _ not _ attempt to estimate the expectation of the payoff if the owner continues to hold the option . instead , we use the sampled paths to evaluate the boundary between the early - exercise region and the hold region in the space of variables entering into the option contract .
the location of this boundary is the crucial piece of information whose knowledge allows the straightforward use of the monte carlo procedure for option price estimation .
we treat the valuation of american options as an _ optimization problem _ of a certain payoff function which depends on the set of sampled paths .
this function depends also on the exercise policy , i.e. the boundary between the early - exercise region and the complementary hold region ( yes - no boundary in the following ) .
maximation of the payoff function provides an estimator of the yes - no boundary which gets arbitrarily close to the true location of the boundary with increasing number of sampled paths .
for the plain vanilla put and call options , the boundary is rather simple at any exercise date , a point for one state variable , a line for a two - dimensional space of state variables and so on .
after the boundary has been estimated the price estimation proceeds as for european options , because the option prices for points on the boundary are known . the only difference to price estimation of european options as in eq .
( [ eq_europt_mc ] ) is that concerning paths crossing the yes - no boundary each pathwise option value is discounted starting from the point at which the path crosses the boundary for the first time : @xmath16 time @xmath17 denotes the earliest time at which the @xmath18-th path crosses the yes - no boundary ( if at all ) .
otherwise @xmath17 equals the time until expiration of the option ( @xmath11 ) .
@xmath19 is the agreed - on payoff from exercising the option at time @xmath17 under the assumption that path @xmath18 represents the realized evolution of the underlying state variables . here
we have tacitly assumed that the starting point lies in the no region .
of course , the algorithm is able to handle the other possibility immediate exercise as well . in this case calculation of
the option price is trivial , however . in a later section of this paper
we will demonstrate that the novel monte carlo algorithm achieves an accuracy in the determination of american option values which compares well with corresponding calculations for european options .
we compare the results for a spectrum of standard american puts with results from binomial tree calculations .
this method originally introduced by cox , ross , rubinstein ( 1979 ) is widely used for option valuation . here
we assume an ideal wiener - type stochastic process for a single state variable , because a binomial tree is well suited to provide very accurate results in this situation .
of course , this comparison serves mostly illustrative purposes to show that the method works at finite @xmath1 .
lateron , we are going to present numerical results for _ path - dependent _ american options .
we consider two cases , puts on either the geometric or the arithmetic average over the lifetime of the option .
assuming again an underlying ideal wiener process we may compare the monte carlo estimations to results from binomial tree calculations . while options on the geometric average can be priced by the tree method to arbitrary accuracy
no such tree method is available for options contingent on the arithmetic average value .
some approximative method has been suggested to express the latter option prices by the corresponding prices in case of geometric averaging .
therefore we will compare the approximative formula to our more accurate procedure .
in general , it is envisioned that the novel monte carlo approach and some variants will be applied for valuation problems which can not or not easily be tackled by other methods .
such problems encompass stochastic processes with non - markov properties , stochastic volatilities , multi - dimensional state variables and other path - dependent american options . |
over the last ten years there has been tremendous progress in our understanding of the ads / cft correspondence @xcite in the presence of unbroken supersymmetry .
we have witnessed the discovery of many highly non - trivial supersymmetric solutions of supergravity , together with a rather detailed understanding of their gauge theory duals .
supersymmetric solutions with an anti - de sitter ( ads ) factor are particularly important , as they are dual to _ superconformal _ field theories , in an suitable limit .
comprehensive studies of general supersymmetric ads geometries , in different dimensions , have been carried out in @xcite and led to a number of interesting developments .
these results have all been obtained using the technique of analysing a canonical @xmath9-structure in order to obtain necessary and sufficient conditions for supersymmetry @xcite . in this paper
we will systematically study the most general class of @xmath0 ads@xmath5 solutions of @xmath1 supergravity .
supersymmetric ads@xmath5 solutions of @xmath1 supergravity have been discussed before in the literature @xcite . however , these references contain errors , and reach incorrect conclusions that miss important classes of solutions .
our main motivation for studying ads@xmath5 solutions of @xmath1 supergravity in particular is that , starting with the seminal work of @xcite , over the past few years there has been considerable progress in understanding the ads@xmath5/cft@xmath10 correspondence in m - theory . in particular , with @xmath11 supersymmetry there is good control on both sides of the duality , and this has led to many new examples of ads@xmath5/cft@xmath10 dualities , including infinite families , along with precise quantitative checks . on the gravity side , the simplest setup is that of freund - rubin ads@xmath12 backgrounds of m - theory , where @xmath13 is a sasaki - einstein manifold include three - sasakian manifolds and orbifolds of the round seven - sphere . ] .
these are conjectured to be dual to the theory on a large number @xmath14 of m2-branes placed at a calabi - yau four - fold singularity .
rather generally , these field theories are believed to be strongly coupled chern - simons - matter theories at a conformal fixed point . with @xmath15 supersymmetry the partition function of such a theory on
the three - sphere localizes @xcite , reducing the infinite - dimensional functional integral exactly to a finite - dimensional matrix integral .
this can then often be computed exactly in the large @xmath14 limit , where @xmath14 is typically related to the rank of the gauge group , and compared to a gravitational dual computation which is purely geometric .
such computations have now been performed in a variety of examples @xcite , with remarkable agreement on each side . thus far , almost all attention has been focused on ads@xmath16 solutions .
this is for the simple reason that very few ads@xmath5 solutions outside this class are known .
an exception is the corrado - pilch - warner solution @xcite , which describes the infrared fixed point of a massive deformation of the maximally supersymmetric abjm theory on @xmath14 m2-branes in flat spacetime .
this solution is topologically ads@xmath4 , but the metric on @xmath17 is not round , and there is a non - trivial warp factor and internal four - form flux on the @xmath17 .
this has more recently been studied in @xcite , and in particular in the last reference the free energy @xmath18 of the @xmath0 superconformal fixed point was shown to match the free energy computed using the gravity dual solution .
the corrado - pilch - warner solution also has a simple generalization to massive deformations of @xmath14 m2-branes at a cy@xmath19 four - fold singularity , where cy@xmath10 denotes an arbitrary calabi - yau three - fold cone singularity . in this paper
we systematically study the most general class of @xmath0 ads@xmath5 solutions of m - theory .
these have an eleven - dimensional metric which is a warped product of ads@xmath5 with a compact riemannian seven - manifold @xmath2 . in order
that the @xmath20 isometry group of ads@xmath5 is a symmetry group of the full solution , the four - form field strength necessarily has an `` electric '' component proportional to the volume form of ads@xmath5 , and a `` magnetic '' component which is a pull - back from @xmath2 .
we show , with the exception of the sasaki - einstein case , that the geometry on @xmath2 admits a canonical local @xmath3-structure , and determine the necessary and sufficient conditions for a supersymmetric solution in terms of this structure .
in particular , @xmath2 is equipped with a canonical killing vector field @xmath21 , which is the geometric counterpart to the @xmath22 r - symmetry of the dual @xmath23 superconformal field theory .
purely magnetic solutions correspond physically to wrapped m5-brane solutions , and we correspondingly recover the supersymmetry equations in @xcite from our analysis .
there is a single known solution in the literature , where @xmath2 is an @xmath24 bundle over a three - manifold @xmath25 equipped with an einstein metric of negative ricci curvature .
on the other hand , solutions with non - vanishing electric flux have a non - zero quantized m2-brane charge @xmath26 , and include the sasaki - einstein manifold solutions as a special case where the magnetic flux vanishes .
for the general class of solutions with non - vanishing m2-brane charge , we show that supersymmetry endows @xmath2 with a canonical contact structure , for which the r - symmetry vector field @xmath21 is the unique reeb vector field . a number of physical quantities
can then be expressed purely in terms of contact volumes , including the gravitational free energy referred to above , and the scaling dimension of bps operators @xmath27 dual to probe m5-branes wrapped on supersymmetric five - submanifolds @xmath28 .
these formulae may be evaluated using topological and localization methods , allowing one to compute the free energy and scaling dimensions of certain bps operators without knowing the detailed form of the supergravity solution . in our analysis
we recover the corrado - pilch - warner solution as a solution to our system of @xmath3-structure equations .
we also show that this solution is in a subclass of solutions which possess an additional killing vector field . for this subclass
the supersymmetry conditions are equivalent to specifying a ( local ) khler - einstein four - metric , together with a solution to a particular second order non - linear ode . we show that this ode admits a solution with the correct boundary conditions to give a gravity dual to the infrared fixed point of cubic deformations of @xmath14 m2-branes at a cy@xmath19 four - fold singularity .
in particular , when cy@xmath29 equipped with its flat metric , this leads to a new , smooth @xmath0 supersymmetric ads@xmath4 solution of m - theory .
the plan of the rest of this paper is as follows . in section [ sec : susy ] we analyse the general conditions for @xmath30 supersymmetry for a warped ads@xmath31 background of eleven - dimensional supergravity , reducing the equations to a local @xmath3-structure when @xmath2 is not sasaki - einstein . in section [ sec : contact ]
we further elaborate on the geometry and physics of solutions with non - vanishing electric flux , in particular showing that solutions admit a canonical contact structure , in terms of which various physical quantities such as the free energy may be expressed .
this section is an expansion of material first presented in @xcite .
finally , in section [ sec : taukilling ] we analyse the supersymmetry conditions under the additional geometric assumption that a certain vector bilinear is killing .
in addition to recovering the corrado - pilch - warner solution , we also numerically find a new class of cubic deformations of general cy@xmath19 backgrounds .
section [ concsec ] briefly concludes .
a number of technical details , as well as the analysis of various special cases , are relegated to four appendices . * note * : shortly after submitting this paper to the arxiv , the paper @xcite appeared , which contains a supersymmetric solution that appears to coincide with the solution we present in section [ sec : taukilling ] .
in this section we analyse the general conditions for @xmath0 supersymmetry for a warped ads@xmath32 background of eleven - dimensional supergravity .
the bosonic fields of eleven - dimensional supergravity consist of a metric @xmath33 and a three - form potential @xmath34 with four - form field strength @xmath35 . the signature of the metric is @xmath36 and
the action is [ action ] s & = & r * _ 11 - g_11 g - cg g , with @xmath37 the eleven - dimensional planck length .
the resulting equations of motion are [ eom ] r_mn-&=&0 , + * _ 11g+gg&=&0 , where @xmath38 denote spacetime indices .
we consider ads@xmath5 solutions of m - theory of the warped product form [ ansatz ] g_11 & = & ^2(g__4+g_7 ) , + g&= & m_4 + f . here
@xmath39 denotes the riemannian volume form on ads@xmath5 , and without loss of generality we take @xmath40.the factor here is chosen to coincide with standard conventions in the case that @xmath2 is a sasaki - einstein seven - manifold .
for example , the ads@xmath5 metric in global coordinates then reads @xmath41 , where @xmath42 denotes the unit round metric on @xmath43 . ] in order to preserve the @xmath44 invariance of ads@xmath5 we take the warp factor @xmath45 to be a function on the compact seven - manifold @xmath2 , and @xmath46 to be the pull - back of a four - form on @xmath2 .
the bianchi identity @xmath47 then requires that @xmath48 is constant .
the case in which @xmath49 will turn out to be quite distinct from that with @xmath50 . in an orthonormal frame ,
the clifford algebra @xmath51 is generated by gamma matrices @xmath52 satisfying @xmath53 , where the frame indices @xmath54 , and @xmath55 , and we choose a representation with @xmath56 . the killing spinor equation is [ killingspinoreqn ] _
m+ ( _ m^ npqr-8_m^n^pqr)g_npqr & = & 0 , where @xmath57 is a majorana spinor
. we may decompose @xmath58 via _ & = & _ 1 ,
_ a+3 = _ 5_a , where @xmath59 and @xmath60 are orthonormal frame indices for ads@xmath5 and @xmath2 respectively , @xmath61 , @xmath62 , and we have defined @xmath63 . notice that our eleven - dimensional conventions imply that @xmath64 .
the spinor ansatz preserving @xmath65 supersymmetry in ads@xmath5 is [ spinoransatz ] & = & ^+^/2 + ( ^+)^c^/2^c , where @xmath66 is a positive chirality killing spinor on ads@xmath5 , so @xmath67 , satisfying [ ads4spinor ] _ ^+ & = & _ ( ^+)^c .
the superscript @xmath68 in ( [ spinoransatz ] ) denotes charge conjugation in the relevant dimension , and the factor of @xmath69 is included for later convenience . substituting ( [ spinoransatz ] ) into the killing spinor equation ( [ killingspinoreqn ] ) leads to the following algebraic and differential equations for the spinor field @xmath70 on @xmath2 [ spinoreqns ] ^n_n-^-3+^-3 f_npqr^npqr+^c&=&0 , + _ m+^-3_m-^-3 f_mpqr^pqr -_m^c&=&0 . for a supergravity solution one must also solve the equations of motion ( [ eom ] ) resulting from ( [ action ] ) , as well as the bianchi identity @xmath47 . motivated by the discussion in the introduction , in this paper we will focus on @xmath0 supersymmetric ads@xmath5 solutions for which there are two independent solutions @xmath71 , @xmath72 to ( [ spinoreqns ] ) .
the general @xmath23 killing spinor ansatz may be written as [ n2spinoransatz ] & = & _ i=1,2 _ i^+^/2_i + ( ^+_i)^c^/2^c_i .
in general the two killing spinors @xmath73 on ads@xmath5 satisfy an equation of the form _ _ i^+ & = & _ j=1 ^ 2 w_ij_(_j^+)^c .
[ adspinors ] multiplying by @xmath74 on the left it is not difficult to show that @xmath75 is necessarily a constant matrix . using the integrability conditions of ( [ adspinors ] ) , _
j w_ij w_jk^ * & = & _ ik , one can verify that , without loss of generality , by a change of basis we may take @xmath76 to be the identity matrix . thus @xmath77 and @xmath78 may both be taken to satisfy ( [ ads4spinor ] ) . in this case with @xmath0 supersymmetry there is a @xmath22 r - symmetry which rotates the spinors as a doublet .
it is then convenient to introduce [ chipm ] _ & & ( _ 1_2 ) , which will turn out to have charges @xmath79 under the abelian r - symmetry . in terms of the new basis ( [ chipm ] ) , the spinor equations ( [ spinoreqns ] ) read [ chispinoreqns ] ^n_n_-^-3_+^-3 f_npqr^npqr_+_^c&=&0 , + _ m_+^-3_m_-^-3 f_mpqr^pqr_-_m_^c&=&0 .
the condition of @xmath0 supersymmetry means that the spinors @xmath71 , @xmath72 in ( [ n2spinoransatz ] ) are linearly independent .
notice that we are free to make @xmath80 transformations of the pair @xmath81 , since this leaves the spinor equations ( [ chispinoreqns ] ) invariant .
we shall make use of this freedom below .
the scalar bilinears are @xmath82 and @xmath83 , which may equivalently be rewritten in the @xmath84 basis ( [ chipm ] ) .
the differential equation in ( [ spinoreqns ] ) immediately gives @xmath85 , so that using @xmath86 we may without loss of generality set @xmath87 .
setting @xmath88 in ( [ useful3 ] ) , the algebraic equation in ( [ spinoreqns ] ) thus leads to 2&= & - ^-3|_i_j , where @xmath89 .
we immediately conclude that for @xmath90 we have [ gammazero ] & = & 0 . when @xmath50 this statement is not necessarily true .
the case with @xmath50 and @xmath91 $ ] not identically zero is discussed separately in appendix [ app : m0 ] , where we show that there are no regular solutions in this class
. we may therefore take ( [ gammazero ] ) to hold in all cases .
it is straightforward to analyse the remaining scalar bilinear equations .
in particular , @xmath92 $ ] is constant , and using the remaining @xmath80 freedom one can without loss of generality set @xmath92=0$].=1 $ ] one can show that @xmath93 , which in turn leads to only @xmath65 supersymmetry . ] in the @xmath84 basis ( [ chipm ] ) we may then summarize the results of this analysis as [ scalars ] |_+_+ & = & 1 = |_-_- , |_+_- = 0 , + a complex function on @xmath2 , while it is convenient to define @xmath94 to be the real function [ zeta ] & & ^-3 .
notice that in the @xmath50 limit we have @xmath95 , while for @xmath49 instead @xmath94 is nowhere zero .
we also define the one - form bilinears [ vectors ] k & & |_+^c_(1)_- , l |_-_(1)_+ , + @xmath96 . _ a priori _ notice that @xmath97 and @xmath98 are complex , while @xmath99 is real . the spinor equations ( [ chispinoreqns ] )
imply that [ k4eqn ] 2 k & = & = 0 , where we have used ( [ gammazero ] ) .
thus in fact @xmath97 is real , and it is then straightforward to show that @xmath97 is a killing one - form for the metric @xmath100 on @xmath2 , and hence that the dual vector field @xmath101 is a killing vector field .
more precisely , one computes [ lie ] _ ( m k_n ) & = & -2 g_7 mn = 0 . using the fierz identity ( [ fierz ] )
one computes the square norm [ norm ] ^2 & & g_7 ( , ) = |s|^2 + ^2 .
in particular when @xmath90 we see from ( [ zeta ] ) that @xmath21 is nowhere zero , and thus defines a one - dimensional foliation of @xmath2 . in the case
that @xmath50 this latter conclusion is no longer true in general , as we will show in section [ m0limit ] via a counterexample .
the algebraic equation in ( [ chispinoreqns ] ) leads immediately to @xmath102 , and using both equations in ( [ chispinoreqns ] ) one can show that ( ^3 |_+^c_(2)_- ) & = & - f .
it follows that _ f & = & ( f)+ f = 0 , provided the bianchi identity @xmath103 holds .
, as we will show shortly in section [ sec : eom ] .
] thus @xmath21 preserves all of the bosonic fields .
one can also show that _ _ & = & 2 _ , [ charge2 ] so that @xmath84 have charges @xmath104 under @xmath21 . perhaps the easiest way to prove
this is to use the remaining non - trivial scalar bilinear equation [ sbilinear ] ^-3(^3s ) & = & 4l , to show that [ scharge ] _ s & = & 4s .
since @xmath21 preserves all of the bosonic fields , we may take the lie derivative of the spinor equations ( [ chispinoreqns ] ) to conclude that @xmath105 satisfy the _
same _ equations , and hence @xmath106 are linear combinations of @xmath84 .
the lie derivatives of the scalar bilinears , in particular ( [ scharge ] ) , then fix ( [ charge2 ] ) . is not identically zero .
however , when @xmath107 we necessarily reduce to the sasaki - einstein case , as shown in appendix [ app : se ] . in that case ( [ charge2 ]
) also holds .
] we thus identify @xmath21 as the canonical vector field dual to the r - symmetry of the @xmath0 scft .
given our ansatz , the equation of motion and bianchi identity for @xmath9 reduce to [ feqns ] ( ^3f ) & = & - m f , f = 0 , where @xmath108 denotes the hodge star operator on @xmath2 .
we begin by showing that supersymmetry implies the equation of motion , and that for @xmath90 it also implies the bianchi identity .
the imaginary part of the bilinear equation for the three - form @xmath109 leads immediately to [ mf ] m f & = & 6 ( ^6 ) .
thus for @xmath90 we deduce that @xmath46 is closed . on the other hand , the bilinear equation for the two - form @xmath110 [ * feqn ] ^3f & = & ( ^6|_+_(2)_+ ) - 6 ^ 6 ,
gives , via taking the exterior derivative , ( ^3f ) & = & -6(^6 ) = -mf , where in the second equality we have combined with equation ( [ mf ] ) .
we thus see that supersymmetry implies the equation of motion in ( [ feqns ] ) .
finally , using the integrability results of @xcite one can now show that the einstein equation is automatically implied as an integrability condition for the supersymmetry conditions , once the @xmath9-field equation and bianchi identity are imposed .
in particular , note that the eleven - dimensional one - form bilinear @xmath111 is dual to a timelike killing vector field , as discussed in @xcite and later in section [ sec : m5 ] .
we thus conclude _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ for the class of @xmath0 supersymmetric @xmath112 solutions of the form ( [ ansatz ] ) , supersymmetry and the bianchi identity @xmath103 imply the equations of motion for @xmath9 and the einstein equations
. moreover , when @xmath49 the bianchi identity @xmath103 is also implied by supersymmetry .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ note that similar results were obtained in both @xcite and @xcite .
in fact we will see in section [ m0limit ] that the @xmath50 supersymmetry conditions also imply the bianchi identity , although the arguments we have presented so far do not allow us to conclude this yet .
provided the three real one - forms @xmath113 defined in ( [ vectors ] ) are linearly independent , we may use them to in turn define a canonical orthonormal three - frame @xmath114 here , as opposed to @xmath98 , since @xmath115 is invariant under the r - symmetry generated by @xmath21 .
in particular , from the definitions in ( [ vectors ] ) , and using ( [ charge2 ] ) , ( [ scharge ] ) , we have that @xmath116 . ] .
more precisely , if these three one - forms are linearly independent at a point in @xmath2 , the stabilizer group @xmath117 of the pair of spinors @xmath84 at that point is @xmath118 , giving a natural identification of the tangent space with @xmath119 . here
the @xmath3 structure group acts on @xmath120 in the vector representation .
if this is true in an open set , it will turn out that we may go further and also introduce three canonical coordinates associated to the three - frame @xmath114 . in ( [ vectors ] )
is a linear combination of @xmath97 and @xmath115 , and indeed one finds that @xmath121 . ]
we study the case that @xmath113 are linearly _ dependent _ in appendix [ app : se ] .
in particular , for @xmath49 we conclude that at least one of @xmath107 or @xmath122 holds at such a point .
if this is the case over the whole of @xmath2 ( or , using analyticity and connectedness , if this is the case on any open subset of @xmath2 ) then we show that @xmath2 is necessary sasaki - einstein with @xmath123 . of course
, in general the three one - forms can become linearly dependent over certain submanifolds of @xmath2 , and here our orthonormal frame and coordinates will break down . structure . ] by analogy with the corresponding situation for ads@xmath124 solutions of type iib string theory studied in @xcite , one expects this locus to be the same as the subspace where a pointlike m2-brane is bps , and thus correspond to the abelian moduli space of the dual cft , although we will not pursue this comment further here .
returning to the generic case in which @xmath113 are linearly independent in some region , we may begin by introducing a coordinate @xmath125 along the orbits of the reeb vector field @xmath21 , so that & & 4 .
the equation ( [ scharge ] ) then implies that we may write [ scoords ] s & = & ^-3 ^(- ) .
this defines the real functions @xmath126 and @xmath127 , which will serve as two additional coordinates on @xmath2 .
the factor of @xmath128 has been included partly for convenience , and partly to agree with conventions defined in @xcite that we will recover from the @xmath50 limit in section [ m0limit ] . using ( [ sbilinear ] ) together with the fierz identity ( [ fierz ] )
, one can then check that e_1 & & k = ( + ) , + e_2 & & s^*l = , + e_3 & & ( k- s^*l ) = ( + ) , [ frame ] are orthonormal . here
@xmath129 is a local one - form that is basic for the foliation defined by the reeb vector field @xmath21 , _ i.e. _ @xmath130 , @xmath131 .
note here that [ reeblength ] ^2 & & g_y_7 ( , ) = ^2 + |s|^2 = ^2 + ^-6 ^ 2 = ( m^2 + 36 ^2 ) , is the square length of the reeb vector field .
the metric on @xmath2 may then be written as [ metric ] g_7 & = & g_su(2 ) + e_1 ^ 2 + e_2 ^ 2 + e_3 ^ 2 .we may now in turn introduce an orthonormal frame @xmath132 for @xmath133 , and define the @xmath3-invariant two - forms j & & j_3 e_1e_2 + e_3e_4 , + & & j_1+j_2 ( e_1+e_2)(e_3+e_4 ) . of course , such a choice is not unique we are free to make @xmath134 rotations , under which @xmath135 , @xmath136 , transform as a triplet , where the structure group is @xmath137 , and @xmath138 is the spin group associated to @xmath133 .
any spinor bilinear may be written in terms of @xmath139 , @xmath135 , having chosen a convenient basis ) , and the fact that @xmath140 , we see that also the @xmath135 are invariant under @xmath21 , namely @xmath141 . ] for the @xmath135 .
having solved for the one - forms in ( [ frame ] ) , the remaining differential conditions arising from @xmath142-form bilinears , for all @xmath143 , then be shown to reduce ( after some lengthy computations ) to the following system of three equations @xmath144 & = & 2 j_3 - 2\|\xi\|e_2\wedge e_3 ~ , \\
{ \mathrm{d}}(\|\xi\|^2 { \mathrm{e}}^{9\delta } j_2 \wedge e_2 ) - { \mathrm{e}}^{3\delta }
|s| { \mathrm{d}}(\|\xi\| { \mathrm{e}}^{6\delta } |s|^{-1 } j_1 \wedge e_3 ) & = & 0~ , \\ { \mathrm{d } } ( { \mathrm{e}}^{6\delta } j_1 \wedge e_2 ) + { \mathrm{e}}^{3\delta } |s| { \mathrm{d}}(\|\xi\| { \mathrm{e}}^{3\delta } |s|^{-1 } j_2 \wedge e_3 ) & = & 0~ , \end{array } } \ ] ] where in addition the flux is determined by the equation ( ^6 j_2 ) & = & - ^3 f - 6 ^ 6 .
notice this is the same equation ( [ * feqn ] ) we already used in proving that the equation of motion for @xmath9 follows from supersymmetry .
the bilinear on the right hand side is given in terms of our frame by & = & |s| j_2 e_2 - j_1 ( e_1 + |s|e_3 ) .
one can invert the expression for the flux using these equations to obtain @xmath145 notice that although we have written these equations in terms of the three real functions @xmath146 , @xmath147 and @xmath94 , in fact they obey ( [ reeblength ] ) , where @xmath94 is given by ( [ zeta ] ) . regarding @xmath126 as a coordinate , there is then really only one independent function in these equations , which may be taken to be the warp factor @xmath45 .
we also note that the connection one - form @xmath129 , defined via the orthonormal frame ( [ frame ] ) , has curvature determined by the first equation in ( [ susys ] ) , giving & = & .
it is important to stress that the set of equations ( [ susys ] ) , where the three - frame @xmath148 is given by ( [ frame ] ) , are both necessary and sufficient for a supersymmetric solution . in order to see this
, we recall that our @xmath3 structure can be thought of in terms of the two @xmath149 structures defined by the spinors @xmath150 , @xmath151 ( or equivalently @xmath71 , @xmath72 ) .
each of these determines a real vector @xmath152 , real two - form @xmath153 , and complex three - form @xmath154 , where recall that also @xmath155 .
in fact @xmath156 , so that the vectors determined by each @xmath149 structure are equal and opposite , and @xmath157 determine two @xmath149 structures on the transverse six - space @xmath158 .
let us now turn to the killing spinor equations in ( [ chispinoreqns ] ) .
we have two copies of these equations , one for each @xmath149 structure determined by the spinors @xmath84 .
we shall refer to the first equation in ( [ chispinoreqns ] ) as the algebraic killing spinor equation ( it contains no derivative acting on the spinor itself ) . using this notice that we may eliminate the @xmath159 term in the second equation , in order to get an equation linear in @xmath84
; we shall refer the resulting equation as the differential killing spinor equation . for each choice of @xmath160
, the latter may be phrased in terms of a generalized connection @xmath161 , where @xmath162 is the levi - civita connection .
the intrinsic torsion is then defined as @xmath163 for each @xmath149 structure , and may be decomposed into irreducible @xmath149-modules as a section of @xmath164 . since @xmath165 , the intrinsic torsion may be identified as a section of @xmath166 .
it is then a fact that the exterior derivatives of @xmath167 , @xmath168 , @xmath169 determine completely the intrinsic torsion @xmath170 the identifications of the irreducible modules are given explicitly in section 2.3 of @xcite .
our equations ( [ susys ] ) certainly imply the exterior derivatives of both @xmath149 structures , since they imply the exterior derivatives of all @xmath142-form bilinears , for @xmath143 .
it follows that from our supersymmetry equations we could ( in principle ) construct both @xmath170 , and hence write down connections @xmath171 which preserve each spinor , so @xmath172 . in other words ,
our conditions then imply the differential killing spinor equations for each of the @xmath0 supersymmetries . for the algebraic killing spinor equation , note first that @xmath173 forms a basis for the spinor space for each @xmath174 . thus in order for the algebraic equation to hold , it is sufficient that the bilinear equations resulting from the contraction of the algebraic killing spinor equation with @xmath175 and @xmath176 hold , where @xmath70 is either of @xmath84 . however , this is precisely how the identities in appendix [ app : useful ] were derived .
we thus find that the @xmath150 algebraic killing spinor equation in ( [ chispinoreqns ] ) is implied by the two zero - form equations -^-3 + 2 |_+_-^c & = & 0 , [ algebraiczero ] + _ + + ^-3|_+_(4)_+ f & = & 0 , and the one - form equations + ^-3_+f & = & 0 , + [ algebraicone ] ^-3 p - 2k + _ + ( ) - ^-3 ( |_+_(3)_+)f & = & 0 , with similar equations for @xmath151 .
notice that the first equation in ( [ algebraiczero ] ) is simply the scalar bilinear in ( [ scalars ] ) which determines @xmath177 .
the reader can find explicit expressions for the real two - form @xmath178 and three - form @xmath179 , in terms of the @xmath3-structure , in appendix [ app : structures ] .
using these expressions , one can show that ( [ susys ] ) imply the remaining scalar equation in ( [ algebraiczero ] ) and both of the equations in ( [ algebraicone ] ) , thus proving that our differential system ( [ susys ] ) also implies the algebraic killing spinor equations .
the computation is somewhat tedious , and is best done by splitting the equations ( [ susys ] ) into components under the @xmath180 decomposition implied by the three - frame ( [ frame ] ) .
this decomposition is performed explicitly in section [ sec : reduction ] . in the second equation in ( [ algebraiczero ] )
we note that each term is in fact separately zero .
we also note that the first equation in ( [ algebraicone ] ) may be rewritten as _ + ( ^6 j_+ ) & = & ( ^6(1-|s|^2 ) ) .
[ leeform ] the left hand side is essentially the _ lee form _ associated to the @xmath149-structure defined by @xmath150 . )
has the geometrical interpretation that the transverse six - dimensional space @xmath158 is conformally balanced . ] to conclude , we have shown that ( [ susys ] ) are necessary and sufficient to satisfy the original killing spinor equations ( [ chispinoreqns ] ) .
it is straightforward to take the @xmath50 limit of the frame ( [ frame ] ) , differential conditions ( [ susys ] ) , and flux @xmath46 given by ( [ flux ] ) . denoting @xmath181 , @xmath182 , @xmath183 and @xmath184 we obtain the metric ^-1g_7 & = & + ^2 + ^2(+^2 ^ 2 ) , with corresponding differential conditions [ m0diff ] ( ^-1)&= & 2 ^ -1/2_3 + 2 ,
+ ( ^-3/2_1-_2)&= & 0 , + ( _ 2+^-3/2 ^ -1_1)&= & 0 . the flux @xmath46 in ( [ flux ] ) then becomes [ fm0 ] f & = & ( ^-1/2_1 ) .
these expressions precisely coincide with those in section 7.2 of @xcite .
of course , this is an important cross - check of our general formulae .
notice that the bianchi identity for @xmath46 is satisfied automatically from the expression in ( [ fm0 ] ) .
in fact for the general @xmath50 class of geometries the bianchi identity and equation of motion for @xmath46 read f & = & 0 , ( ^3f ) = 0 . defining the conformally related metric @xmath185 ,
the equation of motion for @xmath46 becomes @xmath186 .
it follows that @xmath46 is a harmonic four - form on @xmath187 .
in particular , imposing also flux quantization we see that @xmath46 defines a non - trivial cohomology class in @xmath188 , which we may associate with the m5-brane charge of the solution .
when @xmath50 there is no `` electric '' component of the four - form flux @xmath9 , and these ads@xmath5 backgrounds have the physical interpretation of being created by wrapped m5-branes . indeed , as we shall see in section [ sec :
contact ] , when @xmath49 there is always a non - zero quantized m2-brane charge @xmath26 , with the supergravity description being valid in a large @xmath14 limit .
the supergravity free energy then scales universally as @xmath189 .
one would expect the free energy of the m5-brane solutions , sourced by the internal `` magnetic '' flux @xmath46 , to scale as @xmath190 , where the cohomology class in @xmath188 defined by @xmath46 scales as @xmath14 .
however , the lack of a contact structure in this case ( see below ) means that a proof would look rather different from the analysis in section [ sec : contact ] . in section 9.5 of @xcite
the authors found a solution within the @xmath50 class , solving the system ( [ m0diff ] ) , describing the near - horizon limit of m5-branes wrapping a special lagrangian three - cycle @xmath25 .
in fact this is the eleven - dimensional uplift of a seven - dimensional solution found originally in reference @xcite .
the internal seven - manifold @xmath2 takes the form of an @xmath24 fibration over @xmath25 , where the latter is endowed with an einstein metric of constant negative curvature . as one sees explicitly from the solution , the r - symmetry vector field @xmath191 acts on @xmath192 by rotating the @xmath193 factor in the latter decomposition .
in particular , there is a fixed copy of @xmath43 , implying that @xmath191 does _ not _ define a one - dimensional foliation in this @xmath50 case .
notice this also implies there can not be any compatible global contact structure , again in contrast with the @xmath49 geometries .
the flux @xmath46 generates the cohomology group @xmath194 .
as far as we are aware , the solution in section 9.5 of @xcite is the only known solution in this class .
it would certainly be very interesting to know if there are more ads@xmath5 geometries sourced only by m5-branes . in this section we further analyse the system of supersymmetry equations ( [ susys ] ) , extracting information from each component under the natural @xmath180 decomposition implied by the three - frame ( [ frame ] ) . since we have dealt with the @xmath50 equations in the previous section , we henceforth take @xmath90 in the remainder of the paper .
we begin by defining the one - form [ calb ] & & ( + ) , which appears in the frame element @xmath195 in ( [ frame ] ) , so that e_3 & = & , and further decompose & & _ + , where @xmath196 .
since also @xmath197 and @xmath198 are orthogonal to @xmath199 , it follows that @xmath200 is a linear combination of @xmath201 , @xmath202 , the orthonormal frame for the four - metric @xmath133 in ( [ metric ] ) .
it is also convenient to rescale the latter four - metric , together with its @xmath3 structure , via _ i & & j_i , i=1,2,3 , so that correspondingly @xmath203 . , where @xmath50 . ]
notice this makes sense only when @xmath90 , so that @xmath94 is nowhere zero . given the coordinates @xmath204 defined via ( [ frame ] ) , it is then natural to decompose the exterior derivative as & = & + + + , where from now on hatted expressions will ( essentially ) denote four - dimensional quantities
. we may then decompose the exterior derivatives and forms in the supersymmetry equations ( [ susys ] ) under this natural @xmath180 splitting . beginning with the first equation in ( [ susys ] ) ,
the utility of the definition ( [ calb ] ) is that this first supersymmetry equation becomes simply [ susy1 ] & = & 2_3 - , where to simplify resulting equations it is useful to define the function & & .
decomposing as outlined above , this becomes [ susy1dec ] _ - _ & = & 0 , + _ & = & - , + & = & 2_3 .
note here that everything is invariant under @xmath191 .
the integrability condition for ( [ susy1 ] ) immediately implies that @xmath205 , while combining the component ( _ ) - _ ( ) & = & 0 , with the first and last equation in ( [ susy1dec ] ) leads to the conclusion _ & = & 0 = . given ( [ norm ] ) ,
this then implies _ & = & 0 = , so that the warp factor @xmath45 , and the related functions @xmath206 , @xmath94 , @xmath146 and @xmath147 , all depend only on the coordinate @xmath126 ! as an aside comment , we notice that a subset of the equations in ( [ summary ] ) may be re - interpreted as equations for a dynamical _ contact - hypo _ structure on a five - dimensional space @xcite .
here we decompose the seven - dimensional manifold under a @xmath211 split , where the two transverse directions are parametrized by the coordinates @xmath126 and @xmath125 .
the @xmath212 then define a contact - hypo structure ( at fixed @xmath126 ) obeying the equations = 2j_3 , = ( _ u ) , where @xmath213 .
note that when @xmath214_-=0 $ ] these become the conditions characterizing a sasaki - einstein five - manifold .
however , in this paper we will not pursue further this point of view .
we emphasize again that since @xmath45 is a function only of @xmath126 , this implies that the derived functions @xmath206 , @xmath94 , and @xmath147 also depend only on @xmath126 .
we conclude by writing an even more explicit expression for the flux given in ( [ flux ] ) : @xmath215_-\\ & & + \frac{m}{6}{\mathrm{e}}^{3\delta}\frac{(1-\|\xi\|^2)}{\|\xi\|}e_{12}\wedge \big[\partial_\rho \hat{j}_1\big]_-~. \end{array}}\ ] ] this expression is particularly useful for proving sufficiency of the differential system in section [ sec : ns ] .
we shall investigate the general equations ( [ summary ] ) , in a special case , in section [ sec : taukilling ] , reducing them to a single second order ode in @xmath126 .
in this section we further elaborate on the geometry and physics of solutions with @xmath49 .
in particular we show that all such solutions admit a canonical contact structure , for which the r - symmetry killing vector @xmath21 is the reeb vector field .
many physical properties of the solutions , such as the free energy and scaling dimensions of bps wrapped m5-branes , can be expressed purely in terms of this contact structure .
this section is essentially an expansion of the material in @xcite , as advertized in that reference . when @xmath90 we may define a one - form @xmath216 via p & & , where @xmath99 is the one - form bilinear defined in the second line in ( [ vectors ] ) . in terms of our frame ( [ frame ] )
, we then have & = & e_1 + e_3 , + & = & . up to a factor of @xmath217 , the one - form inside the square bracket on the left hand side of the first equation in ( [ susys ] ) is in fact @xmath216 .
thus we read off & = & ^3(j_3 - e_2e_3 ) , and a simple algebraic computation then leads to [ contactvolume ] ( ) ^3 & = & ^9_7 .here _ 7 & & -e_1e_2e_3_4 = -e_1e_2e_3j_3j_3 , denotes the riemannian volume form of @xmath2 ( with a convenient choice of orientation ) .
it follows that when @xmath90 , the seven - form @xmath218 is a nowhere - zero top degree form on @xmath2 , and thus by definition @xmath216 is a contact form on @xmath2 .
again , straightforward algebraic computations using the fierz identity in appendix [ app : useful ] lead to = 1 , = 0 .
this implies that the killing vector field @xmath21 is also the unique reeb vector field for the contact structure defined by @xmath216 .
when @xmath90 , equation ( [ mf ] ) immediately leads to the natural gauge choice f & = & a , where @xmath219 is the global three - form [ a ] a & & ^6 |_+^c _
( 3)_- . in terms of our frame , this reads a&= & ^6 .
notice that , either using the last expression or using ( [ charge2 ] ) , we find that _ a & = & 0 .
of course , one is free to add to @xmath219 any closed three - form @xmath220 , which will result in the same curvature @xmath46 [ gaugea ] a & & a + a . if @xmath220 is exact this is a gauge transformation of @xmath219 and leads to a physically equivalent m - theory background .
in fact more generally if @xmath220 has integer periods then the transformation ( [ gaugea ] ) is a large gauge transformation of @xmath219 , again leading to an equivalent solution .
it follows that only the cohomology class of @xmath220 in the torus @xmath221 is a physically meaningful parameter , and this corresponds to a marginal parameter in the dual cft .
in fact the free energy will be independent of this choice of @xmath220 , which is why we have set @xmath222 in ( [ a ] ) .
there is also the possibility of adding discrete torsion to @xmath219 when @xmath223 is non - trivial , but we will not discuss this here .
the flux quantization condition in eleven dimensions is [ flux11 ] n & = & -_y_7 * _ 11 g+ cg , where @xmath14 is the total m2-brane charge .
dirac quantization requires that @xmath14 is an integer .
substituting our ansatz ( [ ansatz ] ) into ( [ flux11 ] ) leads to n & = & _ y_7 m ^3_7 - af , where @xmath224 denotes the riemannian volume form for @xmath2 . by far
the simplest way to evaluate @xmath225 is to use the identity ( [ useful1 ] ) with @xmath88 . using ( [ a ] )
, this immediately leads to an expression for @xmath225 in terms of @xmath224 , and using ( [ contactvolume ] ) we obtain [ nexplicit ]
n & = & _ y_7()^3 . in particular , we see that @xmath90 leads to a non - zero m2-brane charge @xmath14 .
the effective four - dimensional newton constant @xmath226 is computed by dimensional reduction of eleven - dimensional supergravity on @xmath2 .
more precisely , by definition @xmath227 is the coefficient of the four - dimensional einstein - hilbert term , in einstein frame .
a standard computation leads to the formula [ newton ] & = & .
on the other hand , @xmath226 also determines the gravitational free energy @xmath228 [ f ] _ & & - |z| = . here the left hand side of ( [ f ] ) is the free energy of the unit radius ads@xmath5 computed in euclidean quantum gravity , where @xmath229 is the gravitational partition function .
thus in the supergravity approximation , @xmath228 is simply the four - dimensional on - shell einstein - hilbert action , which has been regularized to give the finite result on the right hand side of ( [ f ] ) using the boundary counterterm subtraction method of @xcite . via the ads / cft correspondence , @xmath230 , where @xmath231 is the free energy of the dual cft on the conformal boundary @xmath232 of ads@xmath5 .
combining ( [ newton ] ) and ( [ f ] ) then leads to the supergravity formula [ merlin ] & = & . combining ( [ nexplicit ] ) , ( [ merlin ] ) and ( [ contactvolume ] )
leads to our final formula @xmath233 we see that the famous @xmath189 scaling behaviour of the free energy of @xmath14 m2-branes continues to hold in the most general @xmath0 supersymmetric case with flux turned on .
moreover , the coefficient is expressed purely in terms of the contact volume of @xmath2 . in the sasaki - einstein case
this agrees with the riemannian volume computed using @xmath224 , but more generally the two volumes are different .
the contact volume has the property , in the sense described precisely in appendix b of @xcite , that it depends only on the reeb vector field @xmath21 determined by the contact structure .
in particular , if we formally consider varying the contact structure of a given solution , the contact volume is a strictly convex function of the reeb vector field @xmath21 .
it is of course natural to conjecture that this function is related as in ( [ f ] ) to minus the logarithm of the field theoretic @xmath234-function defined in @xcite , as a function of a trial r - symmetry in the dual supersymmetric field theory on @xmath232 .
this was conjectured in the sasaki - einstein case in @xcite , and has by now been verified in a large number of examples , including infinite families @xcite .
the contact volume has the desirable property that it can be computed using topological and fixed point theorem methods , so that one can compute the free energy of a solution essentially knowing only its reeb vector field .
we will illustrate this with the class of solutions in section [ sec : taukilling ] . finally ,
the scaling symmetry of eleven - dimensional supergravity in which the metric @xmath33 and four - form @xmath9 have weights two and three , respectively , leads to a symmetry in which one shifts @xmath235 and simultaneously scales @xmath236 , @xmath237 , where @xmath68 is any real constant .
we may then take the metric on @xmath2 to be of order @xmath238 , and conclude from the quantization condition ( [ flux11 ] ) , which has weight 6 on the right hand side , that @xmath239 .
it follows that the ads@xmath5 radius , while dependent on @xmath2 , is @xmath240 , and that the supergravity approximation we have been using is valid only in the @xmath241 limit .
a probe m5-brane whose world - space is wrapped on a generalized calibrated five - submanifold @xmath28 and which moves along a geodesic in ads@xmath5 is expected to correspond to a bps operator @xmath27 in the dual three - dimensional scft . in particular , when @xmath2 is a sasaki - einstein manifold , the scaling dimension of this operator can be calculated from the volume of the five - submanifold @xmath242 @xcite . in this section
we show that a simple generalization of this correspondence holds for the general @xmath23 supersymmetric ads@xmath32 solutions treated in this paper .
given a killing spinor @xmath57 of eleven - dimensional supergravity , it is simple to derive the following bps bound for the m5-brane @xcite ^ l _ _ 5 & & .
[ mainbound ] here @xmath243 is the three - form on the m5-brane , defined by @xmath244 where @xmath245 is closed and @xmath246 denotes the pull - back to the m5-brane world - volume .
the one - form @xmath247 , two - form @xmath248 and five - form @xmath249 are defined @xcite by the eleven - dimensional bilinears k & & |_(1 ) , on the world - space of the m5-brane .
we have defined @xmath250 as usual .
the bound ( [ mainbound ] ) follows from the inequality _
- ^2 & = & ^_- 0 , where @xmath251 is the @xmath206-symmetry projector and @xmath252 is the traceless hermitian product structure & & _ 0 . here @xmath253 , where the two - form @xmath254 is the world - space dual of @xmath243 .
this bound is saturated if and only if @xmath255 and corresponds to a probe m5-brane preserving supersymmetry .
we write the ads@xmath5 metric in global coordinates ( _ cf .
_ footnote [ footing ] ) and choose the static gauge embedding @xmath256 , where @xmath257 is global time in ads@xmath5 and @xmath258 , with @xmath259 , are coordinates on @xmath2 .
the dirac - born - infeld lagrangian @xmath260 is then defined by @xmath261 .
the vector @xmath262 dual to the one - form @xmath142 is a time - like killing vector , which using the explicit form of the eleven - dimensional @xmath23 killing spinor ( [ n2spinoransatz ] ) , and an appropriate choice of ads@xmath5 spinors @xmath263 , reads [ dualk ] k _ & = & _ t + .
accordingly , @xmath264 , and hence the bound ( [ mainbound ] ) is saturated when @xmath265 ( _ i.e. _ the m5-brane is at the centre of ads@xmath5 ) and l _ _ 5 & = & .[satur ] the energy density of an m5-brane can be computed by solving the hamiltonian constraints @xcite . for the static gauge embedding and @xmath266 these lead to & = & p_t = t _
( l _ + _ t ) , where @xmath267 is the m5-brane tension and the contribution from the wess - zumino coupling is @xmath268 , with the potential @xmath269 defined through @xmath270 .
however , from the explicit expression of @xmath34 one can check that we have @xmath271 .
the m5-brane energy is then given by e _ & = & t _ _ _ 5 l _ _ 5 = t _ _ _ 5 ( h)h + j^*h + j^ * , [ trivial ] where we used .
let us briefly discuss this expression for the energy . with our gauge choice ( [ a ] ) for the three - form potential , in general we have @xmath272 , where @xmath245 is a closed three - form . if @xmath245 is exact and invariant and @xmath21 generate symmetries of the m5-brane action .
] under @xmath262 , namely @xmath273 with @xmath274 , then one can check that the integral does not depend on @xmath245 . to see this
, one has to recall that @xmath275 , use the results of @xcite , and apply stokes theorem repeatedly .
if @xmath245 is not exact , _ a priori _ it will contribute to the energy , and hence we expect the dimension of the dual operator to be affected .
we leave an investigation of this interesting possibility for future work , and henceforth set @xmath276 .
in particular , @xmath219 is expressed as a bilinear of @xmath84 in . using the explicit form of the eleven - dimensional @xmath23 killing spinor ( [ n2spinoransatz ] ) and
the static gauge embedding one derives ^*k & = & ^2 k , + ^*&= & 4 ^3 \ { - [ |_+^c _ ( 2 ) _ - ] + [ |^+_1(^+_2)^c ] [ |_+^c _ ( 2 ) _ - ] } , + ^*&= & 4 ^6 \ { [ |_+^c _ ( 2 ) _ - ] + [ |^+_1(^+_2)^c ] [ |_+^c _ ( 2 ) _ - ] } , where @xmath277 denotes a pull - back to @xmath2 , and where the constant scalar bilinear @xmath278 $ ] is rescaled for convenience to @xmath279 .
the @xmath84 bilinears can then be expressed in terms of @xmath280 and @xmath135 .
the non - constant scalar @xmath281 $ ] drops out of the calculation and one arrives at ( j^*kh)h + j^ * h + j^ * = - ( ) ^2 . hence we get the remarkably simple result e _ & = & - t _ _ _ 5 ( ) ^2 .
combining the latter with ( [ nexplicit ] ) , and identifying @xmath282 with the energy @xmath283 in global ads , leads straightforwardly to the formula @xmath284 the scaling dimensions of operators dual to bps wrapped m5-branes are thus also determined purely by the contact structure . as for the contact volume of @xmath2 , the right hand side of ( [ delta ] ) can again be computed from a knowledge of @xmath242 and the reeb vector field @xmath21 .
since the general system of supersymmetry equations presented in section [ sec : reduction ] is rather complicated , in this section we impose a single simplifying assumption , namely that @xmath285 is a _ killing vector _ field for the metric generates a symmetry of the full solution . indeed we will show that in general the flux @xmath46 is _ not _ invariant under @xmath286 . ]
there are two motivations for this .
firstly , it is clearly a natural geometric condition .
secondly , the only solution in the literature in the @xmath90 class that is not sasaki - einstein is the corrado - pilch - warner solution @xcite .
this solution describes the infrared fixed point of a massive deformation of the maximally supersymmetric ads@xmath4 solution , and has the same topology but with non - standard metric on @xmath17 and flux .
we will first show that the assumption that @xmath285 is killing immediately leads to the four - metric @xmath133 being conformal to a khler - einstein metric , and that the supersymmetry conditions then entirely reduce to a single second order non - linear ode .
the corrado - pilch - warner solution is a particular solution to this ode , with @xmath133 being ( conformal to ) the standard fubini - study metric on @xmath288 .
we will then show numerically that there exists a second solution , dual to the infrared fixed point of a cubic deformation of @xmath14 m2-branes at a general @xmath289 singularity , where @xmath290 denotes any calabi - yau three - fold cone .
in particular , when @xmath291 endowed with a flat metric , this leads to a new , smooth @xmath0 supersymmetric ads@xmath4 solution .
let us analyze the conditions ( [ summary ] ) , with the assumption that @xmath285 is killing .
notice that the latter implies _ & = & _ _
= _ _ i + 0
. the left hand side of last equation in ( [ summary ] ) is thus identically zero .
taking the real and imaginary parts of the right hand side then implies that @xmath292 is self - dual .
the plus subscripts may then be dropped in the second line of ( [ summary ] ) , and we see that _ _ i & = & -_i , holds for all @xmath136 . recalling that @xmath206 is always a function only of @xmath126 , we may introduce the rescaled @xmath3 structure _ i & & f ( ) _ i , i=1,2,3 , and see that provided @xmath293 satisfies the differential equation [ diff1rho ] & = & -f , then the @xmath3-structure two - forms @xmath294 are independent of @xmath126 .
similarly , the killing condition on @xmath285 implies that @xmath295 and @xmath200 are independent of @xmath127 , and the first equation in ( [ summary ] ) then implies that @xmath296 depends only on @xmath126 .
we may then similarly solve the second equation in ( [ summary ] ) by rescaling & & f ( ) , and deduce that @xmath297 is independent of both @xmath127 and @xmath126 .
similarly writing & & _
+ , where now @xmath298 is a constant , the remaining equations in ( [ summary ] ) are [ remaining ] & = & 2_3 , ( _ 1+_2 ) = -f(_- ^2)(_1+_2 ) , + _ ( _ 1+_2 ) & = & - u ( _ 1+_2 ) .
since the blackboard script quantities are independent of @xmath126 , the second equation in ( [ remaining ] ) implies that [ diff2rho ] f(_- ^2 ) & = & - , which is _ a priori _ a function of @xmath126 , is in fact a constant . at this point
we should recall the definitions & = & ^-3 , = . in order to remove the explicit factors of @xmath48 , and write everything in terms of a single function ,
it is convenient to rescale r & & , ^2(r ) ( ) ^2 . in terms of these new variables , the differential equations ( [ diff1rho ] ) , ( [ diff2rho ] ) read @xmath299 \frac{\left(r\alpha ' - r^2\alpha^3\right)f}{\sqrt{1+(1+r^2)\alpha^2 } } & = & -\gamma~ , \end{array}}\ ] ] which are a coupled set of first order odes for the functions @xmath300 , @xmath301 , and from henceforth a prime will denote derivative with respect to the coordinate @xmath302 .
the remaining supersymmetry conditions ( [ remaining ] ) now simplify to & = & 2_3 , ( _ 1+_2 ) = ( _ 1+_2 ) , + _ ( _ 1+_2 ) & = & _ ( _ 1+_2 ) . here
both @xmath303 and @xmath298 are constants .
the first line says that the four - metric defined by @xmath304 is khler - einstein with ricci tensor satisfying @xmath305 .
the second equation is solved simply by multiplying @xmath306 by a phase @xmath307 , so that everything is independent of @xmath127 .
to conclude , given any khler - einstein four - metric @xmath308 with ricci curvature @xmath309 , a solution to the ode system ( [ odes ] ) leads to a ( local ) supersymmetric ads@xmath5 solution with internal seven - metric being [ 7dmetric ] g_7 & = & g _
+ , and flux f & = & ( m ^-3(1+r^2 ) - 9 r^2 f ) ( - ) _ 1 + & + & ( + a _ ) ( j_1 + ( - ) j_2 ) , [ newf ] where we have written the latter in terms of three functions @xmath310 in order to simplify the expression slightly . however , recall that the warp factor is related to @xmath311 via ^6 & = & ( ) ^2 ( 1 + r^2 + ^-2 ) . here
we have denoted @xmath312 , and without loss of generality we have set @xmath313 by rescaling the @xmath127 coordinate . from ( [ newf ] ) we see explicitly that @xmath314 , since the holomorphic two - form on the khler - einstein base satisfies @xmath315 . therefore ,
as anticipated at the beginning of this section , @xmath286 does not generate a symmetry of the full solution . if @xmath316 then by rescaling @xmath317 we may also without loss of generality set @xmath318 .
the local one - form @xmath319 is globally a connection on the anti - canonical bundle of the khler - einstein four - space .
notice that we may algebraically eliminate @xmath301 from the first equation in ( [ odes ] ) to obtain the single second order ode for @xmath300 [ 2ndorder ] 3rf^2+f(rf-f)&= & .
we begin by noting that the following is an explicit solution to the ode system ( [ odes ] ) [ cpw ] f(r)&= & ( 2- ) , ( r ) = .
taking the khler - einstein metric to be simply the standard fubini - study metric on @xmath288 , and with @xmath320 $ ] , we claim this is precisely the ads@xmath4 solution described in @xcite .
in fact the authors of @xcite conjectured that one should be able to replace @xmath288 by any other khler - einstein metric ( with positive ricci curvature ) to obtain another supergravity solution .
this was shown in @xcite for the special case in which one uses the khler - einstein metrics associated to the @xmath321 sasaki - einstein manifolds @xcite , @xcite .
we can immediately read off the warp factor [ pwzeta ] ^-3 = & = & = . comparing our @xmath322 component of the metric ( [ 7dmetric ] ) to the @xmath323 component of the metric in @xcite , we are led to the identification r & = & 2 ^ 2 .
it is then straightforward to see that our metric ( [ 7dmetric ] ) coincides with the metric in @xcite , and using ( [ newf ] ) also that the fluxes agree .
the corrado - pilch - warner solution fits into a more general class of solutions obtained by deforming the theory on @xmath14 m2-branes at the conical singularity of the calabi - yau four - fold @xmath324 . in this section
we give a unified treatment , in particular recovering the field theory result in @xcite for the free energy of such theories using our contact volume formula ( [ freeenergy ] ) .
we begin by taking @xmath308 to be the ( local ) khler - einstein metric associated to a sasaki - einstein five - manifold .
the corresponding sasaki - einstein five - metric is [ se5 ] g__5 & = & ( + a_)^2 + g _ , which leads to a calabi - yau four - fold product metric on @xmath289 given by g__4 & = & _ 1 ^ 2 + _ 1 ^ 2 + _ 0 ^ 2 + _ 0 ^ 2_0 ^ 2 . here
@xmath325 are radial variables , and @xmath326 has period @xmath327 .
the corresponding sasaki - einstein seven - metric at unit distance from the conical singularity at @xmath328 is [ se7 ] g__7 & = & r^2 + r^2 + ( 1-r^2)_0 ^ 2 , where @xmath329 .
note that the killing vector fields @xmath330 and @xmath331 vanish at @xmath332 and @xmath333 , respectively , and that the reeb vector field is the sum @xmath334 .
the metric ( [ se7 ] ) is singular at @xmath332 ( which is an @xmath335 locus parametrized by @xmath326 ) unless the original sasaki - einstein five - manifold is @xmath336 equipped with its standard round metric .
this is simply because the calabi - yau four - fold is also singular along @xmath332 , which is the conical singularity of @xmath290 .
it is no coincidence that the sasaki - einstein metric ( [ se7 ] ) resembles our general metric ( [ 7dmetric ] ) .
the ads@xmath16 background is the infrared limit of @xmath14 m2-branes at the conical singularity @xmath328 of @xmath289 .
the holomorphic function @xmath337 leads to a scalar kaluza - klein mode on the sasaki - einstein seven - space , which in turn is dual to a gauge - invariant scalar chiral primary operator @xmath338 in the dual three - dimensional scft .
we may then consider deforming the scft by adding the operator @xmath339 . in three dimensions , this is a relevant deformation for @xmath340 and @xmath341 , as discussed in @xcite .
moreover , such a term can appear in the superpotential of a putative infrared fixed point also only if @xmath340 , @xmath341 , since otherwise one violates the unitarity bound
the r - charge / scaling dimension of @xmath338 would be @xmath342 , and necessarily we have @xmath343 for a unitary cft in three dimensions , with equality only for a free field .
the gravity dual to the infrared fixed point of the massive @xmath340 deformation is the corrado - pilch - warner solution of the previous section , while we will find the @xmath341 solution as a numerical solution to the odes ( [ odes ] ) in the next section .
in @xcite the authors studied @xmath344 , @xmath0 supersymmetric field theories for @xmath14 m2-branes on @xmath289 backgrounds , in particular computing the free energy using localization and matrix model techniques .
this allows one to compute the ratio of uv and ir free energies , where the uv theory is dual to the ads@xmath16 background , while the ir theory is the fixed point of the renormalization group flow induced by the @xmath345 deformation .
they found the universal formula , independent of the choice of cy@xmath10 , [ fieldp ] & = & .
we now show that this field theory result is easily obtained using our contact volume formula ( [ freeenergy ] ) , thus acting as a check of the ads / cft duality for this class of theories .
the cy@xmath19 calabi - yau four - fold has at least a @xmath346 symmetry , in which the first @xmath347 acts on the cy@xmath10 , and under which the cy@xmath10 killing spinors have charge @xmath348 , and the second @xmath347 acts in the obvious way on the copy of @xmath349 with coordinate @xmath350 .
let us denote the components of the reeb vector field in this basis as @xmath351 . in terms of the explicit coordinates introduced above
, this gives the reeb vector field as [ reebbasis ] & = & _ 1 _ + _ 0__0 .
for the calabi - yau four - fold metric , we have already noted that @xmath352 and @xmath353 . in general , the killing spinors have charge @xmath354 , as in equation ( [ charge2 ] ) , precisely when [ weight4 ] _ 1+_0 & = & 4 , which is also equivalent to the holomorphic @xmath355-form @xmath356 having charge 4 . as shown in appendix b of @xcite , in general
the contact volume is a function of the reeb vector field . in our case
the contact volume of @xmath2 is given by the general formula [ volumes ] ( y_7)[_1,_0 ] & = & ( y_5)[_1 ] , where @xmath357 denotes the five - manifold link of cy@xmath10 . using @xmath352 for a sasaki - einstein metric , ( [ volumes ] ) implies the relation @xmath358 between sasaki - einstein volumes .
notice that @xmath353 gives the expected scaling dimension @xmath359 of a free chiral field . in going from the geometric scaling dimension under the euler vector to the scaling dimension @xmath45 in field theory , _ cf .
_ equation ( 2.31 ) of @xcite . ]
let us now consider the ir solution corresponding to the deformation by @xmath345 .
the scaling dimension of @xmath338 necessarily changes from @xmath359 to @xmath342 .
since the coordinate @xmath350 gives rise to the kaluza - klein mode leading to this bps operator , this means the charge of @xmath350 under the reeb vector field at the ir fixed point should be @xmath360 . from ( [ weight4 ] )
we thus have @xmath361 .
we then compute the contact volumes ( y_7^(p ) ) & = & ( y_5)[_1 ] = ( ) ^-3(y_5)[3 ] + & = & ( _ 7 ) . here
we have used that the volume of a contact five - manifold is homogeneous degree @xmath362 in the reeb vector field @xcite , @xcite .
taking the square root and using our free energy formula ( [ freeenergy ] ) , we precisely reproduce the field theory result ( [ fieldp ] ) !
this is a somewhat formal agreement , since the ir fixed point is not expected to exist due to the unitarity bound , as explained above . ]
we conclude by recording that the reeb vector field ( [ reebbasis ] ) at the ir fixed point is [ reebp ] & = & _ + _ _ 0 . this will be crucial in the following sections when we consider the appropriate boundary conditions for the odes ( [ odes ] ) . before moving on to the gravity dual of the cubic @xmath341 deformation ,
let us consider again the explicit @xmath340 corrado - pilch - warner solution .
the analysis in the previous section implies that the reeb vector field should be [ cpwreeb ] & = & 4 _
= _ + 2__0 , where @xmath125 is the coordinate in ( [ 7dmetric ] ) .
this fact is very closely related to the appropriate boundary conditions one needs to impose on the odes ( [ odes ] ) in order to obtain a good supergravity solution . for the explicit solution in section [ sec : cpw ] ,
the coordinate @xmath363 $ ] , and by definition @xmath331 is the killing vector field that vanishes at @xmath332 , while @xmath364 vanishes at @xmath365 .
let us see how this works precisely . without loss of generality we henceforth set & = & 3 .
near to @xmath332 , we may use @xmath366 , @xmath367 to compute a_+ b_^2_r=0 & = & ( a+2b)^2 .
this vanishes only if @xmath368 , so that the vanishing vector field at @xmath332 is _ _ 0 & & -2_+ _ . to determine the proportionality constant we need to examine the rate of collapse . introducing @xmath369 , we have near to @xmath332 that @xmath370 $ ] .
thus @xmath371 measures geodesic distance from @xmath372 , to leading order , and if @xmath331 is such that @xmath326 has period @xmath327 and @xmath331 vanishes at @xmath372 , then the metric will be smooth here only if @xmath373 .
said another way , to leading order near to @xmath372 the metric must be the standard metric @xmath374 on @xmath193 in polar coordinates @xmath375 .
we then compute -2_+ _ ^2 & = & ^2 r^2 + ( r^4 ) .
this fixes [ cpwphi0 ] _ _ 0 & = & 2_- _ .
we may perform a similar analysis near to @xmath365 .
introducing @xmath376 , we have @xmath377 , while near to @xmath378 we have @xmath379 . now
b_^2 & = & , so this vector field vanishes at @xmath365 only if @xmath380 , leading to _ & & _ .
in particular , the coefficient may be computed from @xmath381 $ ] and _ ^2 & = & z^2 = z^2 , where we have used @xmath318 in the last step .
this is indeed the expected result , since for the canonical scaling of @xmath318 the connection term @xmath382 in the metric ( [ 7dmetric ] ) must be the contact one - form @xmath383 for the original sasaki - einstein five - manifold ( [ se5 ] ) , implying that indeed @xmath384 .
the collapsing part of the metric near to @xmath365 is then @xmath385 .
this locally is precisely the cy@xmath10 conical metric , giving a smooth collapse at @xmath386 if and only if the the khler - einstein metric is the standard metric on @xmath288 . more generally , @xmath365 is an @xmath335 locus of cy@xmath10 cone singularities . to summarize , putting ( [ cpwphi0 ] ) together with @xmath384 we have shown 2_&= & _ _ 0 + _ .
recalling that the reeb vector field is @xmath387 , we have thus shown & = & 4 _ = _ + 2__0 .
this precisely coincides with ( [ cpwreeb ] ) , which was derived in the previous section based only on topological and scaling arguments .
we may now use precisely the same arguments as the previous section to deduce the appropriate boundary conditions for the odes ( [ odes ] ) in the case of cubic @xmath341 deformations .
the reeb vector field is now & = & 4 _ = _ + _ _ 0 , where by definition again @xmath388 and @xmath364 are the vanishing vector fields , while @xmath125 is the coordinate in our metric ( [ 7dmetric ] ) .
let us begin by considering the behaviour near to @xmath332 .
suppose that @xmath389 , with @xmath390 a non - zero constant .
then the first ode in ( [ odes ] ) implies ( f) & ~ & - r^1 + 2 , which leads to the leading order solution f(r)~a_0 , where @xmath391 is a constant . the second ode in ( [ odes ] )
is then to leading order & ~ & . for @xmath392
the right hand side tends to zero as @xmath393 , which is a contradiction .
this is also the case for @xmath394 . on the other hand ,
@xmath300 blows up exponentially at @xmath332 unless @xmath395 .
since we do not want the size of the khler - einstein metric to blow up on @xmath2 , a regular solution must hence have @xmath396 .
given this , to leading order the last equation becomes & ~ & -a_0w ( r^-2+w^2)^-1/2 -a_0 .
thus we conclude that @xmath397 .
note that @xmath398 , and that the metric ( [ 7dmetric ] ) is then positive definite only if @xmath399 .
as in the previous section , introducing @xmath400 we compute r^2 & ~ & = r^2 , we now determine the vanishing vector field at @xmath332 , computing [ bob ] a_+ b_^2_r=0 & = & ( a-)^2 , where we have eliminated @xmath401 .
thus the vector field @xmath402 vanishes at @xmath332 . to fix the normalization we need the rate of collapse : -_-_^2&= & r^2+o(r^2 ) , near to @xmath372 .
this fixes _ _ 0 & = & _ + _ . in fact
this is already enough to determine @xmath403 .
recall that @xmath387 is the reeb vector field , so we can also write _
_ 0 & = & + _ .
since the coordinate @xmath350 on @xmath349 has charge @xmath404 under @xmath21 , we thus conclude that in general 1 & = & , so that & = & -1 + .
in particular , the corrado - pilch - warner solution has @xmath405 , while for the cubic deformation we should set @xmath406 .
the boundary condition for @xmath301 near to @xmath332 is in general @xmath407 .
it is important to note that , with this boundary condition on @xmath301 , the metric is completely smooth near to @xmath332 . although @xmath301 is blowing up , the function @xmath408 , so that the khler - einstein factor in ( [ 7dmetric ] ) has finite non - zero size .
the remaining killing vector that is not zero also has finite length at @xmath332 , as one sees from ( [ bob ] ) .
we can now similarly analyse the other collapse .
this is necessarily at a zero of @xmath300 . to see this , note that the khler - einstein part of the metric ( [ 7dmetric ] ) collapses at either a zero of @xmath311 , or a zero of @xmath317 ( potentially both ) .
suppose this is at @xmath409 .
if @xmath410 to leading order , with @xmath411 , then solving the ode for @xmath317 leads to the leading order result f(r ) & ~ & a_1 .
thus @xmath412 is in fact non - zero .
the second ode in ( [ odes ] ) is then consistent near to @xmath409 only if the exponent @xmath413 , which means that @xmath414 is a simple zero .
however , from the metric ( [ 7dmetric ] ) we see that in fact then the entire metric collapses at @xmath409 , which does not give the correct topology .
so we can rule out @xmath301 having a zero at @xmath409 .
thus @xmath415 .
let us suppose that to leading order f(r ) & ~ & q ( r_0-r)^ , with @xmath416 . then from the first ode in ( [ odes ] ) we obtain ( r ) & ~ & . notice that for the corrado - pilch - warner solution we have @xmath417 , and this leading order solution for @xmath301 near to @xmath409 is in fact the _ exact _ solution .
for our cubic @xmath341 solution @xmath301 must instead interpolate between @xmath418 behaviour near to @xmath332 and @xmath419 behaviour near to @xmath409 .
the second ode again fixes the exponent @xmath420 for consistency near to @xmath409 , and we conclude that f(r ) & ~ & q ( r_0-r ) , + ( r ) & ~ & , near to @xmath409
. moreover , the second ode then fixes [ gammaq ] & = & .
finally , we turn to looking at the vanishing vector field .
writing @xmath421 , we find that @xmath422 .
then [ bobagain ] a_+ b_^2 & = & a^2 + ( w^2 ) , so that the vanishing vector field at the root @xmath409 is again proportional to @xmath285 .
we find more precisely that , quite remarkably , _ ^2 & = & ( ) ^2w^2+o(w^2 ) , where we have substituted for @xmath423 using ( [ gammaq ] ) .
this is exactly the same behaviour as for the corrado - pilch - warner solution near to this root .
since this collapsing vector field is by definition @xmath364 , we again conclude that _ & = & _ .
again , this had to be the case for global reasons associated to the form of the connection one - form appearing in the metric .
again one finds that @xmath409 is an @xmath335 family of cy@xmath10 cone singularities , with the analysis being identical to that for the corrado - pilch - warner solution in the previous section .
this completes our analysis of the regularity conditions . setting @xmath318 ,
we have shown that the reeb vector field is & = & -_+4(1+)__0 .
using the fact that @xmath424 , this precisely agrees with our topological analysis in section [ sec : def ] , and in particular the formula ( [ reebp ] ) .
we may summarize the results of the previous sections as follows .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the gravity dual to the infrared fixed point of a deformation of a cy@xmath19 background by the operator @xmath339 may be obtained by solving the coupled set of odes for @xmath301 , @xmath300 : f & = & -f , + & = & -3 .
the boundary conditions are that near to @xmath332 we have @xmath425 , with @xmath399 a constant .
using the second ode above this implies that @xmath426 . then near to @xmath409 , for some @xmath427
, we must impose that @xmath428 , where the odes imply that @xmath429 and @xmath430 . with these boundary conditions
we obtain a smooth supergravity solution , up to the expected @xmath335 locus of cy@xmath10 singularities along @xmath409 . when the cy@xmath10 is simply flat @xmath431 , in particular we obtain a completely smooth @xmath0 supergravity solution with the topology ads@xmath4 . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ the corrado - pilch - warner solution precisely solves this problem for @xmath340 , and physical arguments imply there should also be a solution for @xmath341 .
we have not been able to find this solution analytically , but it is straightforward to solve the odes numerically with the above boundary conditions .
we first change variable to @xmath432 , and then solve the second order ode ( [ 2ndorder ] ) in a taylor expansion in @xmath371 , around @xmath372 , up to some large order . using the constraint that @xmath433 we find [ taylor ] f(r ) = - cr^2 - r^4 + r^6 + r^8 + ( r^10 ) , where @xmath68 is an arbitrary integration constant .
this then implies ( r ) & = & c^1/2r^-2 + c^3/2 - c^-1/2r^2 + ( r^4 ) .
thus @xmath301 has the correct behaviour @xmath434 , where we identify the constant @xmath435 .
we then have a numerical shooting problem : for each choice of integration constant @xmath68 , we solve the second order ode ( [ 2ndorder ] ) ( or equivalently the coupled first order system ) , with initial taylor expansion ( [ taylor ] ) .
we simply require that @xmath415 for some @xmath427 . from the analysis in the previous section ,
the odes themselves imply that a zero of @xmath300 is automatically a simple zero .
we find that there exists a point @xmath427 with @xmath415 for the choice c & & 2.4998 .
the resulting plot of the function @xmath436 , with @xmath437 , is shown in figure [ fig1 ] .
smaller values of @xmath68 lead to @xmath436 remaining positive , while for @xmath438 we find the numerics becomes highly unstable . indeed ,
the numerics is slightly unstable near the zero of @xmath317 for @xmath439 . as a cross check that we really do have a zero
, we note that at a zero of @xmath436 we necessarily have f(r_0 ) & = & - . in figure [ fig2 ]
we numerically plot the function @xmath440 , which should tend to zero at @xmath441 . of course , it is quite tantalizing that the numerical value of @xmath68 is so close to @xmath442 , perhaps suggesting the possibility of an analytic solution , or at least an analytic explanation of @xmath443 .
we leave this question open .
the main result of this paper is the determination of the necessary and sufficient conditions on supersymmetric solutions of @xmath1 supergravity that are dual to @xmath23 three - dimensional superconformal field theories .
the eleven - dimensional metric is taken to be a warped product of ads@xmath5 with a seven - dimensional riemannian metric , and we have allowed for the most general four - form @xmath9 consistent with @xmath20 symmetry .
we showed that generically the supersymmetry conditions may be formulated in terms of a canonical local @xmath3-structure on the seven - dimensional manifold @xmath2 .
the well - known freund - rubin ads@xmath32 solutions where @xmath2 is sasaki - einstein arise as a special case , characterized by an @xmath149-structure . for solutions with non - zero m2-brane charge
, we showed that many geometrical and physical properties of @xmath2 are captured by a contact structure , elaborating on the results presented in @xcite .
we also recovered the class of general solutions with vanishing m2-brane charge , previously discussed in @xcite . by imposing a single additional requirement ,
that a certain vector bilinear is a killing vector , we reduced the conditions to solving a second order non - linear ode .
the seven - dimensional metric on @xmath2 is then fully specified by the choice of a ( local ) four - dimensional khler - einstein metric , and any solution to this ode .
we managed to find an analytic solution of the ode , and showed that this reproduces a class of solutions found originally in @xcite .
in addition , using a combination of analytic and numerical methods , we have discovered a further solution to our ode , yielding a class of new supersymmetric ads@xmath5 solutions with non - trivial four - form flux .
these can be interpreted as holographically dual to certain cubic superpotential deformations of @xmath23 chern - simons gauge theories . when the khler - einstein metric is chosen to be that on @xmath288 , the seven - dimensional metric is a smooth ( non - einstein ) metric on @xmath17 , different from that of @xcite .
we suspect that there are no further regular solutions in this class .
our work may be regarded as providing the foundation for studying more general aspects of @xmath23 three - dimensional superconformal field theories with m - theory duals .
for example , we expect that the geometric characterization of solutions we presented may be used to attack general problems , such as the gravity dual of @xmath444-maximization , similarly to the developments in @xcite .
it is also clear that using our results it will be possible to construct a consistent kaluza - klein truncation to four dimensions , extending that in @xcite .
the ads@xmath5 solutions dual to beta - deformations @xcite of @xmath23 field theories must solve the equations that we presented , and it would interesting to verify this explicitly . of course , it would also be very interesting to use our general equations as a method for finding new solutions ( perhaps numerically ) , outside the classes that have been discovered so far marginal deformations @xcite . ] .
these are all exciting directions for future work .
we would like to thank the isaac newton institute for hospitality .
m. g. is supported by the erc starting independent researcher grant 259133-observablestring , d. m. is supported by the epsrc advanced fellowship
ep / d07150x/3 , a. p. is supported by an a. g. leventis foundation grant and via the act `` scholarship programme of s. s. f. by the procedure of individual assessment , of 2011 - 12 '' by resources of the operational programme for education and lifelong learning , of the european social fund ( esf ) and of the nsrf , 2007 - 2013 .
d. m. also acknowledges partial support from the stfc grant st / j002798/1
. j. f. s. is supported by the royal society and oriel college .
in this appendix we collect a number of useful identities that have been used repeatedly to derive the results presented in the main text . from the algebraic equation in ( [ spinoreqns ] )
one can derive the following useful identities ( |_i^c _ j^c + |_i _ j ) - ^-3 |_i^c _ j + + _ m |_i^c [ , ^m]_- _ j + f_mnpq^-3 |_i^c [ , ^mnpq]_+ _ j & = & 0 , [ useful1 ] + ( |_i^c _ j^c - |_i _ j ) + _ m |_i^c [ , ^m]_+ _ j + f_mnpq^-3 |_i^c [ , ^mnpq]_- _ j & = & 0 , [ useful2 ] where @xmath445 cliff@xmath446 is an arbitrary element of the clifford algebra and @xmath447_\pm$ ] denotes the ( anti)-commutator . similarly we note ( |_i^c _ j - |_i _
j^c ) + ^-3 |_i _ j + - _ m |_i [ , ^m]_- _ j - f_mnpq^-3 |_i [ , ^mnpq]_- _ j & = & 0 , [ useful3 ] + ( |_i^c _ j + |_i _ j^c ) + _ m |_i [ , ^m]_+ _ j + f_mnpq^-3 |_i [ , ^mnpq]_+ _ j & = & 0 .
[ useful4 ] similar identities exist in the alternative basis ( [ chipm ] ) . from the fierz identity for the @xmath448 algebra [ fierz7 ] |_1 _ 2 |_3
_ 4 & = & , where @xmath449 , @xmath202 , are arbitrary spin(7 ) spinors , we derive the useful identity [ fierz ] |^c_1 ^m _ 2 |^c_2 _ m _ 4 & = & |^c_1 _ 4 |^c_2 _ 2 -
in the main text we have presented our results , summarized in the equations in section [ sec : ns ] , in terms of an @xmath3-structure .
this is defined by the three one - forms @xmath197 , @xmath198 , @xmath195 , and three @xmath3-invariant two - forms @xmath451 , @xmath452 , @xmath453 . in arguing that the conditions we write are sufficient
, it is also convenient to think of this in terms two @xmath149-structures , defined by the killing spinors @xmath84 . in this appendix we present explicit formulas for the spinor bilinears in terms of both @xmath3- and @xmath149-structures .
recall that the @xmath3-structure is specified by two spinors @xmath71 , @xmath72 , or equivalently the linear combinations @xmath454 defined in ( [ chipm ] ) .
here we choose to use @xmath84 as our basis .
we then have the following zero - form bilinears [ 0form ] |_+_+ & = & |_-_- = 1 , + s & & |_+^c_+ = ( |_-^c_-)^ * , + & & |_+^c_- = ^-3 , one - form bilinears [ 1form ] k & & |_+^c_(1)_- = e_1 , + l & & |_-_(1)_+ = ( e_1+e_2- e_3 ) , + p & & -|_+_(1)_+ = |_-_(1)_- = e_1 + e_3 , two - form bilinears v _ & & 12 , + v_+ & = & j_2 , + v_- & = & j_3 + 1 e_2 ( |s| e_1 - e_3 ) , + bilinears w _ & & , + & = & - j_3 e_2 , + & = & - j_3 ( e_1- e_3 ) + |s| e_123 , + & = & j_1 ( |s| e_1 - e_3 ) + j_2 e_2 , + & = & - j_1 e_2 - j_2 e_3 , + & = & |s| j_2 e_2 - j_1 ( e_1 + |s|e_3 ) , + & & + j_2 ( e_1 + |s| e_3 ) . notice that the two - forms and three - forms above are an incomplete list we have included only those bilinears that are referred to explicitly in the text .
recall that we defined the two non - canonical @xmath149-structures as real vectors @xmath152 , real two - forms @xmath153 , and complex three - forms @xmath154 .
then we have the one - form bilinears _ + & = & - _ - = - p , two - form bilinears _ & = & v_+v_- , and three - form bilinears _ + & = & w_++w_- , _ - = ( w_+-w_-)^ * .
in this appendix we study the case in which the three one - forms @xmath113 are linearly dependent .
when they are linearly independent we have an @xmath3 structure , and in an open set we can then introduce corresponding coordinates , as described in section [ sec : frame ] . since these one - forms are derived from spinor bilinears , linear dependence implies we have an @xmath149 structure .
focusing on the @xmath90 case for clarity , we will prove that the only solutions for which we have a global @xmath149 structure are sasaki - einstein . in order to proceed ,
we impose the linear relation [ depend ] a k + b s^*l + c s^ * l & = & 0 , with @xmath220 , @xmath455 , @xmath68 not all zero . making use of the fierz identity in ( [ fierz ] )
it is straightforward to compute the dot products of each of @xmath113 into this equation .
an analysis of the resulting three equations then implies that at least one of @xmath456 or @xmath457 must hold . in particular ,
if @xmath456 then necessarily @xmath458 , while if @xmath457 then @xmath459 . the following analysis then treats these cases in turn .
if @xmath456 then of course also @xmath107 . the bilinear equation ( [ sbilinear ] ) then implies that @xmath460 and hence in particular that the one - form @xmath461 .
this says that @xmath71 defines a @xmath462 structure , rather than an @xmath149 structure , and hence that @xmath71 satisfies a reality ( majorana ) condition @xmath463 .
the scalar bilinears determine that @xmath464 , and since @xmath465 we conclude that @xmath466 and the warp factor is constant @xmath467 .
finally , the bilinear equation ( [ * feqn ] ) and its @xmath151 analogue imply ^3f & = & ( ^6|_1_(2)_1 ) - 6 ^ 6 , which in turn immediately implies that @xmath123 .
this is because the majorana condition @xmath468 implies that the two - form bilinear @xmath469 ( there are no @xmath462-invariant two - forms ) , while the three - form bilinear @xmath470 is real ( corresponding to the unique @xmath462-invariant three - form ) .
we conclude that the warp factor is constant and @xmath123 , so that the killing spinor equation for @xmath71 ( [ spinoreqns ] ) leads to weak @xmath462 holonomy and hence an einstein metric .
the second killing spinor @xmath72 ( for which the analysis is essentially the same ) then of course leads to a sasaki - einstein manifold . alternatively ,
if @xmath457 then we immediately have @xmath471 by computing the square length of the latter using ( [ fierz ] ) .
but since also @xmath472 follows from linear dependence , we also have the additional relation @xmath473 from ( [ depend ] ) .
there is thus only one linearly independent vector , as one expects since we must have an @xmath149 structure .
using the exterior derivatives of the one - form bilinears one can then show that where @xmath474 is non - zero we have that @xmath97 is closed , @xmath475 ( recall that @xmath97 is killing in any case , so this implies that @xmath97 is parallel ) . by contracting @xmath97 into the bilinear equation for @xmath476 and making use of a fierz identity one then proves that @xmath477 .
given that @xmath478 by assumption , this immediately implies that @xmath474 is constant , and hence that @xmath460 .
but then all vectors are identically zero , and we have a contradiction .
thus it must be that @xmath107 and we hence reduce to the previous case , which implies that @xmath2 is sasaki - einstein with @xmath123 and @xmath45 constant .
in section [ pre ] we noted that when @xmath50 we can no longer conclude that equation ( [ gammazero ] ) holds . in this appendix
we study the case @xmath50 but @xmath479 $ ] not being identically zero , in particular showing that there are no regular solutions in this class .
note this is different from the class of @xmath50 geometries discussed in section [ m0limit ] , and can not be obtained by taking the @xmath480 limit of the general @xmath49 equations in the main text .
we begin by defining h & & , which is a function on @xmath2 .
equation ( [ k4eqn ] ) now becomes k & = & h , while the imaginary part of equation ( [ lie ] ) reads _
( m ( k)_n ) & = & - 2h g_7 mn . combining the last two equations
gives [ obataeqn ] _
m_n h & = & -t^2 h g_7 mn , where @xmath481 .
notice that @xmath482 is a particular type of gradient conformal killing vector .
equation ( [ obataeqn ] ) was studied by obata in @xcite . in particular
, he proved that if a complete riemannian manifold of dimension @xmath483 admits a non - constant function @xmath245 satisfying ( [ obataeqn ] ) , where @xmath484 is ( without loss of generality ) a positive constant , then it is necessarily isometric to a round sphere of radius @xmath485 . thus we immediately conclude that if @xmath245 is not identically zero , @xmath2 is isometric to the round @xmath17 with radius @xmath486 .
now as in section [ m0limit ] , the bianchi identity and equation of motion for @xmath46 imply that @xmath46 is harmonic on the conformally rescaled manifold @xmath487 , where @xmath488 .
but in the case at hand , @xmath489 and the hodge theorem implies there are no harmonic four - forms since @xmath490 .
thus for a non - singular solution in fact @xmath123 , and hence the m - theory four - form @xmath491 .
the equation of motion ( [ eom ] ) then implies that the eleven - dimensional spacetime must be ricci - flat , but this is a contradiction .
j. m. maldacena , `` the large @xmath14 limit of superconformal field theories and supergravity , '' adv .
* 2 * , 231 ( 1998 ) [ hep - th/9711200 ] .
d. martelli and j. sparks , `` @xmath9-structures , fluxes and calibrations in m theory , '' phys .
d * 68 * , 085014 ( 2003 ) [ arxiv : hep - th/0306225 ] .
gauntlett , d. martelli , j. sparks and d. waldram , `` supersymmetric ads@xmath124 solutions of m - theory , '' class .
* 21 * , 4335 ( 2004 ) [ hep - th/0402153 ] .
h. lin , o. lunin and j. m. maldacena , `` bubbling ads space and 1/2 bps geometries , '' jhep * 0410 * , 025 ( 2004 ) [ hep - th/0409174 ] .
d. lust and d. tsimpis , `` supersymmetric ads@xmath5 compactifications of iia supergravity , '' jhep * 0502 * , 027 ( 2005 ) [ hep - th/0412250 ] . j. p.
gauntlett , d. martelli , j. sparks and d. waldram , `` supersymmetric ads@xmath124 solutions of type iib supergravity , '' class .
* 23 * , 4693 ( 2006 ) [ hep - th/0510125 ] .
e. o colgain , j. -b . wu and h.
yavartanoo , `` on the generality of the llm geometries in m - theory , '' jhep * 1104 * , 002 ( 2011 ) [ arxiv:1010.5982 [ hep - th ] ] .
j. p. gauntlett , d. martelli , s. pakis and d. waldram , `` @xmath492structures and wrapped ns5-branes , '' commun .
* 247 * , 421 ( 2004 ) [ hep - th/0205050 ] .
a. lukas and p. m. saffin , `` m theory compactification , fluxes and ads@xmath5 , '' phys .
d * 71 * , 046005 ( 2005 ) [ arxiv : hep - th/0403235 ] .
k. behrndt , m. cvetic and t. liu , `` classification of supersymmetric flux vacua in m - theory , '' nucl .
b * 749 * , 25 ( 2006 ) [ arxiv : hep - th/0512032 ] .
j. bagger and n. lambert , `` modeling multiple m2 s , '' phys .
* d75 * , 045020 ( 2007 ) [ hep - th/0611108 ] .
o. aharony , o. bergman , d. l. jafferis and j. maldacena , `` @xmath493 superconformal chern - simons - matter theories , m2-branes and their gravity duals , '' jhep * 0810 * , 091 ( 2008 ) [ arxiv:0806.1218 [ hep - th ] ] .
a. kapustin , b. willett and i. yaakov , `` exact results for wilson loops in superconformal chern - simons theories with matter , '' jhep * 1003 * , 089 ( 2010 ) [ arxiv:0909.4559 [ hep - th ] ] .
d. l. jafferis , `` the exact superconformal r - symmetry extremizes @xmath229 , '' jhep * 1205 * , 159 ( 2012 ) [ arxiv:1012.3210 [ hep - th ] ]
. n. hama , k. hosomichi and s. lee , `` notes on susy gauge theories on three - sphere , '' jhep * 1103 * , 127 ( 2011 ) .
[ arxiv:1012.3512 [ hep - th ] ] .
d. martelli and j. sparks , `` the large @xmath14 limit of quiver matrix models and sasaki - einstein manifolds , '' phys . rev .
d * 84 * , 046008 ( 2011 ) [ arxiv:1102.5289 [ hep - th ] ] .
s. cheon , h. kim and n. kim , `` calculating the partition function of @xmath494 gauge theories on @xmath232 and ads / cft correspondence , '' jhep * 1105 * , 134 ( 2011 ) [ arxiv:1102.5565 [ hep - th ] ] .
d. l. jafferis , i. r. klebanov , s. s. pufu and b. r. safdi , `` towards the f - theorem : @xmath494 field theories on the three - sphere , '' jhep * 1106 * , 102 ( 2011 ) [ arxiv:1103.1181 [ hep - th ] ] .
a. amariti and s. franco , `` free energy vs sasaki - einstein volume for infinite families of m2-brane theories , '' arxiv:1204.6040 [ hep - th ] .
r. corrado , k. pilch and n. p. warner , `` an @xmath494 supersymmetric membrane flow , '' nucl .
b * 629 * , 74 ( 2002 ) [ arxiv : hep - th/0107220 ] .
i. klebanov , t. klose and a. murugan , `` ads@xmath5/cft@xmath10 squashed , stretched and warped , '' jhep * 0903 * , 140 ( 2009 ) [ arxiv:0809.3773 [ hep - th ] ] .
c. ahn and k. woo , `` the gauge dual of a warped product of ads@xmath5 and a squashed and stretched seven - manifold , '' class .
* 27 * , 035009 ( 2010 ) [ arxiv:0908.2546 [ hep - th ] ] . c. ahn ,
`` the eleven - dimensional uplift of four - dimensional supersymmetric rg flow , '' j. geom .
* 62 * , 1480 ( 2012 ) [ arxiv:0910.3533 [ hep - th ] ] .
j. p. gauntlett , o. a. p. mac conamhna , t. mateos and d. waldram , `` ads spacetimes from wrapped m5 branes , '' jhep * 0611 * , 053 ( 2006 ) [ hep - th/0605146 ] .
m. gabella , d. martelli , a. passias and j. sparks , `` the free energy of n=2 supersymmetric ads@xmath5 solutions of m - theory , '' jhep * 1110 * , 039 ( 2011 ) [ arxiv:1107.5035 [ hep - th ] ] .
n. halmagyi , k. pilch and n. p. warner , `` on supersymmetric flux solutions of m - theory , '' arxiv:1207.4325 [ hep - th ] . j. p. gauntlett and s. pakis , `` the geometry of @xmath495 killing spinors , '' jhep * 0304 * , 039 ( 2003 ) [ hep - th/0212008 ] .
m. gabella , j. p. gauntlett , e. palti , j. sparks and d. waldram , `` ads@xmath124 solutions of type iib supergravity and generalized complex geometry , '' commun .
phys . * 299 * , 365 ( 2010 ) [ arxiv:0906.4109 [ hep - th ] ] .
m. pernici and e. sezgin , `` spontaneous compactification of seven - dimensional supergravity theories , '' class .
quant . grav .
* 2 * ( 1985 ) 673 .
r. emparan , c. v. johnson and r. c. myers , `` surface terms as counterterms in the ads / cft correspondence , '' phys .
d * 60 * , 104001 ( 1999 ) [ arxiv : hep - th/9903238 ] .
m. gabella and j. sparks , `` generalized geometry in ads / cft and volume minimization , '' nucl . phys .
b * 861 * , 53 ( 2012 ) [ arxiv:1011.4296 [ hep - th ] ] .
d. conti and s. salamon `` generalized killing spinors in dimension 5 '' trans .
( 2007 ) , no .
11 , 5319 - 5343 [ arxiv : math/0508375 ] .
d. berenstein , c. p. herzog and i. r. klebanov , `` baryon spectra and ads / cft correspondence , '' jhep * 0206 * , 047 ( 2002 ) [ arxiv : hep - th/0202150 ] .
i. r. klebanov , s. s. pufu and t. tesileanu , `` membranes with topological charge and ads@xmath5/cft@xmath10 correspondence , '' phys .
d * 81 * , 125011 ( 2010 ) [ arxiv:1004.0413 [ hep - th ] ] .
n. benishti , d. rodriguez - gomez and j. sparks , `` baryonic symmetries and m5 branes in the ads@xmath497/cft@xmath498 correspondence , '' jhep * 1007 * , 024 ( 2010 ) [ arxiv : 1004.2045 [ hep - th ] ] .
o. barwald , n. d. lambert and p. c. west , `` a calibration bound for the m theory five - brane , '' phys .
b * 463 * , 33 ( 1999 ) [ arxiv : hep - th/9907170 ] .
m. cvetic , h. lu , d. n. page and c. n. pope , `` new einstein - sasaki spaces in five and higher dimensions , '' phys .
* 95 * , 071101 ( 2005 ) [ hep - th/0504225 ] .
d. martelli and j. sparks , `` toric sasaki - einstein metrics on @xmath499 , '' phys .
b * 621 * , 208 ( 2005 ) [ hep - th/0505027 ] .
j. p. gauntlett , d. martelli , j. sparks and s. -t .
yau , `` obstructions to the existence of sasaki - einstein metrics , '' commun .
* 273 * , 803 ( 2007 ) [ hep - th/0607080 ] .
d. martelli , j. sparks and s. -t .
yau , `` sasaki - einstein manifolds and volume minimisation , '' commun .
phys . * 280 * , 611 ( 2008 ) [ hep - th/0603021 ] .
m. obata , `` certain conditions for a riemannian manifold to be isometric with a sphere , '' j. math .
japan * 14 * ( 1962 ) 333 - 340 .
j. p. gauntlett and o. varela , `` consistent kaluza - klein reductions for general supersymmetric ads solutions , '' phys .
d * 76 * , 126007 ( 2007 ) [ arxiv:0707.2315 [ hep - th ] ] .
o. lunin and j. m. maldacena , `` deforming field theories with @xmath500 global symmetry and their gravity duals , '' jhep * 0505 * , 033 ( 2005 ) [ hep - th/0502086 ] .
gauntlett , s. lee , t. mateos and d. waldram , `` marginal deformations of field theories with ads@xmath5 duals , '' jhep * 0508 * , 030 ( 2005 ) [ hep - th/0505207 ] .
b. kol , `` on conformal deformations , '' jhep * 0209 * , 046 ( 2002 ) [ hep - th/0205141 ] . | we analyse the most general @xmath0 supersymmetric solutions of @xmath1 supergravity consisting of a warped product of four - dimensional anti - de - sitter space with a seven - dimensional riemannian manifold @xmath2 .
we show that the necessary and sufficient conditions for supersymmetry can be phrased in terms of a local @xmath3-structure on @xmath2 .
solutions with non - zero m2-brane charge also admit a canonical contact structure , in terms of which many physical quantities can be expressed , including the free energy and the scaling dimensions of operators dual to supersymmetric wrapped m5-branes .
we show that a special class of solutions is singled out by imposing an additional symmetry , for which the problem reduces to solving a second order non - linear ode . as well as recovering a known class of solutions , that includes the ir fixed point of a mass deformation of the abjm theory
, we also find new solutions which are dual to cubic deformations .
in particular , we find a new supersymmetric warped ads@xmath4 solution with non - trivial four - form flux .
1.5 cm * @xmath0 supersymmetric ads@xmath5 solutions of m - theory * 1.5 cm maxime gabella@xmath6 , dario martelli@xmath7 , achilleas passias@xmath7 and james sparks@xmath8 + @xmath6__institut de physique thorique , cea / saclay + 91191 gif - sur - yvette cedex , france + _ _ 0.8 cm @xmath7__department of mathematics , king s college , london , + the strand , london wc2r 2ls , united kingdom + _ _ 0.8 cm @xmath8__mathematical institute , university of oxford , + 24 - 29 st giles , oxford ox1 3 lb , united kingdom + _ _ 0.8 cm _ [email protected] , [email protected] , [email protected] , [email protected]_ 1.0 cm |
imaging soft x - ray observations have established that nearly two thirds of clusters of galaxies , particularly those with cd galaxies , show `` central excess x - ray emission '' ( jones & forman 1984 ; edge & stewart 1991ab ; edge , stewart , & fabian 1992 ) .
that is , diffuse x - ray emission from their intracluster medium ( icm ) exhibits centrally peaked surface brightness profiles when compared with @xmath3 models ( e.g. , sarazin 1988 ) which describe an isothermal gas sphere hydrostatically confined in a king - type potential ( king 1972 ) .
since the x - ray volume emissivity from the icm is proportional to the icm density squared , the excess x - ray emission generally implies an excess icm density in the cluster core region .
this may in turn result from two apparently independent ( and mutually non - exclusive ) mechanisms .
one is decrease in the average icm temperature ( or presence of a cooler icm phase ) in the cluster core region , particularly due to cooling flows ( see fabian 1994 for a review ) ; then the @xmath3-model is no longer valid since the gas is not isothermal , and the central icm density increases so as to maintain the pressure .
the other is deviation of the gravitational potential from king - type ones , in such a way that the potential exhibits an additional drop in the cluster center which attracts an excess amount of icm . among these two mechanisms ,
the former effect is usually thought to be significant , because the icm is often multi - phased in the central regions of cd clusters ( fabian et al . 1994 ; fukazawa et al .
1994 ; matsumoto et al .
1996 ) , and the radiative cooling time of the icm in such regions is estimated to be generally shorter than the hubble time ( fabian 1994 ) .
in contrast , the effect of potential shapes on the central x - ray excess has so far remained inconclusive , because the imaging x - ray telescopes were available only in energy ranges below @xmath7 kev where the x - ray brightness is usually more sensitive to the temperature decreases than to the detailed potential shapes . with ( tanaka , inoue & holt 1994 ) ,
the energy range available for x - ray imaging studies has been expanded up to 10 kev , and the spectroscopic capabilities have been drastically improved .
observations of the fornax cluster with have revealed that the strong central excess brightness around its cd galaxy , ngc 1399 , is a consequence of the shape of the gravitational potential , which requires two distinct spatial scales corresponding to the entire cluster and the central galaxy ( makishima 1995 , 1996 ; ikebe 1995 ; ikebe et al .
the optical surface photometry of ngc 1399 shows essentially the same effect ( schombert 1986 ) .
similarly , the prominent central excess emission from the hydra - a cluster , previously known in soft x - rays ( david et al.1990 ) , has been detected in harder x - rays up to 10 kev ( ikebe et al . 1997 ) .
therefore the potential of this cluster is inferred to exhibit a marked central dimple to confine the excess hot icm . even in cd clusters with prominent central cool emission , such as the centaurus cluster ( fukazawa et al .
1994 ; ikebe 1995 ) and the virgo cluster ( matsumoto et al . 1996 ) , non - cooled icm was found to fill a major fraction of the cluster core volume ; two - phase modeling of the icm in the centaurus cluster has yielded a potential profile with a noticeable feature in the central @xmath8 kpc ( ikebe 1995 ) .
similar effects have been observed from smaller systems as well .
the giant elliptical galaxy ngc 4636 was revealed to sit at the center of two concentric x - ray brightness components of similar temperatures , having distinct scales of @xmath9 kpc and @xmath10 kpc ( matsushita 1997 ; matsushita et al .
1996 , 1997 ) .
therefore ngc 4636 must be surrounded by hierarchically nested two gravitating halos , one corresponding to the galaxy itself and the other to a galaxy group which is otherwise difficult to recognize .
the nested two - component x - ray brightness distributions have also been detected with _ rosat _ from several galaxy groups ( mulchaey & zabludoff 1997 ) .
thus , recent x - ray observations suggest that the gravitational potential shape has a significant effect on the phenomenon of central excess x - ray emission
. it may be that the cluster potential generally consists of doubly - nested hierarchical components , corresponding to the cluster and the cd galaxy .
alternatively , the potential shape of a self - gravitating system may normally exhibit a significant central cusp , as suggested by some gravitational lensing measurements ( e.g. , wu & hammer 1993 ; miralda - escude & babul 1995 ) and @xmath11-body simulations ( navarro , frenk , & white 1996 ) .
if either ( or both ) of these possibilities are true , the central excess brightness of clusters should be detectable in hard x - rays as a rule rather than the exception . in this paper , we examine this possibility using the prototypical central - excess cluster , abell 1795 ( hereafter a1795 ) .
a1795 , a richness class 2 cluster at a redshift of @xmath12 , is one of the most equilibrated clusters with nearly circular x - ray isophotes ( briel & henry 1996 ) . the soft x - ray brightness within @xmath13 of its cd galaxy ( mcg + 05 - 33 - 005 ) exhibits a strong excess , above a @xmath3-model fitted to the outer - region brightness profile ( jones & forman 1984 ; briel & henry 1996 ) . in the 0.53 kev band ,
the central excess emission is so strong that it accounts for 3050% of the total luminosity in the same energy band .
mainly based on these morphological results in soft x - rays , the cooling flow rate in this cluster was estimated to be as large as @xmath14 yr@xmath15 with _
( edge , stewart , & fabian 1992 ) , or @xmath16 yr@xmath15 with ( briel & henry ) .
spectroscopically , the icm of a1795 has an emission - weighted mean temperature of @xmath17 kev ( hatsukade 1990 ; tsuru 1992 ) , together with a central cool emission component ( fabian et al .
1994 ; mushotzky et al .
1995 ; briel & henry 1996 ) .
the spectroscopy with ( fabian et al .
1994 ) provided a somewhat lower cooling - flow rate of @xmath18 yr@xmath15 .
outer regions of a1795 are reported to be rather isothermal ( mushotzky 1994 ; ohashi et al .
1997 ) . besides all these x - ray studies , an optical surface photometry of a1795 ( johnstone , naylor , & fabian 1991 ) also illustrates an elongated large - scale halo around the cd galaxy , which is similar to that of the fornax cluster ( schombert 1986 ) .
clearly , this object is ideal to our search for the hard x - ray excess in the central region . using the data of a1795
, we have actually detected the central excess brightness in the hard energy band up to 10 kev , and found that the excess percentage in the 310 kev band is very similar to those found below @xmath7 kev .
we also analyzed the archival data and obtained a consistent result . unless stated otherwise , we use the @xmath19% confidence level throughout this paper , and assume the hubble constant to be @xmath20 @xmath21 km s@xmath15 mpc@xmath15 .
we also employ the solar abundance ratios from anders & grevesse ( 1989 ) , with fe / h@xmath22 by number .
observations of a1795 were conducted for @xmath23 ks on 1993 june 16 - 17 , during the pv ( performance verification ) phase .
the field of view was centered at @xmath24 , @xmath25 ( j2000 ) .
the data from the gis ( gas imaging spectrometer ; ohashi et al .
1996 ; makishima et al .
1996 ) were taken in the normal ph mode , and those of the sis ( solid - state imaging spectrometer ; burke et al .
1994 ) were acquired in 4 ccd bright / faint mode .
the same dataset was already utilized by fabian et al .
( 1994 ) , mushotzky ( 1994 ) , mushotzky et al .
( 1995 ) , and ohashi et al .
( 1997 ) .
we screened the gis and sis events with the standard data selection criteria .
specifically , we required the telescope view direction above the earth s limb to be @xmath26 for the gis data , @xmath26 for the night - time sis data , and @xmath27 for the day - time sis data to avoid the light leakage .
the minimum cutoff rigidity was set to be 8.0 gev c@xmath15 for both the gis and the sis .
these screening processes have provided @xmath28 ks and @xmath29 ks of good exposures with the gis and the sis , respectively .
figure 1 shows gis images of a1795 in the 0.73 kev and 310 kev bands , where the cosmic x - ray background and internal detector background have been subtracted by utilizing observations onto several blank sky fields .
the two images are very similar to each other , both having a good circular symmetry .
the x - ray brightness centroid is found at @xmath30 and @xmath31 ( j2000 ) , with a typical uncertainty of @xmath32 .
this agrees with the optical position of mcg + 05 - 33 - 005 within @xmath33 .
the x - ray emission extends at least to @xmath34 : considering that @xmath32 corresponds to @xmath35 kpc , the icm emission is thus detectable at least up to @xmath36 mpc . a seyfert 1 galaxy , exo 1346.2 + 2645 ,
is located about @xmath37 south of the a1795 center and is included in the field of view of both the gis and sis . since the x - ray flux of this object is about 6% of that of a1795 , we exclude a region around it in the following data analysis .
there is yet another active galactic nucleus ( agn ) to @xmath38 south of a1795 , emitting a flux which is 2% of a1795 .
we accordingly limit our analysis to within @xmath39 from the cluster centroid .
although the extended point spread function ( psf ) of the x - ray telescope ( xrt ; serlemitsos et al . 1995 ; tsusaka et al .
1995 ) causes some fraction of x - rays from these agns to leak into the analysis region , the effect is confirmed to be negligible compared with the photon counting errors .
using both the gis and sis , we accumulated photons in five annular regions of @xmath40 , @xmath41 , @xmath42 , @xmath43 , and @xmath44 , where @xmath45 is the projected radius from the x - ray brightness centroid .
however , due to the smaller field of view of the sis ( @xmath46 ) and an offset of the cluster centroid from the detector center ( fig.1 ) , the sis data only partially covers the regions outside @xmath47 .
the background was subtracted from each spectrum , in a similar way as mentioned in 2 .
corresponding spectra from the two gis detectors ( gis2 and gis3 ) were then combined together ; so were those from the two sis detectors ( sis0 and sis1 ) .
we show these spectra in figure 2 , where we combined four outer - region spectra for presentation .
as seen in figure 2 , all these spectra exhibit strong fe - k emission lines .
in addition , both instruments reveal somewhat enhanced emission at @xmath48 kev in the spectra from the inner region ( @xmath40 ) , which is indicative of an additional cool component as already reported by fabian et al .
however in the outer regions ( @xmath49 ) , the spectral shapes do not depend significantly on the radius .
thus , the icm of this cluster has relatively uniform properties in outer regions , in agreement with reports by mushotzky ( 1994 ) and ohashi et al .
( 1997 ) , who studied temperature and abundance distributions using the same data .
it is well known that the extended psf of the xrt causes difficulties in the spectral analysis of extended sources ( e.g. takahashi et al .
1995 ; markevitch et al .
1996 ; ikebe et al . 1997 ; ohashi et al .
1997 ) . in the case of clusters of galaxies ,
x - rays from bright central regions are scattered by the psf into fainter peripheral regions of the target , thus contaminating the outer - region spectra . furthermore , the scattered photons form a harder spectrum than their original one , because the psf has more prominent outskirts in higher energies .
this often produces an artificial outward temperature increase . in spite of these problems
, we first performed a brief spectral analysis in a conventional way just to grasp gross properties of the x - ray emission .
that is , we jointly fitted the sis and gis spectra at each radius with the single temperature raymond - smith emission model ( raymond & smith 1977 ) , while the psf effects are not taken into account .
the absorption column density and the heavy - element abundances were left free , whereas the abundance ratios were constrained to obey the solar ratios .
this conventional analysis yielded mostly acceptable fits to the spectra , and the temperature turned out to be in the range 5 8 kev with a mild outward increase due to the instrumental artifact ( takahashi et al .
also there is a slight temperature drop at the center , due probably to the cool component ( fabian et al .
1994 ; briel & henry 1996 ) . except these
, the cluster appears to be rather isothermal in agreement with mushotzky ( 1994 ) and ohashi et al .
( 1997 ) who used the same dataset .
the metallicity is consistent with being radially constant at @xmath50 solar .
there is some hint of excess absorption in the inner regions up to a few times @xmath51 @xmath52 , above the galactic line - of - sight value of @xmath53 @xmath52 , but the effect is not as significant as in some other clusters ( white et al . 1991 ) .
> from these fits , we estimate the x - ray luminosity of a1795 to be @xmath54 erg s@xmath15 in 0.53 kev or @xmath55 erg s@xmath15 in 210 kev , both integrated up to @xmath56 ( @xmath57 mpc ) and corrected for absorption . the former value is close to the 0.53 kev luminosity of @xmath58 erg s@xmath15 obtained with within @xmath56 ( jones & forman 1984 ) .
similarly , our 210 kev luminosity is in a good agreement with the value of @xmath59 erg s@xmath15 ( hatsukade 1989 ) , obtained with the lac in the same 210 kev range but over a much wider field of view of @xmath60 .
therefore , contribution to the x - ray luminosity from regions outside @xmath61 is inferred to be negligible .
note that these and measurements both refer to @xmath62 , and that these luminosities are approximately free from the psf effects . in order to derive the icm temperature distributions considering the psf effects , we must solve the spectral couplings among different regions on the focal plane ( e.g. markevitch et al .
1996 ; ikebe et al .
however such a detailed spectral investigation is beyond the scope of this paper .
instead , we examine the spectra against the hypothesis of constant icm temperature , taking the psf effects fully into account . for this purpose
, we performed a monte - carlo simulation ( takahashi et al .
1995 ; ikebe 1995 ) which converts a model cluster into the simulated data .
that is , we generate a number of monte - carlo photons by specifying their spatial and spectral probability distributions .
the photons are then `` reflected '' by the xrt and `` detected '' either by the gis or the sis , according to the known instrumental responses .
this tells us how a `` known '' cluster looks like when observed with .
the xrt+gis angular response is represented by actual images of cygnus x-1 ( a bright galactic x - ray binary ) obtained with the gis ( takahashi et al .
however cygnus x-1 is too bright for the sis , so that this monte - carlo method can not be applied at present to the sis data . in the simulation ,
we employ the raymond - smith model to express the spectral probability distribution of monte - carlo photons .
we assume the model cluster to have a free but spatially constant temperature , and fix the model cluster abundance at @xmath63 solar according to the results of 3.2 .
the spatial probability distribution of monte - carlo photons is modeled by a @xmath3-model , which describes the radial x - ray surface brightness profile as @xmath64^{-3\beta+ \frac{1}{2}}\ ] ] where @xmath65 is the central brightness , @xmath45 is again the projected radius , @xmath66 is the core radius , and @xmath3 is so - called @xmath3 parameter .
we chose @xmath67 and @xmath68 for our simulation .
later we in fact find that the brightness profiles of a1795 can not be described by a single-@xmath3 model .
however , the @xmath3-model employed here can describe the and brightness profiles of a1795 to within 5% as we show later .
the difference between this approximation and the true brightness distribution produces negligible effects in the simulated spectra .
we generated about @xmath69 simulated gis events , which were subsequently sorted into 5 annular - region spectra exactly in the same way as the actual data .
then , at each radius range , we directly compared the actual and simulated gis spectra , by subtracting the latter ( after an appropriate re - scaling ) from the former in channel - by - channel way , and evaluating residuals via the chi - square tests .
the re - scaling factor was adjusted at each radius to minimize the chi - square .
we then varied the model cluster temperature , to optimize the overall fit goodness .
the model cluster with a temperature of @xmath70 kev has been found to best reproduced the gis spectra , yielding @xmath71 = 82/106 , 92/106 , 65/106 , 53/106 and 82/106 , from the inner to outer annular regions .
therefore we conclude that the observed gis spectra are consistent with those expected from an isothermal cluster at a temperature of 6.2 kev , when the psf effects are fully taken into account .
this `` best - fit '' temperature is estimated to be uncertain by @xmath50 kev .
figure 3 shows the ratio ( instead of difference ) between the actual and simulated gis spectra , in the innermost and @xmath72 radius ranges . as expected from the above results , the ratios are approximately constant across the energy range .
although the overall cluster is nearly isothermal as confirmed above , the central cool component detected by fabian et al .
( 1994 ) reveals itself in the form of enhanced fe - l complex ( at @xmath48 kev ) at about 1 kev , in the innermost spectra of figure 2 and the spectral ratios of figure 3a . to investigate this cool component , we extracted the sis and gis spectra from the inner region of @xmath73 .
this region , having by far the highest brightness , is little ( @xmath74% ) affected by photons which are scattered from outer regions by the psf into this region .
( however , outer regions are strongly affected by photons from the central region . )
we are therefore justified to analyze the spectra using ordinary energy responses calculated at the position of the cluster centroid , without considering x - rays from outer regions due to the psf effects .
we tried to fit these spectra jointly with a raymond - smith model whose temperature is fixed at 6.2 kev , according to the result of 3.3 .
however as shown in table 1 , the fit was totally unsuccessful even when the absorbing column density and the overall metallicity were allowed to float .
we then let the temperature to vary freely .
this gave a much improved fit together with a somewhat lower temperature ( table 1 ) , which is close to that measured with _
( 5.3 kev ) , but the fit still remained unacceptable .
we therefore introduced another raymond - smith component of lower temperature , which is constrained to have the same abundance and the same absorption as the first component .
this two - temperature ( 2 t ) fit has been mostly successful as shown in table 1 .
the hot - component temperature ( @xmath75 kev ) , thus derived jointly with the two instruments , agrees with that obtained via the simulation in 3.3 .
we also tried another two - temperature modeling after fabian et al .
( 1994 ) , who assumed that the hot component is absorbed with the galactic line - of - sight column whereas the cool component is further absorbed by an excess column density .
this model gave as good a fit ( @xmath76 ) to the data , and the derived results are essentially the same as obtained above .
> from the two - temperature fit , we calculate the 0.53 kev luminosity of the cool component to be @xmath77 erg s@xmath15 , after removing absorption .
this is @xmath78% of the total 0.53 kev luminosity described in 3.2 .
the alternative modeling used by fabian et al .
( 1994 ) yields a consistent value ; @xmath79 erg s@xmath15 also after removing absorption .
when these values of @xmath80 are substituted into equation ( 2 ) of the review by fabian ( 1994 ) , we get mass deposition rates of @xmath81 yr@xmath15 , which are consistent with those obtained via the direct spectral fitting in terms of the cooling - flow model .
we can not however compare our cool - component luminosities with that derived by fabian et al .
( 1994 ) , which is not given explicitly in their publication . in order to more directly confirm the consistency with the result of fabian et al .
( 1994 ) , we tried an isothermal model plus an absorbed cooling flow model on the central ( @xmath40 ) sis0 spectrum .
the mass deposition rate turned out to be @xmath82 yr@xmath15 , with @xmath83 .
this reconfirms the report by fabian et al .
( 1994 ) who used the same sis data .
in figure 4 , we show radial count - rate profiles of a1795 in two representative energy bands , without correction for the xrt vignetting or the xrt psf effects .
they were derived by azimuthally averaging the counts from gis2+gis3 and sis0 separately around the x - ray centroid , and then subtracting the corresponding backgrounds . in deriving these profiles , we excluded a region of radius @xmath84 centered on the seyfert 1 galaxy exo 1346.2 + 2645 . for comparison
, we also show the instrumental psf of the xrt+gis combination .
although this psf was calculated over the total 0.710 kev band , it can well approximate those in either soft or hard energy band . in order to compare the classical soft - band brightness profiles of a1795 against the hard - band ones that have been obtained for the first time ,
we have normalized the latter to the former at each radius .
the results , or radial profiles of the spectral hardness ratio , are presented in figure 4c for the gis and sis separately .
they are very flat ; the gis profile is almost constant within @xmath85% , except beyond @xmath86 where the instrumental artifact mentioned in 3.2 begins to affect the data .
the sis hardness ratio indicates a slight drop in @xmath87 ; this is presumably because the sis has a higher spectral sensitivity and a better spatial resolution to detect the central cool emission , whose presence was already confirmed in 3.4 .
the constancy of the hardness - ratio profiles implies that the central excess emission that has been observed in the soft x - ray band should exist also in the hard x - ray band .
this inference is consistent with the results obtained in 3 that the spectral changes at the cluster center are rather modest .
two obvious criticisms may be addressed to the above inference .
first , the central excess emission may have too soft a spectrum to be detected with .
this is however unlikely , since the average icm temperature of a1795 derived with through the cluster center , 2.9 kev ( briel & henry 1991 ) , is clearly high enough for the bandpass .
alternatively , the central soft x - ray excess may be undetectable with because of the heavy smearing by the psf . in order to examine this issue in the simplest manner ,
an appropriately scaled 310 kev radial profile of a point source was added to , or subtracted from , the hard - band radial profile of a1795 .
this is because any central excess emission should appear almost pointlike to .
for the point source , we used the actual gis data of cygnus x-1 observed at the same position as a1795 ( takahashi et al .
the hard - band profile of a1795 , thus modified , was then divided again by the soft - band gis profile of a1795 .
the results shown in figure 4d clearly confirm that the hardness ratios deviate significantly from the actually measured values when the point - source contribution ( either positive or negative ) exceeds @xmath7% of the a1795 counts integrated within @xmath39 .
> from these considerations , we conclude that the hardness - ratio profile obtained with is very sensitive to small differences in the fractional central brightness excess between the two energy bands .
given that the relative excess is 3050% in energies below @xmath7 kev ( jones & forman 1984 ; briel & henry 1996 ) , the relative excess should be at least 25% in hard x - rays . to strengthen the conclusion derived in 4.1 , we need to quantify the x - ray count - rate profiles .
however , an x - ray image obtained with is a very complex integral transform ( not a simple convolution ) of the original surface brightness distribution on the sky plane , because the psf of xrt is axially non - symmetric and heavily position dependent .
accordingly , instead of trying to inversely solve this integral transform , we again employ the monte - carlo simulation used in 3.3 .
that is , we generate simulated cluster data according to the @xmath3 model of equation ( 1 ) , in which the xrt+gis psfs are again represented by the actual gis image of cygnus x-1 observed at various detector positions ( takahashi et al .
we then derive simulated gis count - rate profiles in several energy bands , in the same way as has been done with the actual gis data . by evaluating their agreements with the actual gis radial profiles , we search for the best - fit @xmath3 model .
we added a systematic error of 3% to the simulated radial profiles .
this technique has been developed and applied to the data by takahashi et al .
( 1995 ) , ikebe ( 1995 ) , markevitch et al .
( 1996 ) , honda et al .
( 1996 ) , matsuzawa et al .
( 1996 ) , ohashi et al .
( 1997 ) , and ikebe et al .
( 1997 ) .
we first applied this @xmath3-model to the gis radial count - rate profiles over the whole radius range of @xmath88 .
as summarized in table 2 , rather flat @xmath3 models with small core radii provided the best - fit solutions in the soft ( 0.73 kev ) and hard ( 310 kev ) energy bands . however , none of these fits are acceptable , exhibiting subtle but statistically significant fit residuals as shown in figure 5a-5b . these failures of the fit can be ascribed obviously to the presence of the central excess emission , as already pointed out by jones & forman ( 1984 ) and briel & henry ( 1996 ) .
we also analyzed the processed archival pspc data of a1795 , provided by the guest observer facility at the nasa / gsfc .
these data , acquired on 1991 july 1 and 2 , comprise a part of the datasets used by henry & briel ( 1996 ) .
the pointing position is @xmath89 , @xmath90 ( j2000 ) , just near the center of the cluster . by using the exposure map
, we corrected the broad - band ( 0.22 kev ) image for vignetting and variations in the exposure across the field of view .
we then derived the azimuthally - averaged radial count - rate profile in the 0.22 kev range , in which the seyfert 1 galaxy ( exo 1346.2 + 2645 ) was omitted by a circular mask of @xmath91 radius .
the background is not subtracted , but considered in the latter model fitting .
we performed a single-@xmath3 model fitting to the derived count - rate profile , over the radius range of @xmath92 .
we approximated the psf with a delta - function , which is justifiable when we ignore spatial structures much finer than @xmath93 ( h. bhringer , a private communication ) .
we included a background component into the fit model assuming it to be free but spatially constant .
as shown in table 2 and figure 5c , the fit turned out to be unacceptable , confirming the report by briel & henry ( 1996 ) . in spite of these failures , we note that the single-@xmath3 model can approximate the observed x - ray profiles to a reasonable accuracy ( e.g. within 510 % ) ; this allowed us to utilize the single-@xmath3 approximation of the brightness profile in running the monte - carlo spectral simulation ( 3.3 ) .
it is still more important to note that the best - fit @xmath3-model parameters ( though unacceptable ) are quite similar between the pspc and the gis , and among the two energy bands of ( table 2 ) .
furthermore , the fit residuals behave essentially in the same manner in the three panels of figure 5 : the data exceed the model at @xmath94 , while the opposite occurs in other regions .
these results support our inference made in the previous subsection that the count - rate profiles do not depend on the energy significantly .
when the central cluster region of @xmath95 radius is ignored , the single-@xmath3 model has given acceptable fits to the gis profiles in the three energy bands ( soft , hard , and total ) , as well as to the 0.22 kev profile .
we present these fits in figure 6 , and summarize the best - fit parameters in table 3 .
clearly , the @xmath3-model parameters derived with the gis are again nearly energy - independent within the respective errors .
they also agree with those derived with the pspc which in turn are close to those obtained by briel & henry ( 1996 ) who used a larger datasets , discarding the central @xmath95 .
furthermore , the previous results from the ipc ( jones & forman 1984 ) are roughly consistent with those obtained here , as we can see in table 3 . in a word ,
regardless of the used instrument or the employed energy range , we obtain a core radius @xmath96 and a slope of @xmath97 for a1795 , as long as the central @xmath95 region is ignored . when extrapolated inwards , these best - fit @xmath3 models fall significantly below the observed brightness ( fig .
6 ) , thus reconfirming the prominent central excess emission .
the last column of table 3 gives a rough measure of this central excess , integrated within @xmath98 and normalized to the total counts integrated up to a projected radius of @xmath39 .
thus , the measurements with , , and consistently indicate a central excess emission that reaches @xmath99% of the total x - ray emission within @xmath39 , although the excess may decrease mildly with energy , which is most likely to be caused by the cool emission component . combining these results with those obtained in the previous subsections , we draw the following two conclusions of basic importance .
one is that the central excess emission of a1795 it too strong to be smeared out by the psf of , and the other is that the central x - ray excess emission of this cluster is significantly seen over the entire 0.210 kev energy range spanned by and . in order to describe the central excess emission more quantitatively
, we further attempted to fit the count - rate profiles with a sum of two @xmath3 model components , a more extended one representing the emission outside @xmath95 , and a compact one representing the central excess .
we have found that this double-@xmath3 model gives acceptable fits to the profiles obtained with the pspc and the gis , as summarized in table 4 and shown in figure 7 .
all the observed count - rate profiles have thus yielded very similar parameter values for the two @xmath3-model components .
the more extended component is similar to that obtained in the single-@xmath3 fits excluding the central region ( table 3 ) , while the compact component has a core radius which is close to that obtained with the single-@xmath3 fits including the central region ( table 2 ) .
we remark that the very central ( @xmath100 ) region of the cluster may remain unresolved even with the pspc .
consequently , the core radius of @xmath101 for the compact component may be subject to instrumental angular resolutions , and the finer - resolution x - ray brightness profiles , e.g. those obtained with the hri , may no longer be adequately described by our double-@xmath3 model .
nevertheless , our principal conclusions should remain unchanged , because they all deal with angular scales of @xmath93 or larger . in table 5 , we summarize the luminosity , @xmath102 , carried by the narrower @xmath3 component .
it amounts to @xmath103% of the total ( i.e. , narrow plus wide ) luminosity , @xmath104 , in both the soft and hard energy bands .
this is reflected in the fact that the two @xmath3 components crossing over in figure 7 have similar normalization ratios in the two energy bands .
the excess luminosity defined in this way is systematically larger than those given in table 3 , by up to a factor of @xmath105 .
however , this is simply the matter of how to define the `` excess ''
. through extensive @xmath11-body numerical experiments of structure formation in the cold dark matter ( cdm ) scenario , navarro , frenk , & white ( 1996 ) reported that the computed cdm mass density @xmath106 takes a `` universal halo shape '' of @xmath107 where @xmath108 and @xmath109 are parameters .
makino , sasaki , & suto ( 1997 ) showed that the density of an isothermal gas hydrostatically confined in the potential of equation ( 2 ) can be expressed analytically as @xmath110 where @xmath111 and @xmath112 are also parameters . by numerically projecting the square of this @xmath113 onto 2 dimensions , we can obtain the x - ray surface brightness , @xmath114 , emitted by an isothermal gas sphere confined in a universal cdm halo .
since the cdm distribution of equation ( 2 ) is more centrally peaked than the king - type distribution , the idea of universal cdm halo may explain the central excess of a1795 . in order to examine this possibility , we fitted this @xmath114 to the pspc profile .
however as shown in figure 7d , the fit turned out to be as poor as that with the single-@xmath3 model ( fig.5c ) , with @xmath115 .
the gis profiles could not be fitted with this new model , either .
this result is not surprising , since @xmath114 is in fact fairly close to the ordinary @xmath3-model profile @xmath116 given by equation ( 1 ) , with only moderate excess towards the center ( makino , sasaki , & suto 1997 ) . in short ,
the central excess emission from a1795 is too strong to be explained by the universal halo model as long as the isothermality holds .
through a series of analysis of the radial count - rate profiles of a1795 obtained with the gis and the pspc , we have confirmed that the central excess brightness of this cluster , which was well known previously in the soft x - ray band , does exist also in the 310 kev band with a similar fraction .
if employing @xmath3-models , two distinct components are required to express the projected radial count - rate profiles obtained with the gis in the energy range both below and above 3 kev , as well as that obtained with the pspc .
we also carried out spectroscopic studies of the icm in a1795 using the data .
in essence , the icm in a1795 is isothermal at a temperature of 6.2 kev , except in the innermost region of radius @xmath117 ( @xmath118 @xmath2 kpc ) where an additional cool emission component is seen as already report by fabian et al .
( 1994 ) . employing the two - temperature fit obtained in 3.4
, we calculated the luminosities of the cool component in various energy bands , as listed in table 5 . in the 0.53 kev band ,
these morphological and spectroscopic results compare as @xmath119 ( according to table 3 ) , or @xmath120 ( according to table 5 ) .
these ratios may not necessarily imply a discrepancy between @xmath80 and @xmath102 , since our estimation of @xmath80 depends on a particular modeling ( i.e. , two - temperature assumption ) of the icm .
however the discrepancy becomes much more significant when the comparison is made in the newly explored 310 kev band ; @xmath121 in reference to table 3 , or @xmath122 in reference to table 5
. these small ratios directly result from the fact that the spectral cool component has a temperature of at most @xmath123 kev ( table 1 ) , whereas the morphological excess component has a temperature rather close to the cluster - average value of 6.2 kev .
our results consistently indicate that the whole cluster volume of a1795 is permeated with the hot isothermal icm , including the central region where a small amount of cool component coexists .
this is similar to the case of several other cd clusters ( fukazawa et al .
1994 ; ikebe 1995 ; matsumoto et al . 1996
; ikebe et al .
the central excess brightness of a1795 is inferred to be produced primarily by a central excess in the density , hence pressure , of this hot icm itself , with the cool emission component giving subsidiary effects . to confine this excess icm pressure ,
the gravitational potential in a1795 must exhibit an additional central drop as compared to the king - type potential . as indicated by figure 7d
, this central drop may be much deeper than is predicted by the universal halo model of equation ( 2 ) , as long as the isothermality approximation is valid . referring to the double-@xmath3 solutions which provide acceptable fits to the observed brightness profiles , we quantify the shape of the gravitational potential . specifically , we employ the best - fit double-@xmath3 model obtained in the gis 0.710 kev band ( table 4 ) , and back - project the constituent @xmath3-model components individually into 3-dimensions .
this provides emissivity profiles of the extended and compact components , denoted as @xmath124 and @xmath125 , respectively .
these can be converted to the icm density profile @xmath113 through the relation @xmath126 where @xmath127 is the 3 dimensional radius and @xmath128 is the cooling function for a given temperature @xmath129 and abundance @xmath130 .
we then calculate the radially integrated total gravitating mass , @xmath131 , using the usual equation of @xmath132 the total mass profiles calculated via equation ( 4 ) are presented in figure 8 , where we assumed two alternative cases for the temperature profiles @xmath133 ; either a constant icm temperature at 6.2 kev , or a centrally decreasing profile of the form of @xmath134 , \ ] ] which approximately reproduces the pspc result of briel and henry ( 1996 ) . here
the 3 dimensional radius @xmath127 is in arcminute and the temperature @xmath133 is in kev .
we also present the icm mass profile , @xmath135 , calculated by integrating @xmath113 of equation ( 3 ) . inside @xmath136 kpc ,
the mass curves in figure 8 should be regarded as subject to the limited angular resolution .
finer details of the central density profile could be studied using the hri data , but the lack of associated spectral information hampers the calculation of gravitational potential there . in this paper , we therefore limit our study to the information available with the pspc and .
figure 8 reveals a shoulder - like structure in the @xmath131 profile at @xmath137 kpc , regardless of the assumed temperature profile . as a result
, the total gravitating mass within @xmath138 apparently exhibits an excess by @xmath139 , above the large - scale smooth mass distribution carried by the extended @xmath3-component .
this @xmath140 is thought to be responsible for the central dimple in the gravitational potential . the shoulder - like structure seen in the @xmath141 curve of figure 8 may allow two slightly different interpretations .
one possibility is that the overall mass distribution is made up of a single mass component , presumably dominated by dark matter , whose density keeps rising to the center instead of forming a flat core . although equation ( 2 ) , representing one particular example of such cases , failed to explain the x - ray data under the isothermal assumption , some other numerical simulations ( moor et al .
1997 ) suggest even a steeper cdm halo profile .
when taking into account the effect of the cool component and baryonic mass associated with the cd galaxy , such a centrally - steepening cdm halo may still be consistent with the observation .
alternatively , the @xmath131 profile in figure 8 may be interpreted as exhibiting a gravitational hierarchy between the overall cluster and the cd galaxy .
such a nested potential structure was suggested or assumed previously ( e.g. thomas et al .
1987 ) , and has been confirmed in the fornax cluster ( makishima 1995 ; ikebe et al .
1996 ) , the elliptical galaxy ngc 4636 ( matsushita 1997 ; matsushita et al . 1996 , 1997 ) , and possibly in about 8 galaxy groups ( mulchaey & zabludoff 1997 ) .
then , the region inside @xmath138 is considered to belong to the cd galaxy , while the outside region to the entire cluster . in reference to the above @xmath140 and the optical b - band luminosity of the cd galaxy ( @xmath142 in unit of the solar b band luminosity ) , the mass - to - light ratio within @xmath138 becomes @xmath8 .
the hierarchical potential structure is also suggested by the optical observations by johnstone , naylor , & fabian ( 1991 ) , who reported that the _ i_-band surface brightness of a1795 consists of two components ( though the extended one is quite elongated ) , crossing over at @xmath143 kpc ; this is close to @xmath144 .
furthermore , @xmath140 estimated above is similar to the virial mass of the cd galaxy ( @xmath145 ) , derived from the stellar velocity dispersion ( @xmath146 km sec@xmath15 ; oegerle & hoessel 1991 ; mcnamara & oconnell 1993 ) . in comparison , we have @xmath147 kpc and @xmath148 in the fornax cluster .
the hierarchy interpretation relies much on the fact that the logarithmic curvature of @xmath131 becomes apparently concave over the region where the shoulder - like structure is present ( @xmath149 kpc ) .
however , this may be a model dependent artifact , produced e.g. by cross - over between the two components in the employed double-@xmath3 model . as a matter of fact
, it is rather difficult to assess the reality of this feature , due to complex error propagation from the data to the mass curve . instead of attempting to do so
, we here resort to fit the observed 0.710 kev gis brightness profile by a rather arbitrary empirical function of the form @xmath150
\left [ 1 + \left ( \frac{r}{r_{2 } } \right)^2 \right]^{-c_{2}}\ ] ] where @xmath65 , @xmath151 , @xmath152 , @xmath153 , and @xmath154 are parameters . the best - fit of this empirical model is found with @xmath155 , @xmath156 , @xmath157 , and @xmath158 , yielding @xmath159 , which falls between those from the single-@xmath3 fit and the double-@xmath3 fit .
the mass profile derived from this ad - hoc model , via back projection and equation ( 4 ) , is also plotted in figure 8 .
it reveals a very similar feature with concave curvature , even though equation ( 6 ) does not involve particular angular scales corresponding to @xmath160 and the logarithmic derivative of equation ( 6 ) changes monotonically with @xmath45 .
this adds to the reliability of the concave feature seen in figure 8 , and reinforce the interpretation of our results in terms of the hierarchical potential structure between the cluster and the cd galaxy .
finally , some mention may be made on cooling flows .
the dominance of hot icm in the cluster center region , with a smaller amount of cooler component coexisting with it , is in itself consistent with the picture of inhomogeneous cooling flows ( fabian 1994 ) .
the short cooling time of 2.5 gyr within the central 80 kpc , as estimated by edge , stewart , & fabian ( 1992 ) , remains essentially unchanged by our results .
however , causal relation between the nested potential structure revealed here and the cooling flows in a1795 remains unclear .
the hierarchical mass distribution may have formed first during the cluster formation , and provided a favorable environment for cooling flows to take place .
if so , the cooling flow may be regarded as the result , rather than the cause , of the central excess x - ray emission .
alternatively , a certain fraction of the excess mass within @xmath138 may have deposited from the cooling flow itself . in summary
, we have discovered that the central excess x - ray emission from a1795 is seen in the hard x - ray band above 3 kev , with a similar fraction as seen in the soft x - rays .
this property is mainly due to a central excess in the gravitating mass , by @xmath161 , which is likely to be associated wit the cd galaxy .
we thank all the members of the team .
we would also like to thank hans bhringer , wolfgang voges , and doris neumann for their helpful discussions and comments on the analysis of the pspc data .
we are also grateful to a. c. fabian , r. f. mushotzky , s. sasaki , and y. suto for helpful comments .
this research has made use of data obtained through the high energy astrophysics science archive research center online service , provided by the nasa - goddard space flight center .
anders , e. , & grevesse , n. 1989 , geochim . cosmochim .
acta 53 , 197 figure 1 : background - subtracted gis2 + gis3 images of a1795 in ( a ) 0.73 kev and ( b ) 310 kev .
the images have been smoothed with two - dimensional gaussian filters ( @xmath162 ) .
no correction has been made for the xrt vignetting or partial shadows due to the detector support ribs .
the contours are with the same logarithmic steps in the two images .
the sis field of view is indicated by the dashed line in each plot , and the position of the seyfert 1 galaxy exo 1346.2 + 2645 is marked by a cross .
figure 2 : ( a ) spatially sorted gis2 + gis3 spectra of a1795 for @xmath163 and @xmath164 regions . the background has been subtracted , but the detector response has not been removed .
( b ) same as ( a ) but for the sis0 + sis1 spectra .
figure 3 : ( a ) the gis spectrum in the @xmath40 region ( same as shown in fig.2a ) , displayed in the form of channel - by - channel ratios to the simulated gis spectrum .
the monte - carlo simulation was done taking fully into account the psf effects , assuming a constant icm temperature of 6.2 kev and a constant abundance of 0.3 solar .
the model birghtness profile was represented by a single @xmath3-model with a core radius of @xmath165 and @xmath166 .
eerror bars represent @xmath167 due to photon counting statistics in the actual spectrum , while those associated with the simulated spectrum is negligbly small .
( b ) same as ( a ) , but for the @xmath168 region .
figure 4 : ( a ) background - subtracted radial count - rate profiles of a1795 obtained with gis2 + gis3 in 0.73 kev and 310 kev .
the profiles are corrected for neither the xrt vignetting nor the xrt psf effects .
the 0.710 kev xrt psf at the detector center is also plotted for comparison .
( b ) same as ( a ) but for sis0 .
( c ) the radial profiles of the 310 kev to 0.73 kev brightness ratio ( or the spectral hardness ratio ) of a1795 , derived with gis2 + gis3 and sis0 .
( d ) the same gis ratio as in ( c ) , but with a 310 kev radial profile of a point source , whose integrated count is 3% of that of total cluster , added to or subtracted from the 310 kev gis data .
figure 5 : single-@xmath3 fittings to the radial count - rate profiles of a1795 over the whole radius range of @xmath88 .
( a ) the gis profile in 0.73 kev , where crosses represent the data while circles show the best - fit model convolved with the xrt + gis psf .
the lower panel shows the fit residuals .
( b ) the same as panel ( a ) but , in the 310 kev range .
( c ) the pspc profile in 0.22 kev fitted in the same manner .
figure 7 : two-@xmath3 fittings to the brightness profiles of a1795 over the whole radius range of @xmath88 .
meanings of panels ( a ) through ( c ) are the same as those in figure 5 .
squares and triangles show the wider and narrower @xmath3-components , respectively , while circles indicate their sum .
panel ( d ) is the fit to the pspc data , using the x - ray surface brightness profile which is predicted by the universal halo model of equation ( 2 ) in the text .
figure 8 : radial distributions of the integrated total gravitating mass and the icm mass .
cases 1 and 2 are based on the best - fit double-@xmath3 model in 0.710 kev ( table 4 ) , assuming either a constant icm temperature of 6.2 kev ( case 1 ) or a temperature gradient given by equation ( 5 ) in the text ( case 2 ) .
case 3 is derived by replacing the double-@xmath3 model with another analytical surface - brightness model described with equation ( 6 ) in the text , with the temperature assumed to be constant at 6.2 kev . | using the x - ray data from , spectral and spatial properties of the intra - cluster medium ( icm ) of the cd cluster abell 1795 are studied , up to a radial distance of @xmath0 ( @xmath1 @xmath2 kpc ) .
the icm temperature and abundance are spatially rather constant , although the cool emission component is reconfirmed in the central region .
the azimuthally - averaged radial x - ray surface brightness profiles are very similar between soft ( 0.73 kev ) and hard ( 310 kev ) energy bands , and neither can be fitted with a single-@xmath3 model due to a strong data excess within @xmath4 of the cluster center .
in contrast , double-@xmath3 models can successfully reproduce the overall brightness profiles both in the soft and hard energy bands , as well as that derived with the pspc .
properties of the central excess brightness are very similar over the 0.210 kev energy range spanned by and .
thus , the excess x - ray emission from the core region of this cluster is confirmed for the first time in hard x - rays above 3 kev .
this indicates that the shape of the gravitational potential becomes deeper than the king - type one towards the cluster center .
radial profiles of the total gravitating matter , calculated using the double-@xmath3 model , reveal an excess mass of @xmath5 within @xmath6 kpc of the cluster center .
this suggests a hierarchy in the gravitational potential corresponding to the cd galaxy and the entire cluster . received ; accepted _ subject headings _ : galaxies : clusters : individual ( abell 1795 ) galaxies : elliptical and lenticular , cd x - rays : galaxies |
many bright x - ray sources have been discovered in nearby galaxies . those with luminosities @xmath2 exceeding 10@xmath4 erg s@xmath5 , the eddington limit for a black - hole x - ray binary ( bhbs ) with a black - hole ( bh ) mass of @xmath9 , are often referred to as ultraluminous x - ray sources ( ulxs , for a recent review see * ? ? ?
* ) . while a very small number of them appear as hyperluminous x - ray sources ( hlxs , especially ) with @xmath10 erg s@xmath5 and are strong candidates for intermediate - mass bhs ( imbhs ) with @xmath11@xmath12 @xcite , most ulxs have @xmath13 erg s@xmath5
the nature of these low - luminosity ulxs is still under strong debate
. they could be imbhs ( e.g. , * ? ? ? * ) , super - eddington low - mass bhs ( e.g. , * ? ? ?
* ) , beamed emission from geometrically thick disks at high accretion rates @xcite or from relativistic jets @xcite , or combination of these effects @xcite . due to their large distances
, there are no confident dynamical mass measurements for ulxs yet @xcite , unlike galactic bhbs , whose dynamical masses of the bhs are constrained to below 20 @xcite .
the spectral and timing properties of galactic bhbs have been well studied , and they are known to exhibit three characteristic x - ray spectral states : the thermal state has a dominant thermal disk and weak fast variability , typically at luminosities above a few percent of the eddington limit ; the hard state has a dominant hard powerlaw ( pl , with photon index @xmath14 typically within 1.42.1 ) and strong fast variability , generally seen at low luminosities ; and the steep - pl state has a powerful pl with @xmath15 and commonly - occuring quasi - periodic oscillations , often seen at very high luminosities @xcite .
these states also often show correlated radio emission @xcite .
thus studies of ulxs often compare their properties with those of galactic bhbs to infer their possible nature .
two transient ulxs were discovered in m31 recently @xcite .
when they reached luminosities above 10@xmath4 erg s@xmath5 , their x - ray spectra appeared harder than those from a standard geometrically thin optically thick disk .
this is often seen in low - luminosity ulxs , and one possible explanation is compton up - scattering of disc photons in a wind or photosphere @xcite .
the above similarity led @xcite to conclude that these two transient ulxs in m31 are members of low - luminosity ulxs , though the latter are mostly persistent with small long - term variability @xcite .
however , these two ulxs are also similar to bhbs with stellar - mass bhs in other aspects , such as the outburst profile and the coupled x - ray and radio behavior .
thus @xcite concluded that many low - luminosity ulxs are stellar - mass bhbs . in @xcite , we classified 4330 sources from the 2xmmi - dr3 catalog @xcite .
in this project , many sources showing interesting behavior but poorly studied in the literature were also discoverd .
we are devoting a series of papers ( e.g. , * ? ? ?
* ; * ? ? ?
* ; * ? ? ?
* ) to present the properties of these sources in detail .
here we continue our study and concentrate on a highly variable ulx in the spiral galaxy m94 ( ngc 4736 ) , which shows some evidence of dips , periodic oscillations , and a state transition , and has a possible optical counterpart from the _ hubble space telescope _ ( _ hst _ ) images .
the source , whose 1@xmath16 positional error is 035 per coordinate from the 2xmmi - dr3 catalog , is only 026 away from ( ra=12h50m48.605s , [email protected]@xmath18 , and the 95% error is 029 ) in the _ chandra _ source catalog ( csc , release 1.1 , * ? ? ?
thus we conclude that they are the same source .
it should also be the off - nuclear source detected in a _
rosat_/pspcb observation in @xcite , because they coincide within 5@xmath18 in position and there are no other bright sources within 20@xmath18 from _ chandra _ observations .
we designate the source as ulx2 hereafter , considering that there is a different ulx ( u30 in * ? ? ?
* ) , which is hardly variable and has a flux comparable to the peak value of our source . in section
[ sec : reduction ] , we describe the analysis of x - ray and optical observations . in section [ sec : res ] , we present the results . in section [ sec : dis ] , we discuss the nature of the source and the cause of its short - term x - ray variability .
the conclusions of our studies are given in section [ sec : conclusion ] . throughout the paper ,
we assume a source distance of 5.2 mpc @xcite .
rccccccccc + 0094360601(x1 ) & 2002 - 05 - 23 & pn / m1/m2 & 2.3@xmath17 & 15/20/20 & 15@xmath18 & 5 + 0094360701(x2 ) & 2002 - 06 - 26 & pn / m1/m2 & 2.5@xmath17 & 4/12/12 & 15@xmath18 & 5 + 0404980101(x3 ) & 2006 - 11 - 28 & pn / m1/m2 & 1.6@xmath17 & 34/46/46 & 15@xmath18 & 112 + + 808(c1 ) & 2000 - 05 - 13 & acis - s3 & 0.8@xmath17 & 47 & 1.5@xmath18 & 17 + 9553(c2 ) & 2008 - 02 - 16 &
acis - i2 & 0.5@xmath17 & 24 & 1.5@xmath18 & 18 + + rp600050n00(r1 ) & 1991 - 06 - 05 & pspcb & 1.2@xmath17 & 91 & 20@xmath18 & 32 + rh600678n00(r2 ) & 1994 - 12 - 07 & hri & 0.9@xmath17 & 112 & 7@xmath18 & 1 + rh600769n00(r3 ) & 1994 - 12 - 25 & hri & 0.9@xmath17 & 27 & 7@xmath18 & 0 lccccccccccc u96r060[1@xmath192]m & 2005 - 05 - 24 & wf3/f336w & 1800 & 6 & [email protected] & 24.65 & 23.93 + u6ea800[1r@xmath192 m ] & 2001 - 07 - 02 & wf3/f450w & 460 & 24 & [email protected] & 23.05 & 22.96 + u96r060[3@xmath194]m & 2005 - 05 - 25 & wf3/f555w & 400 & 40 & [email protected] & 22.16 & 22.21 + u671260[1 - 5]r & 2001 - 04 - 21 & wf2/f656n & 1700 & 8 & [email protected] & 21.13 & 21.20 + u6ea800[3@xmath194]r & 2001 - 07 - 02 & wf3/f814w & 460 & 46 & [email protected] & 21.07 & 21.05 m94 has been observed by many x - ray observatories .
however , because ulx2 is faint and is only 1@xmath17 away from the nucleus around which there are strong diffuse emission and many bright point sources , including the low - ionization - nuclear - emission - region ( liner ) nucleus , we only studied the eight observations ( table [ tbl : obslog ] ) from _ xmm - newton _ , _ chandra _ and _ rosat _ , thanks to their relatively large effective areas and high angular resolutions .
we designate their observations as x1-x3 , c1-c2 , and r1-r3 , respectively ( refer to table [ tbl : hst ] ) . in the _ xmm - newton _ observations , all the three european photon imaging cameras ,
i.e. , pn , mos1 ( or m1 ) , and mos2 ( or m2 ) @xcite were active .
we used sas 12.0.1 and the calibration files of 2013 january for reprocessing the x - ray event files and follow - up analysis .
the data in strong background flare intervals are excluded following the sas thread for the filtering against high backgrounds .
the event selection criteria followed the default values in the pipeline ( see table 5 in * ? ? ? * ) .
that is , for pn spectra , we used events with pattern @xmath21 and flag @xmath22 ( we also extracted pn light curves , which used events with ( flag & 0xfffffef ) @xmath22 ) , and for mos spectra , we used events with pattern @xmath23 and ( flag & 0xfffffeff ) @xmath22 .
the two _ chandra _
observations used the imaging array of the axaf ccd imaging spectrometer ( acis ; * ? ? ? * ) .
ulx2 falls in the back - illuminated chip s3 and the front - illuminated chip i2 in observations c1 and c2 , respectively .
we analyzed the data with the ciao ( version 4.5 ) package and the latest calibration ( caldb 4.5.5.1 ) .
the _ rosat _
@xcite observations were reduced with ftools 6.12 .
we extracted the source emission from a circular region with the radius for each observation given in table [ tbl : obslog ] .
the background in the xmm - newton and rosat observations strongly depends on the distance to the nucleus due to the presence of bright sources near the nucleus and the relatively large point spread functions ( psfs ) of these two observatories , and we estimated it using four circular regions with the same radius and the same distance to the nucleus as the source region .
the _ chandra _
observations have no such problem thanks to their supreme angular resolutions , and we used a single circular region of 10@xmath18 radius near the source to estimate the background .
we extracted both the source and background spectra / light curves from the source and background regions , respectively .
the response files were constructed for the spectral fitting .
we note that we took special care to construct the response file for the _ rosat_/pspcb observation r1 .
the problem is that we used a small source extraction region ( a radius of 20@xmath18 ) in order to minimize the contamination from bright sources near the nucleus , leading to significant psf loss , which is energy dependent .
the response file ( we used the default on - axis version in the heasarc archive ) does not take into account the psf loss , and there is no ftools tool to correct for this .
we estimate the psf loss using the rosat observation rp700130n00 of mkn 501 ( it used the same gain epoch and has a similar ( low ) column density and photon index ( @xmath12.6 ) as our source ) , and correct the count rate in each channel of the spectrum of our source .
such a correction is probably very approximate .
thus the spectral fits to this spectrum might be subject to some systematic error due to this and will be mostly used to infer the luminosity .
we rebinned the spectra to have at least 20 counts in each bin so as to adopt the @xmath24 statistic for spectral fits .
all spectral models used include the absorption described by the wabs model in xspec , with the lower limit of @xmath25 set to be the galactic line - of - sight value of 10@xmath26 @xmath27 @xcite . to help to identify the optical counterpart ( section [ sec : hstdata ] ) , we used x - ray sources within 2@xmath17 from the nucleus to improve the relative astrometry between _ chandra _ and _ hst_. to have good relative positions for these x - ray sources , we carried out the source detection using the ciao task ` wavdetect ` @xcite .
we used only the _ chandra _ observation c1 so that all sources of interest are in a single ccd , which is not the case for observation c2 .
we performed the detection on the 0.38 kev image , which was binned at a subpixel resolution ( 1@xmath288 sky pixel ) to improve the source positions .
the sky region around ulx2 was observed by the wide field and planetary camera 2 ( wfpc2 ) aboard the _ hst _ several times ( table [ tbl : hst ] ) .
we aligned their drizzled images by matching the point sources to those in the f555w image , resulting in the rms deviations @xmath290.04@xmath18 in all cases . to map the x - ray positions onto the _ hst _ images
, we assumed that in the csc and the uv source at the center of the _ hst _ images are the nuclear source and coincident with each other in position ( see , e.g. , * ? ? ? * ) .
because the nuclear uv source is stronger and less subject to stellar contamination at shorter wavelength and thus appears more clear in the f336w image than in the other images , we used the f336w image to determine its position with the astrolib idl procedure ` cntrd ` .
we found that it happens to be only 002 away from the position of determined by us ( section [ sec : redxrayobs ] ) . after correcting for this small relative astrometry
, we found an optical point source that is only 0@xmath301 away from ulx2 and was detected in all filters ( figure [ fig : hstimg ] ) .
we also found three other x - ray sources with optical counterpart candidates within 0@xmath301 , supporting our astrometry correction .
we performed the photometry on the c0f images with the hstphot 1.1 package @xcite , which outputs aperture - corrected vega magnitudes .
erg s@xmath5 , has only the error bar seen and is very close to the second one ( r2 , both in 1994 december ) .
[ fig : srcltlc],scaledwidth=48.0% ] table [ tbl : obslog ] lists the signal to noise ratio ( s / n ) of ulx2 in each observation .
the source was significantly detected in observations x3 , c1 , c2 , and r1 ( s / n@xmath3117 ) , but it was only marginally detected in observations x1 and x2 ( s / n@xmath325 ) and was not detected in observations r2 and r3 ( s / n@xmath331 ) .
figure [ fig : srcltlc ] plots the long - term 0.210 kev absorbed luminosity curve of ulx2 .
the spectral models that we assumed to derive the luminosities are those shown in figure [ fig : indlcsp ] for the bright observations ( x3 , c1 , c2 and r1 ) and a pl for the other faint observations , which will be described in section [ sec : spmod ] .
the luminosity reached @xmath34 erg s@xmath5 in observation x3 , supporting ulx2 as a low - luminosity ulx .
this luminosity is a factor of @xmath1100 of that of observation x1 ( we did not use observations r2 and r3 for comparison because their luminosities have larger uncertainties ) . considering that observations r2 and r3 are close in time and that x1 and x2 are close in time too ,
ulx2 was observed in essentially six epochs and detected in four ( or in five if the detections in x1 and x2 are really from ulx2 ) of them over nearly two decades .
thus the source has a large duty cycle ( @xmath170% or more ) and is probably a highly variable persistent source or a recurrent transient with probably at least three outbursts in the last two decades ( see figure [ fig : srcltlc ] ) .
considering that observations x3 and c2 are about one year apart , the outburst should last more than one year if they are in the same outburst .
the left panels of figure [ fig : indlcsp ] show the light curves of the four observations when ulx2 was significantly detected , in order of decreasing luminosity from the top to the bottom panels .
the brightest observation x3 seems to experience dipping behavior sporadically ( 15 ks and 25 ks into the observation ) , with duration a few hundred seconds . to understand the spectral properties in the dipping periods
, we extracted the spectrum in the dipping period when the pn 1 - 10 kev count rate is @xmath290.06 counts s@xmath5 ( refer to figure [ fig : indlcsp ] ) .
the spectrum turns out to be soft .
when we fitted it with a multi - color disk ( mcd , diskbb in xspec ) , we obtained the inner disk temperature @xmath35 kev ( the column density is @xmath36 @xmath27 , and the reduced @xmath24 is relatively high ( 2.0 ) for 6 degrees of freedom ( d.o.f . ) ) . in comparison ,
the total spectrum gives @xmath37 kev ( section [ sec : spmod ] ) .
the hardness ratio , defined as the pn count rate in 110 kev divided by that in 0.2 - 1 kev , is [email protected] in the above dipping period , significantly lower than that ( [email protected] ) from the whole observation .
we also calculated the ratio of the pn count rate in the dipping period to that from the whole observation and obtained @xmath38 using 0.21 kev events and @xmath39 using 12 kev events .
thus the dips are accompanied by spectral softening .
because of this , we created the light curve using a high - energy band ( 110 kev ) for this bright observation to show the dips more clearly .
dips might also be present in the two fainter observations r1 and c2 , but not clearly seen due to low counting statistics .
in contrast , the light curve of the fainter observation c1 might show three flares that could possibly indicate @xmath17-hr ( quasi-)periodic modulations over two cycles . using a 60 ks window to calculate the count rates at different times
, we found that the count rates in the three flares are about 3.3 , 5.0 and 6.8 times of those in the minima ( the count rates in the two minima are similar ) , corresponding to about @xmath40 , @xmath41 , and @xmath42 , respectively .
we also extracted the spectra in the bright and faint intervals , depending on whether the count rate is larger or smaller than 0.005 counts s@xmath5 .
when we fitted them with a pl , we obtained their photon indices to be @[email protected] and [email protected] , respectively ( we fixed @xmath25 at @xmath44 @xmath27 obtained from the fit to the total spectrum ( section [ sec : spmod ] ) because of the low counting statistics of the spectra )
. we also calculated the hardness ratio , defined as the count rate in 18 kev divided by that in 0.31 kev and obtained the values of [email protected] and [email protected] for the bright and faint intervals , respectively .
thus we see no clear spectral variability associated with the flares in observation c1 .
r|c|c|l|c|cc 0094360601(x1 ) & pl & 0.06 & @xmath43=2.5 & & [email protected] & [email protected] + 0094360701(x2 ) & pl & 0.06 & @xmath43=2.5 & & [email protected] & [email protected] + & pl & @xmath45 & @xmath43=@xmath46 , @xmath47=@xmath48 & @xmath49 & @xmath50 & @xmath51 + & mcd & @xmath52 & @xmath53=@xmath54 , @xmath55=@xmath56 & @xmath57 & @xmath58 & @xmath59 + & mcd+pl & @xmath60 & @xmath53=@xmath61 , @xmath55=@xmath62 , @xmath43=@xmath63 , @xmath47=@xmath64 & @xmath65 & @xmath66 & @xmath67 + & mcd+nthcomp & @xmath68 & & @xmath69 & @xmath59 & @xmath70 + & mcd+comptt & @xmath52 & & @xmath69 & @xmath59 & @xmath71 + & pl & @xmath72 & @xmath43=@xmath73 , @xmath47=@xmath74 & @xmath75 & @xmath76 & @xmath77 + & mcd & @xmath78 & @xmath53=@xmath79 , @xmath55=@xmath80 & @xmath81 & @xmath82 & @xmath82 + & mcd & @xmath83 & @xmath53=@xmath84 , @xmath55=@xmath85 & @xmath86 & @xmath87 & @xmath88 + & pl & @xmath89 & @xmath43=@xmath90 , @xmath47=@xmath91 & @xmath92 & @xmath93 & @xmath94 + & mcd & @xmath52 & @xmath53=@xmath95 , @xmath55=@xmath96 & @xmath97 & @xmath98 & @xmath99 + & pl & @xmath100 & @xmath43=@xmath63 , @xmath47=@xmath101 & @xmath102 & @xmath103 & @xmath104 + rh600678n00(r2 ) & pl & 0.06 & @xmath43=2.5 & & [email protected] & [email protected] + rh600769n00(r3 ) & pl & 0.06 & @xmath43=2.5 & & [email protected] & [email protected] the brightest observation x3 has the best quality and allows for relatively detailed modeling .
following @xcite , we fitted the 210 kev spectrum with a pl and a broken pl ( both unabsorbed ) and found that the latter fit ( the total @xmath24 is 98.2 for 118 d.o.f . ) showed a significant ( @xmath105 ) improvement over the former one ( the total @xmath24 is 125.2 for 120 d.o.f . ) , indicating the presence of spectral curvature at several kev ( the inferred break energy is @xmath106 kev ) .
the fits over the whole spectrum ( 0.210 kev ) with a pl is very poor , with a reduced @xmath24 of 1.31 ( 560 d.o.f . ,
table [ tbl : spfit ] ) .
the fit with the mcd model is better ( the reduced @xmath24 is 1.15 ) , but there are still clear residuals at high energies , as is often seen in the spectral fits to ulxs and was explained as due to the advection effect and/or the compton scattering of disk photons in an optically thick wind , photosphere or corona in many studies , such as @xcite and @xcite . similar to these studies , we tested two - component models to include a comptonization component , i.e. mcd+pl and mcd+comptt ( comptt @xcite is available in xspec ) .
the fit with the mcd+pl model is much better than that with the mcd model , decreasing the total @xmath24 by 114 . the mcd+comptt model , with the seed photons tied to the inner disk temperature , decreases the total @xmath24 further , by 32 compared with the mcd+pl model for one more d.o.f .
the mcd+comptt model inferred a population of cool ( 1.0 kev ) comptonizing electrons with a high optical depth ( @xmath107 ) .
for the disk , we inferred @xmath108 kev ( 90% error ) or @xmath109 kev ( @xmath110 error ) ( here we report two error bars because we found that @xmath24 varied slowly for a relatively large range of @xmath53 ) .
we also tested the comptonization model nthcomp ( in xspec , * ? ? ? * ; * ? ? ?
* ; * ? ? ?
* ) , whose seed photons can be assumed to have the mcd shape .
this model was used in the fits to very bright ( around the eddington limit ) spectra from the @xmath111 state of the bhb by @xcite and those from the luminous neutron - star ( ns ) low - mass x - ray binary ( lmxb ) by @xcite as a substitution of the simpl model @xcite to account for the high - energy cutoff in the spectra .
the fitting results with mcd+nthcomp model are given in table [ tbl : spfit ] and the unfolded spectrum is shown in figure [ fig : indlcsp ] .
the mcd normalization is not well constrained , with the 90%-confidence lower limit reaching zero , which was also seen in @xcite and @xcite .
following @xcite , we roughly estimated the pre - comptonization disk normalization @xmath112=@xmath113 , where @xmath114 is the inner disk radius , @xmath115 the source distance , and @xmath116 the disk inclination , by adding the photons compton scattered ( the nthcomp component ) and those unscattered ( the mcd component ) , assuming that the photons are conserved .
we obtained @xmath117=@xmath118 .
the optical depth of the corona , which is not an explicit parameter of the model but can be derived using equation a1 in @xcite , is @xmath122 , which is high , similar to that inferred from the mcd+comptt model .
therefore we have made the simple assumption that while the optically thick corona can have seed photons from the disk it does not strongly affect the thermal disk emission in turn in both the mcd+comptt and mcd+nthcomp models .
the other spectra have much lower quality , and we only test the single - component models , i.e. , a mcd and a pl .
the unfolded spectra are plotted in figure [ fig : indlcsp ] .
rosat_/pspcb observation r1 can be fitted with either model .
we caution that the fit for this observation is limited by the narrow energy band coverage ( 0.22.4 kev ) and calibration uncertainty .
observation c2 can also be fitted with either model , but the pl model requires a relatively strong absorption ( @xmath119=0.24@xmath120 @xmath27 ) and a steep pl ( @xmath14=2.9 ) .
observation c1 , the one with ( quasi-)periodic flares ( figure [ fig : indlcsp ] ) , can be fitted with a pl , but not well with the mcd model ( the reduced @xmath24 is 1.9 ) .
the inferred photon index is @xmath121
. we did not carry out spectral fits to the other four observations ( r2 , r3 , x1 , and x2 ) when the source was not significantly detected and subject to strong contamination from bright sources near the nucleus .
instead , we estimated their source luminosities assuming the best - fitting pl model from observation c1 , since c1 was the closest in flux to these observations ( table [ tbl : spfit ] ) .
we note that for observations r2 and r3 , in which the source was not detected , we calculated the 90% confidence intervals using bayesian statistics @xcite .
the candidate optical counterpart to ulx2 fortunately appears on the outskirts of a star - forming region devoid of bright sources . `
hstphot ` indicates this counterpart as point - like in all filters .
its vega magnitudes are given in table [ tbl : hst ] .
the counterpart seems relatively red .
it was the most significantly detected in the f555w and f814w filters . further considering that these two filters are less subject to possible emission from accretion activity than the other filters
, we used the synphot package to compare the color of these two filters with the stellar spectra in @xcite , assuming the galactic extinction of @xmath122(b@xmath123v)=0.02 @xcite .
we found that the counterpart has the f555w@xmath123f814w color ( 1.21 mag ) the closest to those of g8i ( 1.09 mag ) and k3v ( 1.16 mag ) stars among supergiants and dwarfs in @xcite , respectively .
we fitted the f450w , f555w , and f814w fluxes of the counterpart with these two stellar spectra by minimizing the total @xmath24 , with the normalization of the spectra as a free parameter .
the corresponding apparent magnitudes of the best - fitting spectra are listed in table [ tbl : hst ] .
the fit residuals are @xmath1240.1 mag in these filters . in the @xmath125 filter f656n ,
the deviation is slightly larger , with our source brighter than the fits by 0.30.4 mag ( 23@xmath16 ) . in the uv filter f336w ,
the excess is much larger , by 1.5 and 0.7 mag compared with the fits with a g8i star and a k3v star , respectively , though in this filter the source was detected only at 6@xmath16 .
the apparent visual magnitude of our source is @xmath126=22.22 , corresponding to an absolute magnitude of @xmath127 at the distance of m94 , which is close to that expected for a g8i star or that for a red globular cluster containing @xmath1@xmath128 k3v stars . in the above
, we have neglected the possible effect of binary evolution on the color of the donor star .
ulx2 was fortunately captured at different flux levels , showing clear spectral evolution . with similar spectral models tested
, we can easily compare ulx2 with the second ulx in m31 ( ) studied by @xcite .
the brightest observation x3 of ulx2 is very similar to the brightest _ xmm - newton _
observations xmm3 and xmm4 of .
all these observations have @xmath129 erg s@xmath5 and are better fitted by the mcd+comptt model than by a pl , a mcd , or their combination .
the fits with the mcd+comptt model to these observations all inferred a relatively cool ( @xmath291 kev , which was not well constrained ) disk and comptonization in a cool ( @xmath11 kev ) optically thick ( @xmath130 ) corona .
is probably a stellar - mass bhb with @xmath9 , based on the joint radio / x - ray behavior and the observation of a disk - dominated state at a very low luminosity , supporting that the above spectra are probably characteristic of accretion at the eddington limit @xcite .
then ulx2 could be a stellar - mass bhb as well , with an accretion rate around the eddington limit in observation x3 .
many low - luminosity ulxs show similar spectra and could also be explained as supercritically accreting stellar - mass bhbs @xcite .
if ulx2 is really a stellar - mass bhb , we would expect it to behave similar to typical galactic bhbs when it is at well sub - eddington luminosities . in the fainter observation c2 ( @xmath131 erg s@xmath5 ) ,
the spectrum is softer than observation x3 and can be described with a standard thermal accretion disk , thus consistent with being in the thermal state of galactic bhbs .
the thermal state in galactic bhbs tends to occur above @xmath132 of the eddington limit @xcite . assuming observation c2 to be in this state and using its 0.210 kev unabsorbed luminosity from the mcd model , we constrained the mass of the bh in ulx2 to be @xmath124100 , supporting the identification as a stellar - mass bhb .
we note that for this observation we can not rule out a steep pl model with relatively strong absorption from the fit . in the even fainter observation c1 ( @xmath133 erg s@xmath5 )
, the spectrum becomes harder again and can be described with a pl with @xmath134 .
such a photon index is often seen in the steep - pl state of galactic bhbs , and the source might be in such a state in this observation , but considering its large uncertainty , we can not rule out the source being in the hard state instead . in any case , we might have observed a state transition often seen in bhbs .
we clearly need higher quality data for more detailed comparison .
therefore transient / highly variable ulxs like ulx2 and serve a link between persistent low - luminosity ulxs and classical stellar - mass bhbs .
this is reminiscent of the first known transient z source linking z and atoll sources , the two main classes of weakly magnetized ns lmxbs @xcite .
here we briefly discuss the ns lmxb case to gain some insights into the physics involved in the bhb case .
z sources reach the eddington limit or above and are mostly persistent with variability factors of only a few , while atolls have luminosities typically less than 50% of the eddington limit and tend to be transient or highly variable .
the proof that z sources accrete at near or super - eddington limit and is the same class of object as atolls but at different accretion rates based on is very convincing in several ways @xcite .
first , z and atoll sources have well - known distinct timing and spectral properties , and showed z - source properties at high luminosities and then atoll - source ones during the decay in its 2006 - 2007 outburst .
secondly , ns lmxbs have a unique way to infer the eddington luminosity , i.e. radius expansion type-1 x - ray bursts , which were detected in , and the luminosities when behaved as a z source were indeed near or above the eddington luminosity inferred from the bursts @xcite . finally , the spectral fitting by @xcite also gathered some evidence that the eddington limit was reached in the z - source stage , especially the result that the disk in the z - source lower vertex had a relatively constant temperature and an increasing inner disk radius with increasing luminosity , deviating from the @xmath135 trend observed in the atoll - source stage .
the increase in the inner disk radius with luminosity is expected because as the mass accretion rate into the disk increases , more and more inner part of the disk reaches the local eddington limit , leading to increasing radial advection flow and/or mass outflow @xcite
. therefore studies of ns lmxbs not only demonstrate that near or super - eddington accretion is possible but also assure us that objects with different outbursting behaviors and luminosities can belong to the same class and have overall spectral / timing properties mainly determined by the accretion rate .
all this supports the possibility that most low - luminosity ulxs are in fact near or super - eddington accreting stellar - mass bhbs .
galactic bhbs are known to show many similarities to atolls , such as sub - eddington luminosity , transient behavior , broad - band noise , and possibly also disk spectral evolution @xcite , though there are also some differences , such as hotter thermal spectra and stronger millisecond variability in atolls , which could be reasonably ascribed to emission from the impact of materials onto the ns surface .
then there is a question of whether ulxs that are supercritically accreting stellar - mass bhbs also show some properties observed in z sources .
one possible interesting similarity is their generally small long - term variability .
classical z sources ( , , , , , and ) , all with a low - mass companion , are persistent with long - term variation factors of only a few ( only two z sources ( and ) , discovered recently , are transients ) .
most known ulxs are also persistent .
@xcite found only 15 with a long - term variation factor @xmath13610 in their 100 ulxs , which is in contrast with the general transient behavior of galactic bhbs , especially those with a low - mass companion .
this could be because most of these ulxs might have a high - mass companion , but from z sources we can not rule out that the long - time activity might be related to high accretion rates .
it would also be worthwhile to search for similarity of the disk behavior between ulxs and z sources .
@xcite had reported similar disk evolution in the bhb in a very bright state , i.e. , the @xmath111 state , to that seen in in the z - source stage .
a lot of work is still needed in the future to establish the possible connection between z sources and low - luminosity ulxs .
the observations of dips from ulx2 make it one of the dipping sources , which include about 20 galactic x - ray binaries and a few ulxs , such as the ulx in ngc 55 @xcite and ngc 5408 x-1 @xcite .
long dips lasting for 10%30% of the orbital phase are mostly found in ns lmxbs and can be explained as due to absorption by a bulge on the edge of the accretion disk at the point where the gas stream impacts the disk ( e.g. , * ? ? ?
short dips less than a few hundred seconds ( thus typically @xmath291% of the orbital period ) are commonly observed in two bhbs ( with a low - mass companion ) and ( with a supergiant companion ) .
they show orbital phase dependence and could be due to absorption in accretion streams from the companion @xcite .
it is not clear whether the dipping ulxs in ngc 55 and ngc 5408 also show orbital phase dependence , but they might have similar origin .
dips in ulx2 are short and are thus probably also due to absorption in accretion streams .
the spectrum becomes soft in the dips in observation x3 , which could be because a hot component is obscured and/or because the absorbing matter is partially ionized . in terms of the mcd+nthcomp model that we used to fit observation x3 ,
the dipping spectrum that we created from this observation could be very roughly accounted for with either complete obscuration of the nthcomp component ( the reduced @xmath24 is 3.7 for 7 d.o.f . ) or with the presence of absorbing matter with a column density of @xmath137 @xmath27 and ionization parameter of @xmath138 ( the reduced @xmath24 is 4.3 for 7 d.o.f . , using the ionized absorption model zxipcf in xspec ) .
we do not have enough dips to search for periodicity , which could otherwise provide information of the orbital period . in any case , the dips often imply that the system is at a high inclination .
detection of ulxs at high inclinations is important , because it poses a problem to use the beaming effect to explain the ultraluminous nature of these sources .
the possible large ( quasi-)periodic x - ray modulations / flares in observation c1 is also a special property of ulx2 .
similar modulations were also observed in some ulxs such as in circinus ( @xmath17.5
hr , * ? ? ?
* ) , ( @xmath12 hr , * ? ? ?
* ) , and in m74 ( @xmath12 hr , * ? ? ?
* ; * ? ? ?
the main difference is that ulx2 has @xmath139 erg s@xmath5 while those sources have @xmath140 erg s@xmath5 when such large modulations were observed .
various explanations for such modulations were discussed by those studies , including the binary eclipse , modulations of the accretion rate due to some instability , and variability at the base of the jet . because ulx2 might have a high inclination considering the detection of dips
, if the flares are periodic , the binary eclipsing explanation seems plausible , with the @xmath17-hr period being the orbital period .
the main problem for this explanation is that we found no spectral change accompanying the flux variation .
moreover , such modulations seem to be flux and/or spectral state dependent because they were not seen in the brighter / softer observations . therefore the disk or jet instability explanations are favored .
another interesting aspect of ulx2 is our identification of its point - like red optical counterpart candidate .
some ulxs have been reported to have optical counterpart candidates ( * ? ? ?
* ; * ? ? ?
* e.g. , ) .
most of them appear blue and are probably contaminated by accretion activity , while some have a clear red component like our source , such as @xcite and @xcite . for , there is also an extra blue component which was shown by @xcite to be probably from disk emission , while the red component could be an asymptotic giant branch star .
there is some possible uv excess in ulx2 , and the most likely explanation is the disk emission too .
we note that our source is highly variable while the _ hst _ observations are not simultaneous with each other or with x - ray observations .
therefore , one explanation for why the counterpart to ulx2 appears red while those to other ulxs are mostly blue is that the _ hst _ observations of ulx2 were made when it was not x - ray ultraluminous , though the alternative explanation that ulx2 resides in an old cluster is also possible .
we have shown many intriguing properties of ulx2 . it is either a highly variable persistent source or a recurrent transient , with a long - term variation factor of at least @xmath1100 and a duty cycle of at least @xmath170% . in the brightest observation x3 , the source exhibited as a member of low - luminosity ulxs , with @xmath141 erg s@xmath5 and had a spectrum showing a high - energy cutoff but too hard to be described by thermal emission from a standard accretion disk .
short dips , accompanied by spectral softening , are also clearly seen . in the fainter observation c2 ( @xmath142 erg s@xmath5 ) ,
the spectrum became softer and is consistent with thermal emission from a standard accretion disk with @xmath143 kev , supporting it as in the thermal state as seen in galactic bhbs . in the even fainter observation c1 ( @xmath144 erg s@xmath5 ) , the spectrum becomes harder again , with @xmath145 , and could be in the steep - pl state or the hard state of galactic bhbs . in this observation
, we also observe possible flares that might indicate @xmath17-hr periodic x - ray modulations .
we also identify a possible point - like optical counterpart , whose colors and luminosity resemble a g8 supergiant or a compact red globular cluster containing @xmath8 k dwarfs except for some possible uv excess . combining all the above properties , we suggest that ulx2 is a stellar - mass bhb with a supergiant companion or with a dwarf companion residing in a globular cluster .
the high variability and large duty cycle of ulx2 makes it easy for multiwavelength follow - up , which is important for confirmation of some properties that we report here , such as the state transition and possible periodic behavior at low luminosities , and reveal more properties to further constrain its nature . , m. w. , pivovaroff , m. , baganoff , f. , et al .
1998 , in society of photo - optical instrumentation engineers ( spie ) conference series , vol . 3444 , society of photo - optical instrumentation engineers ( spie ) conference series , ed .
r. b. hoover & a. b. walker , 210224 | we report the discovery of a new ultraluminous x - ray source ( ulx ) within the spiral galaxy m94 .
the source has been observed by _ rosat _ , _ chandra _ , and _ xmm - newton _ on several occasions , exhibiting as a highly variable persistent source or a recurrent transient with a flux variation factor of @xmath0100 , a high duty cycle ( at least @xmath170% ) , and a peak luminosity of @xmath2@xmath12@xmath310@xmath4 erg s@xmath5 ( 0.210 kev , absorbed ) . in the brightest observation ,
the source is similar to typical low - luminosity ulxs , with the spectrum showing a high - energy cutoff but harder than that from a standard accretion disk .
there are also sporadical short dips , accompanied by spectral softening . in a fainter observation with @[email protected]@xmath310@xmath6 erg s@xmath5
, the source appears softer and is probably in the thermal state seen in galactic black - hole x - ray binaries ( bhbs ) . in an even fainter observation ( @xmath2@xmath19@xmath310@xmath7 erg s@xmath5 ) ,
the spectrum is harder again , and the source might be in the steep - powerlaw state or the hard state of bhbs
. in this observation , the light curve might exhibit @xmath17 hr ( quasi-)periodic large modulations over two cycles .
the source also has a possible point - like optical counterpart from _ hst _ images . in terms of the colors and the luminosity ,
the counterpart is probably a g8 supergiant or a compact red globular cluster containing @xmath8 k dwarfs , with some possible weak uv excess that might be ascribed to accretion activity .
thus our source is a candidate stellar - mass bhb with a supergiant companion or with a dwarf companion residing in a globular cluster .
our study supports that some low - luminosity ulxs are supercritically accreting stellar - mass bhbs . |
polarization in @xmath10 decays should be sensitive to the @xmath11 structure of the standard model due to the power suppression associated with the helicity - flip of a collinear quark .
for example , in the standard model the factorizable graphs for @xmath12 are due to transition operators with chirality structures @xmath13 , see figure .
1 . there are three helicity amplitudes , @xmath14 , @xmath15 , and @xmath16 , in which both vectors are longitudinally , negatively , and positively polarized , respectively . in @xmath17 a collinear @xmath18 or @xmath19 quark with positive helicity ends up in the negatively polarized @xmath20 , whereas in @xmath16 a second quark
helicity - flip is required in the form factor transition . in the case of new right - handed currents , e.g. ,
@xmath21 , the hierarchy is inverted , with @xmath16 and @xmath15 requiring one and two helicity - flips , respectively .
helicity - flip suppression can be estimated by recalling that the probability for a positive helicity free fermion to have negative spin along some axis is given by @xmath22 , where @xmath23 is the angle between the axis and the momentum vector . for a @xmath20 meson in a symmetric configuration
the transverse momentum of the valence quarks is @xmath24 , implying that the helicity suppression in @xmath25 is @xmath26 .
the form factor helicity suppression in @xmath27 should be approximately @xmath28 , where @xmath29 is the transverse momentum of the outgoing @xmath18 quark .
the latter can be estimated by identifying it with the transverse momentum of the @xmath30 quark . in the fermi momentum model of @xcite @xmath31 . using the equivalence of this model to a particular hqet based shape function ansatz @xcite and for illustration taking @xmath32 mev and @xmath33 gev@xmath34 yields @xmath35 mev , or a helicity suppression of @xmath36 .
these simple estimates should be compared to naive factorization , supplemented by the large energy form factor relations @xcite ( also see @xcite ) . for @xmath37 , @xmath38 the coefficient @xmath39 , where the @xmath40 are the usual naive factorization coefficients , see e.g. @xcite , and @xmath41 .
the large energy relations imply @xmath42 we use the sign convention @xmath43 . @xmath44 and @xmath45 are the @xmath46 form factors in the large energy limit @xcite . both scale as @xmath47 in the heavy quark limit , implying that helicity suppression in @xmath48 is @xmath49 which is consistent with our estimate ( the form factor transition contributes @xmath50 in @xmath51 ) .
@xmath52 parametrizes the form factor helicity suppression .
it is given by @xmath53 where @xmath54 and @xmath55 are the axial - vector and vector current form factors , respectively .
the large energy relations imply that it vanishes at leading power , because helicity suppression is @xmath56 .
light - cone qcd sum rules @xcite , and lattice form factor determinations scaled to low @xmath57 using the sum rule approach @xcite , give @xmath58 ; qcd sum rules give @xmath59 @xcite ; and the bsw model gives @xmath60 @xcite .
these results are consistent with our simple estimate for form factor helicity suppression .
-0.2 cm the large energy relations giving rise to ( [ av1v2scet ] ) are strictly valid for the _ soft _ parts of the form factors , at leading power and at leading order in @xmath61 . however , the soft form factors are not significantly sudakov suppressed in the soft collinear effective theory ( scet ) @xcite .
the results of @xcite thus imply that the @xmath62 form factor contributions , particularly the symmetry breaking corrections to the large energy relations , can be neglected .
in fact , @xmath52 does not receive any perturbative corrections _ at leading power _
@xcite ; again , this is because form factor helicity suppression is @xmath56 .
furthermore , power corrections to all of the form factor relations begin at @xmath63 ( rather than @xmath64 ) in scet @xcite .
therefore , the above discussion of helicity suppression in naive factorization will not be significantly modified by perturbative and power corrections to the form factors . in the transversity basis
@xcite the transvese amplitudes are @xmath65 ( @xmath66 ) for @xmath67 ( @xmath0 ) decays .
the polarization fractions satisfy @xmath68 in naive factorization , where the subscript @xmath69 refers to longitudinal polarization , @xmath70 , and @xmath71 .
the measured longitudinal fractions for @xmath72 are close to 1 @xcite .
this is clearly not the case for @xmath73 , for which full angular analyses yield @xmath74 naively averaging the belle and babar measurements ( without taking large correlations into account ) yields @xmath75 . in the charged mode
, babar has measured @xmath76 @xcite .
we must go beyond naive factorization in order to determine if the small values of @xmath77 could be due to the dominance of qcd penguin operators in @xmath3 decays .
in qcd factorization @xcite exclusive two - body decay amplitudes are given in terms of convolutions of hard scattering kernels with meson light - cone distribution amplitudes . at leading power
this leads to factorization of short and long - distance physics .
this factorization breaks down at sub - leading powers with the appearance of logarithmic infrared divergences .
nevertheless , the power - counting for all amplitudes can be obtained .
the extent to which it holds numerically can be checked by introducing an infrared hadronic scale cutoff , and assigning large uncertainties .
non - perturbative quantities are thus roughly estimated via single gluon exchange . in general
, large uncertainties should be expected for polarization predictions , given that the transverse amplitudes begin at @xmath56 .
however , we will find that this is not the case for certain polarization observables , particularly after experimental constraints , e.g. , total rate or total transverse rate , are imposed .
our results differ substantially from previous studies of @xmath10 in qcd factorization @xcite .
of particular note is the inclusion of annihilation topologies .
the complete expressions for the helicity amplitudes are lengthy and will be given in @xcite .
expressions for a few contributions are included below . in qcd factorization ,
the standard model effective hamiltonian matrix elements can be written as @xcite @xmath78 where @xmath79 labels the vector meson helicity , and @xmath80 for @xmath81 .
@xmath82 gives rise to annihilation topoplogy amplitudes , to be discussed shortly , and @xmath83 where @xmath84 , and @xmath85 or @xmath86 .
the coefficients @xmath87 contain contributions from naive factorization , vertex corrections , penguin contractions , and hard spectator interactions .
the transition operators @xmath88 are @xmath89 ( i=1 ) ; @xmath90 ( i=2 ) ; @xmath91 ( i=3,5)]@xmath92 ( i=9,7 ) ; @xmath93 ( i=4)]@xmath92 ( i=10 ) ; @xmath94 ( i=6)]@xmath92 ( i=8 ) , where @xmath95 is summed over @xmath96 . for i=1,2 ; 3 - 6 ; and 7 - 10 they originate from the current - current @xmath97 ; qcd penguin @xmath98 ; and electroweak penguin operators @xmath99 , respectively . for @xmath100 ,
@xmath101 is defined at leading - order as @xmath102 the @xmath103 coefficients contain factors of @xmath104 , @xmath105 , arising from the vector meson flavor structures .
@xmath106 ( @xmath107 ) is the
emission ( form factor ) vector meson , see figure 1 .
the @xmath108 matrix elements vanish at tree - level , i.e. , at leading order in @xmath61 , as local scalar current vacuum - to - vector matrix elements vanish . due to the underlying flavor structure ,
the effects of @xmath109-@xmath110 are describable in terms of a reduced set of coefficients @xcite @xmath111 where @xmath112 ) , and @xmath113 are next - to - leading order matrix elements in @xmath61 , in which @xmath114 again forms the emission particle @xmath115 .
the arguments @xmath116 are understood throughout . at
next - to - leading order , the coefficients @xmath117 can be written as @xcite @xmath118 + p_i^{h , p } ( v_2)\,,\ ] ] where the upper ( lower ) signs apply when @xmath119 is odd ( even ) .
the superscript
p appears for @xmath120 .
the @xmath121 are tree - level naive factorization coefficients ( figure 1 ) ; at next - to - leading order the @xmath122 account for one - loop vertex corrections , the @xmath123 for penguin contractions ( figure 2 ) , and the @xmath124 for hard spectator interactions ( figure 3 ) .
they are given in terms of convolutions of hard scattering kernels with vector meson and @xmath0 meson light - cone distribution amplitudes .
for each @xmath119 , the corresponding graphs have the same quark helicity structure .
two twist-2 light - cone distribution amplitudes @xmath125 and @xmath126 , and four two - particle twist-3 distributions ( and their derivatives ) enter the longitudinal and tranverse vector meson projections @xcite .
the argument @xmath127 ( @xmath128 ) is the quark ( antiquark ) light - cone momentum fraction .
the two - particle twist-3 distributions can be expressed in terms of @xmath129 via wandura - wilzcek type equations of motion @xcite , if higher fock states are ignored .
the twist-3 vector meson projections then depend on the three distributions , @xmath130 @xmath131 and @xmath132 project onto transversely polarized vectors in which the quark and antiquark flips helicity , respectively .
@xmath133 , defined in @xcite , projects onto longitudinally polarized vectors in which either the quark or antiquark flips helicity .
light quark mass effects are ignored , and a discussion of twist-4 distribution amplitudes and higher fock state effects is deferred @xcite .
the leading - twist distribution amplitudes are given in terms of an expansion in gegenbauer polynomials @xcite , @xmath134 , \qquad i=\perp , \parallel\,.\ ] ] our numerical results include the first two moments @xmath135 , @xmath136 .
the asymptotic forms of the twist-3 distribution amplitudes are @xmath137 , @xmath138 , and @xmath139 for @xmath140 , @xmath141 , and @xmath142 , respectively .
the @xmath0 light - cone distribution amplitude @xmath143 @xcite enters the hard spectator interactions through its inverse moment , @xmath144 . -.35 cm -0.2 cm the naive - factorization coefficients are @xmath145 ( i=6,8 ) , @xmath146 ( i@xmath1476,8 ) .
the longitudinal ( transverse ) helicity amplitudes arise at twist-2 ( twist-3 ) in naive facotrization .
the vertex corrections are negligible compared to the theoretical uncertainties .
note that @xmath148 for all @xmath149 . at @xmath150
the longitudinal penguin contractions are , respectively , @xmath151 at twist-2 and @xmath152 at twist-3 .
the quantities @xmath153 are the @xmath154 counterparts defined in @xcite , and @xmath155 is the scale - dependent tensor - current decay constant
. the transverse penguin contractions are @xmath156 to twist-4 , and at twist-3 , @xmath157\ !
+ c_3 [ 2 l - g^\pm_{v_2}(0 ) \!-\ !
g^\pm_{v_2}(1 ) ] \nonumber\\ & + & ( c_4\!+\!c_6 ) [ \tilde{l}\
! - \!3\,g^\pm_{v_2}(0)\ ! -\ !
g^\pm_{v_2}(s_c)\ ! -\ !
g^\pm_{v_2}(1 ) ] \,),\nonumber \\ p_{10}^{\pm,\ , p}(v_2)&\!=\ ! & ( { \alpha}/9\pi)\ , ( c_1 /n_c + c_2 ) [ l- g^\pm_{v_2}(s_p ) ] , \end{aligned}\ ] ] where @xmath158 , @xmath159 , @xmath160 ( @xmath161 , @xmath162 ) , @xmath163 and @xmath164 is the well known penguin function , see e.g. , @xcite .
the penguin contractions account for approximately 30% and 20% of the magnitudes of @xmath165 and @xmath166 ( for default input parameters ) , respectively , before including the hard spectator interactions .
the dipole operators @xmath167 , @xmath168 do not contribute to the _ transverse _ penguin amplitudes at @xmath169 due to angular momentum conservation : the dipole tensor current couples to a transverse gluon , but a ` helicity - flip ' for @xmath95 or @xmath170 in figure 2 would require a longitudinal gluon coupling .
formally , this result follows from the wandura - wilczek relations and the large energy relations between the tensor - current and vector - current form factors @xcite .
for example , the integrand of the convolution integral for @xmath171 vanishes identically , whereas @xmath172 . note that transverse amplitudes in which a vector meson contains a collinear higher fock state gluon also vanish at @xmath169 .
this can be seen from the vanishing of the corresponding partonic dipole operator graphs in the same momentum configurations .
transverse @xmath173 spectator interaction contributions are highly suppressed and are studied in @xcite .
the hard spectator interaction quantities @xmath174 contain logarithmically divergent integrals beyond twist-2 , corresponding to the soft spectator limit in @xmath175 , see e.g. , figure 3 .
we integrate the quark light - cone momentum fraction in @xmath175 over the range @xmath176 $ ] , and replace the divergent quantities @xmath177 with complex parameters @xmath178 . as in @xcite , these are modeled as @xmath179 , with @xmath180 and @xmath181 gev .
this reflects the physical @xmath182 infrared cutoff , and allows for large strong phases from soft rescattering . for @xmath183
: @xmath184 first arises at twist-2 ; @xmath185 arises via a twist-3@xmath186 twist-2@xmath187 projection ; and @xmath188 arises via a ( twist-3)@xmath34 projection . for @xmath189 : @xmath190 to twist-3 ; @xmath191 arises via a twist-3@xmath192 twist-2@xmath193 projection and is infrared finite ; and @xmath194 arises via a twist-4@xmath195 twist-2@xmath187 projection .
the basic building blocks for annihilation are matrix elements of the operators @xmath196 , @xmath197 , and @xmath198 , denoted @xmath199 , @xmath200 , and @xmath201 , respectively , see e.g. , figure 3 .
the first quark bilinear corresponds to the @xmath67 meson , the superscript @xmath202 indicates a gluon attached to the initial ( final ) state quarks in the weak vertex , and by convention @xmath115 ( @xmath175 ) contains a quark ( antiquark ) from the weak vertex . @xmath203 and
@xmath204 dominate the @xmath3 qcd penguin annihilation amplitudes .
the latter are expressed as @xmath205 , where @xmath206\,.\ ] ] the arguments @xmath116 have been suppressed .
the @xmath103 coefficients are again determined by the vector meson flavor structures . for the electroweak penguin annihilation amplitude @xmath207 ,
substitute @xmath208 , respectively . @xmath203 and
@xmath204 arise at twist-3 .
@xmath209 is given by @xmath210 for @xmath211 substitute @xmath212 , @xmath213 , @xmath214 , and change the sign of the second term .
the integrals over @xmath215 and @xmath216 are logarithmically divergent , corresponding to the soft gluon limit @xmath217 . for simplicity ,
the asymptotic distribution amplitudes are used , as in @xcite ; non - asymptotic @xmath7 violating effects will be discussed shortly .
the logarithmic divergences are again replaced with complex parameters , @xmath218 , yielding @xmath219 for @xmath211 substitute @xmath212 and @xmath220 .
the contributions of @xmath221 and @xmath222 to the helicity amplitudes are formally of @xmath223 ( @xmath224 scales like @xmath225 ) .
however , note that as @xmath226 and @xmath227 are varied from 0 to 1 , @xmath209 and @xmath211 increase by more than an _ order of magnitude_. a summary of power counting at next - to - leading order is given in @xcite .
as expected , each quark helicity - flip costs @xmath228 in association with either one unit of twist , or form factor suppression .
a @xmath229 change in vector meson helicity due to a collinear gluon in a higher fock state also costs one unit of twist , or @xmath228 .
in addition , annihilaton graphs receive an overall @xmath228 suppression .
an apparent exception is provided by the ( twist-3)@xmath34 contributions to @xmath230 ; they contain a linear infrared divergence which would break the power counting .
( @xmath230 would be promoted to @xmath231 but would remain numerically small , as can been by parametrizing the divergence as @xmath232 with , e.g. , @xmath233 ) .
however , the divergence should be canceled by twist-4@xmath234twist-2 effects , see below .
regardless , ( [ smpred ] ) remains formally true in qcd factorization .
the first relation in ( [ smpred ] ) has also been confirmed recently in scet @xcite .
we expect that the power counting obtained in qcd factorization will be reproduced for all corresponding graphs in scet .
amplitudes involving twist-4 vector meson projections remain to be explicitly evaluated @xcite .
twist-4@xmath235twist-2 projections give rise to @xmath236 .
however , these effects should be similar in magnitude to ( twist-3)@xmath34 contributions to the positive helicity hard spectator amplitudes , which were found to be small .
the twist-4@xmath234twist-2 contributions to @xmath237 must cancel the non - vanishing ( twist-3)@xmath34 contributions , since @xmath237 must vanish by equations of motion .
this condition leads to new wandura - wilczek type relations between the products of twist-4@xmath235twist-2 and ( twist-3)@xmath34 light - cone distribution amplitudes .
these relations should insure cancelation of the aforementioned linear divergence in @xmath238 by twist-4@xmath235twist-2 effects @xcite .
finally , twist-4@xmath234twist-3 projections give rise to @xmath239 and @xmath240 ; however , these amplitudes should be both formally and numerically suppressed by @xmath231 compared to @xmath241 and @xmath209 , respectively .
we have also not explicitly considered graphs in which higher twist two - body vector meson projections are replaced with higher fock - state projections of same twist containing collinear gluons , e.g. , @xmath242 .
the latter are expected to receive additional suppression at each twist , e.g. , 20% @xcite .
these corrections , especially the tree - level twist-3 contributions to the coefficients @xmath243 , should be included @xcite .
however , they will not alter our conclusions , given the large uncertainties that have already been assigned to the power corrections .
expressions for a few @xmath244 amplitudes are given below , @xmath245\ , , \nonumber\\ { \cal a}^h_{b^-\rightarrow \rho^{- } \bar{k}^{*0 } } & = & -i { g_f \over \sqrt{2 } } \sum_{p = u , c } \lambda_p^{(s ) } a_{\rho k^ * } ^h \left [ \alpha_4^{p,\,h } -\frac12 \alpha_{4,\,{\rm ew}}^{p,\,h } + \delta_{pu } \beta_2^h + \beta_3^h + \beta_{3,\,{\rm ew}}^h \right]\ , , \nonumber\\ { \cal a}^h_{\bar{b}^0\rightarrow \rho^{+ } \rho^{- } } & = & -i { g_f \over \sqrt{2 } } \sum_{p = u , c } \lambda_p^{(d ) } a_{\rho \rho } ^h \left [ \delta_{pu } a_1^h \!+\ !
\alpha_4^{p,\,h } \!+\ !
\alpha_{4,\,{\rm ew}}^{p,\,h}\ ! + \ !
\delta_{pu } \beta_1^h \!+\ ! \beta_3^h \!+\ !
2 \beta_4^h\ ! -\!\frac12 \beta_{3,\,{\rm ew}}^h \ !
+ \!\frac12 \beta_{4,\,{\rm ew}}^{p,\,h } \right ] , \nonumber\\ { \cal a}^h_{{b}^- \rightarrow \rho^- \rho^{0 } } & = & -i { g_f \over { 2 } } \sum_{p = u , c } \lambda_p^{(d ) } a_{\rho \rho } ^h \left [ \delta_{pu } ( a_1^h + a_2^h ) + \frac32 \alpha_{3,\,{\rm ew}}^{h } + \frac32 \alpha_{4,\,{\rm ew}}^{p,\,h } \right]\,,\end{aligned}\ ] ] where @xmath246 . for @xmath247 add the term @xmath248 to @xmath249 .
the arguments @xmath250 of @xmath251 , @xmath252 , @xmath253 have been suppressed , but it is understood that they are to be identified with the subscripts @xmath250 of the prefactors @xmath254 . the new annihilation coefficients @xmath255 , and amplitudes for other decays are given in @xcite .
( to first approximation , all annihilation coefficients except @xmath256 can be ignored in the above amplitudes . ) @xmath257 arise from current - current operator annihilation graphs .
@xmath258 arise from qcd annihilation graphs with different flavor topology than @xmath256 , and @xmath259 are the analogous electroweak annihilation coefficients . -.35
cm -0.2 cm
in our numerical analysis the inputs of @xcite are used , except for the form factors for which we take @xmath260 and @xmath261 .
the expanded ranges reflect the lower values recently obtained for @xmath262 from the lattice @xcite and from @xmath263 @xcite .
we also take @xmath264 , spanning existing model determinations @xcite@xcite . in the evaluation of the hard - scattering and annihilation graphs , @xmath61 and the wilson coefficients
are evaluated at an intermediated scale @xmath265 , with @xmath266 $ ] .
the renormalization scales are varied independently in the three classes of graphs : hard - scattering ( @xmath267 ) , annihilation ( @xmath268 ) , and those containing form factors .
the quantities @xmath269 and @xmath270 parametrizing the logarithmic divergences are varied independently for unrelated convolution integrals , with @xmath271 $ ] and @xmath272 $ ] .
the default values are @xmath273 .
the form factor dependence can be approximately factored out of the ( cp - averaged ) @xmath72 branching ratios , yielding @xmath274 ^ 2 \,.\label{rhorhoresults}\ ] ] the error bars include the uncertainties due to variation of all remaining inputs , added in quadrature .
the measured values are @xmath275 ( average of @xcite ) , and @xmath276 @xcite .
we also obtain @xmath277 and @xmath278 .
the first and second sets of error bars are due to form factor and remaining hadronic uncertainties , respectively , again added in quadrature .
the form factor errors are reduced substantially when the branching ratios are constrained to lie in their measured ranges , to approximately @xmath279 for @xmath280 and @xmath281 for @xmath282 @xcite .
the relatively small polarization uncertainties are due to the absence of ( in @xmath283 ) , or ckm suppression of ( in @xmath284 ) the qcd penguin amplitudes .
the measured polarizations are @xmath285 ( average of @xcite ) and @xmath286 @xcite .
thus , the predicted longitudinal polarizations are in good agreement with experiment , and with naive power counting . -.35 cm -0.2 cm averaging the belle and babar @xmath287 ( cp - averaged ) measurements @xcite yields @xmath288 and @xmath289 , or @xmath290 and @xmath291 , where @xmath292 is the total transverse branching ratio . without annihilation
we obtain @xmath293 and @xmath294 , where the second ( first ) set of error bars is due to variations of @xmath270 ( all other inputs ) . in figure 4 ,
@xmath295 and @xmath296 are plotted versus the annihilation parameters @xmath297 and @xmath298 entering @xmath299 and @xmath300 , respectively .
the black curves are obtained for central values of all inputs , with default values for all annihilation and hard spectator interaction parameters other than @xmath301 .
the blue bands are obtained by adding the uncertainties due to variations of the inputs in quadrature , keeping default annihilation and hard spectator interaction parameters .
the widths of the bands are dominated by the form factor uncertainties .
the yellow bands also include , in quadrature , the uncertainties due to variations of all @xmath302 and @xmath268 .
the thick curve gives the maximum values obtained for @xmath296 under simultaneous variation of all inputs .
the absolute branching ratios suffer from large theoretical uncertainties , as is usually the case .
nevertheless , it is clear that the contributions of @xmath299 and @xmath300 to the qcd annihilation amplitudes can be @xmath8 numerically even though they are formally @xmath231 .
this can be traced to the quadratic dependence on the divergences ( @xmath303 ) and the large coefficient @xmath304 in @xmath305 .
the quantities @xmath306 and @xmath307 , as well as the renormalization scales and form factors entering @xmath308 and @xmath309 are , a - priori , unrelated .
figure 4 therefore implies that the measurements of @xmath310 and @xmath296 can easily be accounted for simultaneously . according to figure 4
, @xmath311 can also be accounted for given that the @xmath312 and @xmath313 amplitudes only differ by a small current - current operator annihilation graph .
_ a test for right - handed currents _ in figure 5 ( left ) the predicted ranges for @xmath314 and @xmath296 are studied simultaneaously for @xmath315 in the standard model .
default curve is again obtained by varying @xmath316 in the range @xmath317 $ ] , keeping all other inputs at their default values , and the error bands are obtained by adding uncertainties in quadrature as in figure 4 .
evidently , the second relation in ( [ smpred ] ) holds at next - to - leading order , particularly at larger values of @xmath296 where qcd annihilation dominates @xmath318 and @xmath319 .
we also plot the maximum values attained for @xmath314 under simultaneous variation of all inputs .
the result is sensitive to @xmath52 , as it largely determines the relative signs and magnitudes of the form factor terms in @xmath320 and @xmath321 , see ( [ av1v2scet ] ) .
the thick black curve ( corresponding to @xmath322 in figure 4 ) and blue curve give maxima for @xmath323 , in accord with existing model determinations , and @xmath324 , respectively . a ratio in excess of the standard model range , e.g. , @xmath325
if @xmath326 , would signal the presence of new right - handed currents .
we mention that non - vanishing cp - violating triple products in pure penguin decays like @xmath327 would not be a signal for right - handed currents if significant strong phase differences ( @xmath328 ) existed between @xmath329 and @xmath330 @xcite .
there is some experimental indication for such phase differences @xcite , which is to be expected if annihilation amplitudes are important . -.35
cm -0.2 cm right - handed currents are conventionally associated with effective operators @xmath331 , obtained from the standard model operators @xmath332 by interchanging @xmath333 .
the final states in @xmath334 ( @xmath335 ) are parity - even ( parity - odd ) , so that the ith pair of wilson coefficients enters as @xcite @xmath336 the different combinations allow for large modifications to ( [ smpred ] ) .
@xmath337 suffers from prohibitively large theoretical uncertainties .
however , @xmath314 is a much cleaner observable . for illustration ,
new contributions to the qcd penguin operators are considered in figure 5 ( right ) . at the new physics matching scale @xmath338 , these can be parametrized as @xmath339 . for simplicity , we take @xmath340 and consider two cases : @xmath341 ( lower bands ) or @xmath342 ( upper bands ) , corresponding to @xmath343 or @xmath344 , and @xmath345 or @xmath346 , respectively .
the default curves and error bands are obtained as in the standard model case .
clearly , moderately sized right - handed currents could increase @xmath347 well beyond the standard model range if @xmath323 . however , new left - handed currents would have little effect .
_ dipole operators versus four - quark operators _ the suppression of dipole operator effects in the transverse modes has important implications for new physics searches . for example , in pure penguin decays to cp - conjugate final states @xmath348 , e.g. , @xmath349 , if the transversity basis time - dependent cp asymmetry parameters @xmath350 and @xmath351 are consistent with @xmath352 , and @xmath353 is not , then this would signal new cp violating contributions to the chromomagnetic dipole operators .
however , deviations in @xmath354 or @xmath355 would signal new cp violating four - quark operator contributions .
if the triple - products @xmath356 and @xmath357 @xcite do not vanish and vanish , respectively , in pure penguin decays , then this would also signal new cp violating contributions to the chromomagnetic dipole operators .
( this assumes that a significant strong phase difference is measured between @xmath358 and @xmath330 . ) however , non - vanishing @xmath357 , or non - vanishing transverse direct cp asymmetries would signal the intervention of four - quark operators .
the above would help to discriminate between different explanations for an anomalous time - dependent cp asymmetry in @xmath359 , i.e. , @xmath360 , which fall broadly into two categories : radiatively generated dipole operators , e.g. , supersymmetric loops ; or tree - level four - quark operators , e.g. , flavor changing ( leptophobic ) @xmath361 exchange @xcite , @xmath362-parity violating couplings @xcite , or color - octet exchange @xcite . finally , a large @xmath314 would be a signal for right - handed _ four - quark _ operators .
_ @xmath7 violation and @xmath363 _ we have seen that the large transverse @xmath364 polarization can be accounted for in the standard model via the qcd penguin annihilation graphs .
would this necessarily imply large transverse @xmath365 polarizations ? to answer this question we need to address @xmath366 flavor symmetry breaking in annihilation . for simplicity
, we have thus far estimated the annihilation amplitudes using asymptotic light - cone distribution amplitudes @xcite . however , for light mesons containing a single strange quark , non - asymptotic effects should shift the weighting of the distribution amplitudes towards larger strange quark momenta .
@xmath7 violation in processes involving @xmath367 popping versus light quark popping , e.g. , annihilation , can therefore be much larger than the canonical 20% , due to the appearance of inverse moments of the distribution amplitudes @xcite .
this can account for the order of magnitude hierarchy between the @xmath368 and @xmath369 rates @xcite .
similar considerations may also explain the @xmath8 flavor violation empirically observed in high energy @xmath370 fragmentation , e.g. , in kaon versus pion multiplicities , or @xmath371 versus @xmath372 multiplicities at the @xmath373 . in particular , the relative probability for @xmath374 popping versus @xmath375 or @xmath376 popping in jetset fragmentation monte carlo s must be tuned to @xmath377 @xcite . -.35 cm
-0.2 cm the dominant @xmath10 qcd annihilation amplitudes @xmath378 and @xmath211 involve products of inverse moments , see ( [ a3f ] ) . the @xmath7 violation discussed above can be estimated by including the second and third terms in the gegenbauer expansions for the distribution amplitudes .
the first gegenbauer moments @xmath379 , @xmath380 determine the asymmetries of the corresponding leading - twist distribution amplitudes , i.e. , the inverse moments of @xmath381 are given by @xmath382 , and similarly for @xmath383 .
note that the first moments vanish for the symmetric @xmath384 and @xmath372 mesons . for illustration ,
two sets of intervals for the moments are considered : @xmath385 , @xmath386 , @xmath387 , as in @xcite ; and a more restrictive set with same central values but halved intervals .
@xmath388 is required , since the @xmath18-quark should carry the larger fraction of the @xmath371 light - cone momentum .
similarly , we require that the light - cone momentum fraction of the @xmath18-quark ( light antiquark ) in the @xmath371 is greater than ( less than ) that of the quark ( antiquark ) in the @xmath372 and @xmath20 or , equivalently , @xmath389 which imposes the constraints @xmath390 .
the logarithmic divergences in the inverse moments are parametrized as before . for simplicity ,
the @xmath269 are taken equal and independent of the final state , with @xmath391 and @xmath392 $ ] . in @xcite ,
@xmath7 violation was studied with asymptotic distribution amplitudes by varying the @xmath269 .
the scatter plots in figure 6 illustrate @xmath7 violation in the qcd penguin annihilation amplitudes @xmath393 and @xmath394 , see ( [ vvamplitudes ] ) . for simplicity ,
only the contributions of @xmath209 are included ( to good approximation , the other terms in @xmath395 can be neglected ) .
two cases are shown : arbitrary strong phase @xmath396 $ ] ; and vanishing strong phase @xmath397 .
the gegenbauer moments are sampled in the intervals given above , subject to the constraint ( [ constraint ] ) ; @xmath268 lies in the usual range , @xmath398 $ ] , and the remaining inputs are set to their default values . for comparison , the default non - annihilation @xmath399 amplitude ( in units of @xmath400 )
is @xmath401 , with negligible strong phase .
the total negative helicity @xmath402 amplitude observed is about a factor of three larger , corresponding to @xmath403 and , according to figure 6 , @xmath404 .
we therefore expect @xmath405 in the standard model .
the other @xmath406 modes , containing ` tree - level ' amplitudes , will be discussed in @xcite .
babar has measured @xmath407 @xcite .
given the large errors , this is still consistent with the low @xmath408 longitudinal polarization .
we have presented an analysis of polarization in @xmath409 decays to light vector meson pairs beyond naive factorization , using qcd factorization .
formally , the longitudinal polarization satisfies @xmath410 , as in naive factorization .
however , we saw that the contributions of a particular qcd penguin annihilation graph which is formally @xmath223 can be @xmath8 numerically in longitudinal and negative helicity @xmath411 @xmath67 decays .
consequently , the observation of @xmath412 can be accounted for in the standard model , with large theoretical errors .
however , @xmath413 and @xmath414 are predicted to be close to 1 with small theoretical errors , in agreement with observation .
we have shown that the ratio of transverse rates in the transversity basis satisfies @xmath415 , in agreement with naive power counting .
a ratio in excess of the predicted standard model range would signal the presence of new right - handed currents in dimension-6 four - quark operators .
the maximum ratio attainable in the standard model is sensitive to the @xmath416 form factor combination @xmath417 or @xmath52 , see ( [ rperp ] ) , which controls helicity suppression in form factor transitions .
existing model determinations give a positive sign for @xmath52 , which would imply @xmath418 in the standard model .
however , the maximum would increase for negative values .
the magnitude and especially the sign of @xmath419 is an important issue which needs to be clarified with dedicated lattice studies .
the contributions of the @xmath420 dipole operators to the transverse modes were found to be highly suppressed , due to angular momentum conservation .
comparison of cp - violation involving the longitudinal modes with cp - violation only involving the transverse modes , in pure penguin @xmath3 decays , could therefore distinguish between new contributions to the dipole and four - quark operators .
more broadly , this could distinguish between scenarios in which new physics effects are loop induced and scenarios in which they are tree - level induced , as it is difficult to obtain @xmath8 cp - violating effects from dimension-6 operators beyond tree - level .
we have seen that the asymmetry of the @xmath421 meson light - cone distributions generically leads to @xmath8 @xmath7 flavor symmetry violation in annihilation amplitudes , as pointed out in @xcite .
in particular , @xmath367 popping can be substantially suppressed relative to light quark popping .
this implies that the longitudinal polarizations should satisfy @xmath422 in the standard model .
consequently , @xmath423 would indicate that @xmath424-spin violating new physics entering mainly in the @xmath425 channel is at least partially responsible for the small values of @xmath426 .
one possibility would be right - handed vector currents ; they could interfere constructively ( destructively ) in the perpendicular ( longitudinal and parallel ) transversity amplitudes .
alternatively , a parity symmetric realization would only affect , and increase the perpendicular amplitude @xcite .
either case would lead to @xmath427 , and could thus be ruled out .
a more exotic possibility is tensor currents ; they would contribute to the longitudinal and transverse amplitudes at subleading and leading power , respectively .
if left - handed , i.e. , of the form @xmath428 , then @xmath429 would be maintained .
finally , @xmath8 @xmath7 violation is possible in all qcd penguin amplitudes , given that the annihilation topology components can be comparable to , or greater than the penguin topology components .
this is especially true of decays to @xmath430 and @xmath154 final states which , unlike decays to @xmath431 final states , do not receive large contributions from @xmath432 chirality penguin topology matrix elements .
certain applications of @xmath7 symmetry in @xmath0 decays should therefore be reexamined .
this work originated during the 2003 summer workshop on flavor physics at the aspen center for physics .
i would like to thank martin beneke , alakhabba datta , keith ellis , andrei gritsan , yuval grossman , dan pirjol , and ian stewart for useful conversations .
i am especially grateful to matthias neubert for discussions of many points addressed in this paper .
this work was supported by the department of energy under grant de - fg02 - 84er40153 .
j. charles , a. le yaouanc , l. oliver , o. pene and j. c. raynal , phys .
d * 60 * , 014001 ( 1999 ) [ arxiv : hep - ph/9812358 ] .
m. beneke and t. feldmann , nucl .
b * 592 * , 3 ( 2001 ) [ hep - ph/0008255 ] .
a. ali , g. kramer and c .- d .
lu , phys .
d * 58 * , 094009 ( 1998 ) .
p. ball and v. m. braun , phys .
d * 58 * , 094016 ( 1998 ) [ hep - ph/9805422 ] .
l. del debbio , j. m. flynn , l. lellouch and j. nieves , phys .
b * 416 * , 392 ( 1998 ) [ hep - lat/9708008 ] .
p. colangelo , f. de fazio , p. santorelli and e. scrimieri , phys . rev .
d * 53 * , 3672 ( 1996 ) [ erratum - ibid .
d * 57 * , 3186 ( 1998 ) ] [ arxiv : hep - ph/9510403 ] . m. bauer , b. stech and m. wirbel , z. phys .
c * 34 * , 103 ( 1987 ) .
b. o. lange and m. neubert , arxiv : hep - ph/0311345 .
r. j. hill , t. becher , s. j. lee and m. neubert , arxiv : hep - ph/0404217 . g. burdman and g. hiller , phys .
d * 63 * , 113008 ( 2001 ) [ arxiv : hep - ph/0011266 ] .
m. beneke and t. feldmann , nucl .
b * 685 * , 249 ( 2004 ) [ arxiv : hep - ph/0311335 ] .
b. aubert _ et al . _ [ babar collaboration ] , phys .
lett . * 91 * , 171802 ( 2003 ) [ arxiv : hep - ex/0307026 ] , and hep - ex/0303020 .
b. aubert _ et al . _ [ babar collaboration ] , arxiv : hep - ex/0404029 , submitted to phys
lett .
j. g. smith [ babar collaboration ] , talk at rencontres de moriond qcd04 , march 2004 , arxiv : hep - ex/0406063 .
a. gritsan [ babar collaboration ] , lbnl seminar , april 2004 [ http://costard.lbl.gov/ gritsan / rpm / babar - coll-0028.pdf ] .
m. beneke , g. buchalla , m. neubert and c. t. sachrajda , phys .
* 83 * , 1914 ( 1999 ) [ hep - ph/9905312 ] ; nucl .
b * 591 * , 313 ( 2000 ) [ hep - ph/0006124 ] ; m. beneke , g. buchalla , m. neubert and c. t. sachrajda , nucl .
b * 606 * , 245 ( 2001 ) [ arxiv : hep - ph/0104110 ] .
m. beneke and m. neubert , nucl .
b * 675 * , 333 ( 2003 ) [ arxiv : hep - ph/0308039 ] .
h. y. cheng and k. c. yang , phys .
b * 511 * , 40 ( 2001 ) [ arxiv : hep - ph/0104090 ] .
x. q. li , g. r. lu and y. d. yang , phys .
d * 68 * , 114015 ( 2003 ) [ arxiv : hep - ph/0309136 ] .
a. l. kagan , talks at super b factory workshop , honolulu , january 2004 ; les rencontres de physique de la valle daoste , february 2004 . c. w. bauer , d. pirjol , i. z. rothstein and i. w. stewart , arxiv : hep - ph/0401188 .
a. hardmeier , e. lunghi , d. pirjol and d. wyler , nucl .
b * 682 * , 150 ( 2004 ) [ arxiv : hep - ph/0307171 ] .
d. becirevic , arxiv : hep - ph/0211340 .
v. barger , c. w. chiang , p. langacker and h. s. lee , phys .
b * 580 * , 186 ( 2004 ) [ arxiv : hep - ph/0310073 ] .
a. datta , phys .
d * 66 * , 071702 ( 2002 ) [ arxiv : hep - ph/0208016 ] ; b. dutta , c. s. kim and s. oh , phys .
lett . * 90 * , 011801 ( 2003 ) [ arxiv : hep - ph/0208226 ] . | factorizable amplitudes in @xmath0 decays to light vector meson pairs give a longitudinal polarization satisfying @xmath1 .
this remains _ formally _ true when non - factorizable graphs are included in qcd factorization , and is numerically realized in @xmath2 . in @xmath3 decays a qcd penguin annihilation graph can effectively contribute at leading power to the transverse and longitudinal amplitudes .
the observed longitudinal polarization , @xmath4 , can therefore be accounted for in the sm .
the ratio of transverse rates @xmath5 provides a sensitive test for new right - handed currents .
the _ transverse _
@xmath6 dipole operator amplitudes are highly suppressed .
cp violation measurements can therefore discriminate between new contributions to the dipole and four quark operators .
@xmath7 violation in qcd penguin amplitudes can easily be @xmath8 , in general , due to annihilation .
implications for @xmath9 polarization and new physics searches are pointed out .
uctp-102 - 04 + hep - ph/0405134 * polarization in @xmath10 decays * alexander l. kagan + |
influenza epidemics are observed around the world during the wintertime and with a strong seasonal component in temperate regions @xcite@xcite .
influenza is a disease caused by the influenza virus , an rna virus belonging to the orthomyxoviridae @xcite .
many features are common with those of the paramyxovirus infections of the acute upper respiratory tract
. typical symptoms of the disease are characterized by fever , myalgia , severe malaise , non - productive cough , and sore throats .
the disease spreads when an infected individual coughs or sneezes and sends the virus into the air , and other susceptible individuals inhale the virus .
the virus is also believed to be transmitted when a person touches a surface that is contaminated with the virus ( _ e.g. _ door knob , etc . ) and then touches the nose or eyes .
infected individuals can transmit the virus almost within a day following infection ( _ i.e. _ latent period ) .
although it is generally believed that infected individuals can pass the virus for 3 - 7 days following symptom onset , there is some uncertainty on the duration of the infectious period .
the generation time ( _ i.e. _ sum of latent and infectious periods ) for influenza , reported and assumed in the literature , ranges from 3 days @xcite@xcite@xcite to about 6 days @xcite@xcite .
individuals that are infected with influenza are believed to become permanently immune against the specific virus strain .
hence , the virus is able to persist in the human population through relatively minor ( single point ) mutations in the virus composition known as drifts .
influenza ( sub)types a / h3n2 , a / h1n1 and b are currently co - circulating in the human population @xcite .
major changes in the virus composition via recombination or gene reassortment processes ( known as genetic shifts ) can lead to the emergence of novel influenza viruses with the potential of generating dramatic morbidity and mortality levels around the world @xcite .
the 1918 - 19 influenza pandemic known as the * spanish influenza * caused by the influenza virus a ( h1n1 ) has been the most devastating in recent history with estimated worldwide mortality ranging from 20 to 100 million deaths @xcite@xcite with a case fatality of @xmath0-@xmath1 percent @xcite@xcite .
the worldwide 1918 influenza pandemic spread in three waves starting from midwestern united states in the spring of 1918 @xcite@xcite .
the deadly second wave began in late august probably in france while the third wave is generally considered as part of normal more scattered winter outbreaks similar to those observed after the 1889/90 pandemic @xcite .
subsequent pandemics during the 20th century are attributed to subtyes a ( h2n2 ) from 1957 - 58 ( asian influenza ) and a ( h3n2 ) in 1968 ( hong kong influenza ) @xcite .
the ability to quickly detect and institute control efforts at the early stage of an influenza pandemic is directly linked to the final levels of morbidity and mortality in the population @xcite . to appropriately assess the disaster size of a probable future pandemic
, we have to quantify the transmission potential ( and its associated uncertainty ) .
although it is difficult to directly measure the transmissibility of a future pandemic , historical epidemiologic data is readily available from previous pandemics , and as a reference quantity for future pandemic planning , mathematical and statistical analyses of the historical data can offer various insights .
in particular , because many historical records tend to document only the temporal distribution of cases or deaths ( _ i.e. _ epidemic curve ) , we modelers have faced with a difficult need to clarify the mechansms of the spread of influenza using such time - evolution data alone . in this paper
, we review a number of mathematical and statistical methods for the estimation of the transmission potential of pandemic influenza , focusing on theoretical techniques to maximize the utility of the temporal distribution of influenza cases .
the methods that have been incorporated in this review include the applications of epidemiologically structured epidemic models , explicitly duration - structured epidemic system , and stochastic processes ( _ i.e. _ branching and counting processes ) .
whereas this review does not cover the spread of influenza in space , spatial heterogeneity in transmission and the growing interest in the role of contact network are briefly discussed as the future challenge .
the basic reproduction number , @xmath2 ( pronounced as _ r nought _ ) , is a key quantity used to estimate transmissibility of infectious diseases .
theoretically , @xmath2 is defined as the average number of secondary cases generated by a single primary case during its entire period of infectiousness in a completely susceptible population @xcite .
as the epidemic progresses , the number of susceptible individuals is decreased due to infection , and the reproduction number decays following @xmath3 where @xmath4 and @xmath5 are , respectively , the number of susceptible individuals at time @xmath6 and before the epidemic starts ; the latter is equivalent to the total population size @xmath7 given that all individuals are susceptible before the beginning of an epidemic .
clearly , this definition only applies to ( novel ) emerging infectious diseases ( _ e.g. _ the epidemic of severe acute respiratory syndrome ( sars ) from 2002 - 3 ) or re - emerging infectious diseases that had not circulated in the population in question for long enough to allow for residual immunity in the population to disappear due to births and deaths .
+ the reproduction number is directly related to the type and intensity of interventions necessary to control an epidemic since the objective is to make @xmath8 as soon as possible . to achieve @xmath8 ,
one or a combination of control strategies may be implemented .
for example , one of the best known uses of @xmath2 is in determining the critical coverage of immunization required to eradicate a disease in a randomly mixing population .
that is , when vaccine is available against a disease in question , it is of interest to estimate the critical proportion of the population that needs to be vaccinated ( _ i.e. _ vaccination coverage ) in order to attain @xmath9 @xcite@xcite .
for example , in the u.s prior to 1963 , a vaccine against measles was not available and hence recurrent epidemics of measles were observed with approximately @xmath10 million cases per year and a mean of @xmath11 deaths .
the introduction of the vaccine in the u.s . reduced the incidence by 98 percent .
+ the critical vaccination coverage , @xmath12 ( in a randomly mixing population ) can be estimated from the @xmath2 of the disease in question as follows @xcite : + @xmath13 where @xmath14 is the efficacy ( _ i.e. _ direct effectiveness ) of vaccination @xcite .
@xmath12 given in ( [ eqn_intro1 ] ) suggests that the disease could be eradicated even when all susceptible individuals are not vaccinated .
the protection conferred to the population by achieving a critical vaccination coverage is known as * herd immunity * @xcite@xcite . + a brief history of the theoretical developments on the basic reproduction number and its analytical computation via epidemic modeling is given elsewhere @xcite@xcite@xcite .
the mathematical definition and calculation of @xmath2 using next - generation arguments is described initially by odo diekmann and his colleagues @xcite@xcite , where @xmath2 is the dominant eigenvalue of the resulting next generation matrix .
further elaborations and reviews can be found elsewhere @xcite@xcite@xcite@xcite@xcite@xcite@xcite .
classically , rather than the threshold phenomena , @xmath2 was used to suggest the _ severity _ of an epidemic , because the proportion of those experiencing infection at the end of an epidemic depends only on @xmath2 @xcite ( see section 3 ) .
+ statistical methods to quantitatively estimate @xmath2 have been reviewed by klaus dietz @xcite .
depending on the characteristics of data and underlying assumptions of the models , @xmath2 can be estimated using various different approaches @xcite .
in addition to the final size equation , @xmath2 of an epidemic of newly emerging disease can be estimated from the * intrinsic growth rate * @xcite@xcite@xcite@xcite@xcite@xcite , which is also referred to as the _ rate of natural increase _ , suggesting the natural growth rate of infected individuals in a fully susceptible population ( discussed in section 4 ) . moreover , for simple epidemic models with relatively few parameters , @xmath2 can be estimated with other unobservable quantities by rigorous curve fitting of model equations to the observed epidemic data ( discussed in section 3)@xcite@xcite@xcite . not only @xmath2 but also @xmath15 can be estimated from the temporal distribution of infectious diseases , reconstructing the transmission network or inferring the time - inhomogeneous number of secondary transmissions @xcite@xcite@xcite@xcite .
+ to estimate the basic reproduction number of endemic diseases , different approaches are taken .
one would need first to carry out serological surveys to quantify the fraction of the population that is effectively protected against infection ( _ i.e. _ age- and/or time - specific proportion of those possessing acquired immunity needs to be estimated ) . through this effort ,
the * force of infection * , the rate at which susceptible individuals get infected , is estimated @xcite .
for example , this is the case for rubella , mumps and measles ( that are still circulating in some regions of the world even when high effective vaccination coverage is achieved ) .
although the estimation of @xmath2 for endemic diseases is out of the scope of this review , methodological details and the applications to estimate the force of infection and @xmath2 can be found elsewhere @xcite@xcite@xcite@xcite@xcite@xcite@xcite .
+ in practice , the reproduction number denoted simply by @xmath16 and defined as the number of secondary cases generated by a primary infectious cases in a partially protected population might be useful . @xmath16 can also be estimated from the initial growth phase of an epidemic in such a partially immunized population . in a randomly mixing population ,
the relationship between the basic reproduction number ( @xmath2 ) and the reproduction number ( @xmath16 ) is given by @xmath17 where @xmath18 is the proportion of the population that is effectively protected against infection ( in the beginning of an epidemic ) .
besides , for many recurrent infectious diseases including seasonal influenza , estimating the background immunity @xmath18 in the population is extremely difficult due to cross - immunity of antigenically - related influenza strains and vaccination campaigns . + with reagard to seasonal influenza , the reproduction number ( @xmath16 ) over the last 3 decades has been estimated at @xmath19 ( se @xmath20 ) in the united states , france , and australia with an overall range of @xmath21 @xcite
. an @xmath16 estimate of @xmath22 has been reported for a single a / h3n2 season in france @xcite , and some estimates have been reported in the range @xmath23-@xmath24 for several consecutive influenza seasons in england and wales @xcite@xcite .
a particularly high estimate of @xmath16 has been suggested for the 1951 influenza epidemic in england and canada @xcite .
+ because influenza pandemics such as the spanish flu from 1918 - 19 are associated to the emergence of novel influenza strains to which most of the population is susceptible , it might be reasonable to assume that the reproduction number @xmath25 .
previous studies have estimated that @xmath2 of the 1918 - 19 influenza pandemic ranged between @xmath22 and @xmath26 @xcite@xcite@xcite@xcite@xcite@xcite@xcite@xcite@xcite@xcite@xcite@xcite depending on the specific location and pandemic wave considered , type of data , estimation method , and level of spatial aggregation , which has ranged from small towns to entire nations with several million inhabitants .
table [ tablewaves ] lists estimates of @xmath2 in recent studies .
the variability of @xmath2 estimates suggests that local factors , including geographic and demographic conditions , could play an important role in disease spread @xcite@xcite@xcite . in the following sections , we review how these estimates are obtained and how we shall interpret the estimates , starting from a simple structured epidemic model proposed in 1927
mathematical models provide a unique way to analyze the transmission dynamics and study various different scenarios associated to the spread of communicable diseases in population(s ) @xcite@xcite@xcite .
the history of the mathematical modeling of infectious diseases greatly remounts to the study of sir ronald ross in 1911 @xcite who invented a classic malaria model and also discovered the mosquito - borne transmission mechanisms of malaria . employing a mass action principle for the spread of malaria ,
ross explored the effects of controling the mosquito population using simple mathematical models @xcite .
following his effort , kermack and mckendrick introduced a classical sir ( susceptible - infectious - removed ) epidemic model in 1927 , which is most frequently utilized in the present day , given by the following system of nonlinear ordinary differential equations ( odes ) @xcite : @xmath27 where @xmath4 denotes susceptible individuals at time @xmath6 ; @xmath28 , infected ( assumed infectious ) individuals at time @xmath6 ; and @xmath15 , recovered ( assumed permanently immune ) individuals at time @xmath6 ; @xmath29 is the transmission rate ; @xmath30 the recovery rate ; and @xmath7 is the total population size which is assumed constant for a closed population ( _ i.e. _ a population without immigration and emmigration ) .
susceptible individuals in contact with the virus enter the infectious class ( @xmath31 ) at the rate @xmath32 .
that is , homogeneous mixing between individuals is assumed . + the basic reproduction number , @xmath2 , for the epidemic system ( [ eq1 ] ) is given by the product of the transmission rate and the mean infectious period .
that is : @xmath33 classically , @xmath2 has been known as a quantity to suggest * severity * of an epidemic @xcite .
indeed , analytical expression of @xmath2 in ( [ eq_r0 ] ) is derived simply by solving the above system ( [ eq1 ] ) .
replacing @xmath28 in the right hand side of @xmath34 by ( @xmath35)@xmath36 , we get @xmath37 integrating both sizes of ( [ fsize1 ] ) from 0 to infinity , @xmath38 since @xmath39 , and because we assume @xmath40 and @xmath41 , equation ( [ fsize2 ] ) can be rewritten as @xmath42 in the above equation ( [ fsize3 ] ) , * final size * , _ i.e. _ , the proportion of those experiencing infection among a total number of individuals in a community following a large scale epidemic , is defined as @xmath43 .
that is , @xmath44 therefore , the following * final size equation * of an autonomous sir ( or seir ) model is obtained : @xmath45 equation ( [ fsize5 ] ) can be analytically derived using both deterministic ( models governed by odes or partial differential equations ( pdes))@xcite and stochastic models @xcite . despite the usefulness of ( [ eq1 ] ) , sir assumptions given by odes are not always directly applicable to real data .
one of the reasons include that there is no disease where an infected individual can cause secondary transmission immediately after his / her infection
. + accordingly , we have used slightly extended compartmental models in the previous studies to describe the transmission dynamics of the 1918 - 19 influenza pandemic and estimate the reproduction number @xcite@xcite .
we now describe two different seir ( susceptible - exposed - infectious - removed ) models that have been used to estimate the reproduction number .
the first model is the simple seir model , and the second model accounts for asymptomatic and hospitalized individuals .
+ the simple seir model classifies individuals as susceptible ( s ) , exposed ( e ) , infectious ( i ) , recovered ( r ) , and dead ( d ) @xcite .
susceptible individuals in contact with the virus enter the exposed class at the rate @xmath46 , where @xmath29 is the transmission rate , @xmath28 is the number of infectious individuals at time @xmath6 and @xmath47 is the total population for any @xmath6 .
the entire population is assumed to be susceptible at the beginning of the epidemic .
individuals in latent period ( e ) progress to the infectious class at the rate @xmath48 ( where @xmath49 suggests the mean latent period ) .
we assume homogeneous mixing ( _ i.e. _ random mixing ) between individuals and , therefore , the fraction @xmath50 is the probability of a random contact with an infectious individual in a population of size @xmath7 .
since we assume that the time - scale of the epidemic is much faster than characteristic times for demographic processes ( natural birth and death ) , background demographic processes are not included in the model .
infectious individuals either recover or die from influenza at the mean rates @xmath30 and @xmath51 , respectively .
recovered individuals are assumed protected for the duration of the outbreak .
the mortality rate is given by @xmath52 [ cfp/(1-cfp ) ] , where cfp is the mean case fatality proportion .
the transmission process can be modeled using the system of nonlinear differential equations : @xmath53 where @xmath54 is the cumulative number of infectious individuals .
the basic reproduction number of the above system ( [ eqn12 ] ) is given by the product of the mean transmission rate and the mean infectious period , @xmath55 .
+ a more complex seir model ( figure [ figdiagram ] ) classifies individuals as susceptible ( @xmath56 ) , exposed ( @xmath57 ) , clinically ill and infectious ( @xmath31 ) , asymptomatic and partially infectious ( @xmath58 ) , diagnosed and reported ( @xmath59 ) , recovered ( @xmath16 ) , and death ( @xmath60 ) .
the birth and natural death rates are assumed to have a common rate @xmath61 ( 60-year life expectancy as in @xcite ) .
the entire population is assumed susceptible at the beginning of the pandemic wave .
susceptible individuals in contact with the virus progress to the latent class at the rate @xmath62 where @xmath29 is the transmission rate , and @xmath63 is a reduction factor in the transmissibility of the asymptomatic class ( @xmath58 ) . since there is no explicit evidence estimating and proving the effectiveness of public health interventions , and because a high burden was placed upon the sanitary and medical sectors , diagnosed / hospitalized individuals ( @xmath59 ) are assumed equally infectious .
although it is difficult to explicitly evaluate the difference in infectiousness between general community and hospital , we roughly made this assumption since 78 percent of the nurses of the san francisco hospital contracted influenza @xcite .
a more rigorous assumption requires either statistical analysis of more detailed time - series data @xcite or an epidemiological comparison of specific groups by contact frequency @xcite .
the total population size at time @xmath6 is given by @xmath64 .
we assumed homogeneous mixing of the population and , therefore , the fraction @xmath65 is the probability of a random contact with an infectious individual . a proportion @xmath66 of latent individuals progress to the clinically infectious class ( @xmath31 ) at the rate @xmath48 while the remaining ( @xmath67 ) progress to the asymptomatic partially infectious class ( @xmath58 ) at the same rate @xmath48 ( fixed to 1/1.9 days@xmath68 @xcite ) .
asymptomatic cases progress to the recovered class at the rate @xmath69 .
clinically infectious individuals ( class @xmath31 ) are diagnosed ( reported ) at the rate @xmath70 or recover without being diagnosed ( e.g. , mild infections , hospital refusals ) at the rate @xmath69 .
diagnosed individuals recover at the rate @xmath71 or die at rate @xmath51 .
the mortality rates were adjusted according to the case fatality proportion ( cfp ) , such that @xmath72 . + the transmission process can be modeled using the following system of nonlinear differential equations : @xmath73 we assume the cumulative number of influenza notifications , our observed epidemic data , is given by @xmath54 .
seven model parameters ( @xmath29 , @xmath74 , @xmath70 , @xmath75 , @xmath76 , @xmath77 , @xmath78 ) are estimated from the epidemic curve by least squares fitting as explained below . the reproduction number for model ( [ eqn2 ] )
is given by ( see @xcite ) : @xmath79 and the clinical reporting proportion is given by : @xmath80 in the simplest manner , model parameters can be estimated via least - square fitting of the model solution to the observed data .
that is , one looks for the set of parameters @xmath81 whose model solution best fits the epidemic data by minimizing the sum of the squared differences between the observed data @xmath82 and the model solution @xmath83 .
that is , we minimize : @xmath84 the standard deviation of the parameters can be estimated by computing the asymptotic variance - covariance @xmath85 matrix of the least - squares estimate by @xcite : @xmath86 which can be estimated by @xmath87 where @xmath88 is the total number of observations , @xmath89 is the estimated variance , and @xmath90 are numerical derivatives of @xmath91 .
estimates of @xmath92 can be obtained by substituting the corresponding individual parameter estimates into an analytical formula of @xmath2 .
further , using the delta method @xcite , we can derive an expression for the variance of the estimated basic reproduction number @xmath92 .
an expression for the variance of @xmath2 for the simple seir model ( equations [ eqn12 ] ) is given by : @xmath93 this expression depends on the variance ( denoted by @xmath94 ) of the individual parameter estimates as well as their covariance ( denoted by @xmath95 ) .
another method to generate uncertainty bounds on the reproduction number is generating bootstrap confidence intervals by generating sets of realizations of the best - fit curve @xmath54 @xcite .
each realization of the cumulative number of case notifications @xmath96 ( @xmath97 , @xmath0 , @xmath98 , @xmath99 ) is generated as follows : for each observation @xmath54 for @xmath100 , @xmath101 , @xmath98 , @xmath88 days generate a new observation @xmath102 for @xmath103 ( @xmath104 ) that is sampled from a _ poisson _ distribution with mean : @xmath105 ( the daily increment in @xmath54 from day @xmath106 to day @xmath6 ) .
the corresponding realization of the cumulative number of influenza notifications is given by @xmath107 where @xmath108 , @xmath0 , @xmath101 , @xmath98 , @xmath88 .
the reproduction number was then estimated from each of @xmath109 simulated epidemic curves to generate a distribution of @xmath16 estimates from which simple statistics can be computed including @xmath110 confidence intervals .
these statistics need to be interpreted with caution .
for example , @xmath110 confidence intervals for @xmath16 derived from our bootstrap sample of @xmath16 should be interpreted as containing @xmath110 of future estimates when the same assumptions are made and the only noise source is observation error .
it is tempting but incorrect to interpret these confidence intervals as containing the _ true _ parameters with probability @xmath111 .
figure [ figr - seijr ] shows the temporal distributions of the reproduction number and the proportion of the clinical reporting obtained by the bootstrap method after fitting the complex seir epidemic model to the initial phase of the fall influenza wave using 17 epidemic days of the spanish flu pandemic in san francisco , california .
in addition to the estimation of @xmath2 , it is of practical importance to evaluate time - dependent variations in the transmission potential .
explanation of the time course of an epidemic can be partly achieved by estimating the effective reproduction number , @xmath15 , defined as the actual average number of secondary cases per primary case at time @xmath6 ( for @xmath112 ) @xcite@xcite@xcite@xcite@xcite should not be confused with the number of removed individuals using the same notation . in the following arguments of this paper , @xmath15 denotes the effective reproduction number . ] .
although effective interventions against spanish influenza may have been limited in the early 20th century , it is plausible that the contact frequency leading to infection varied with time owing to the huge number of deaths and dissemination of information through local media ( _ e.g. _ newspapers )
. @xmath15 shows time - dependent variation with a decline in susceptible individuals ( intrinsic factors ) and with the implementation of control measures ( extrinsic factors ) .
if @xmath8 , it suggests that the epidemic is in decline and may be regarded as being _ under control _ at time @xmath6 ( vice versa , if @xmath113 ) . to appropriately understand the theoretical concept of @xmath15 ,
let us firstly consider an explicitly infection - age structured epidemic model .
whereas kermack - mckendrick model governed by odes ( i.e. sir and seir models as discussed above ) has been well - known , kermack and mckendrick had actually proposed an infection - age structured model in their initial publication in 1927 @xcite , the mathematical importance of which was recognized only after 1970s @xcite@xcite .
let us denote the numbers of susceptible and recovered individuals by @xmath4 and @xmath114 .
further , let @xmath115 be the density of infectious individuals at time @xmath6 and * infection - age * @xmath116 ( _ i.e. _ time since infection ) .
the model is given by @xmath117 where @xmath118 is referred to as the force of infection ( foi ) ( _ i.e. _ as discussed in section 2 , foi is defined as the rate at which susceptible individuals get infected ) which is given by : @xmath119 and @xmath120 is the rate of recovery at infection - age @xmath116 . it should be noted that the above model has not taken into account the background host demography ( _ i.e. _ birth and death ) . in a closed population ,
the total population size @xmath7 is thus given by : @xmath121 it should also be noted that , although @xmath122 is referred to as _ density _ , it is not meant to be a normalized density ( _ i.e. _ integral of @xmath122 over @xmath6 and @xmath116 does not sum up to 1 ) .
rather , we use density to mathematically refer to the absolute frequency in the infection - age space .
+ the system ( [ eqn_hn1 ] ) can be reasonably integrated @xmath123 where @xmath124 and @xmath125 suggests the density of initially infected individuals at the beginning of an epidemic . in the following arguments , we call @xmath126 as * incidence of infection
* ( _ i.e. _ new infections at a given point of time @xmath6 ) .
it is not difficult to derive @xmath127 from ( [ eqn_hn1])- ( [ eqn_hn5 ] ) .
thus , the subequation of @xmath128 in system ( [ eqn_hn1 ] ) is rewritten as @xmath129 \quad\ ] ] taking into account the initial condition in ( [ eqn_hn4 ] ) , equation ( [ eqn_hn6 ] ) is rewritten as @xmath130\left [ g(t)+\int_{0}^{t } \psi(\tau)j(t-\tau)\,d\tau \right ] \quad\ ] ] where @xmath131 considering the initial invasion phase ( _ i.e. _ initial growth phase of an epidemic ) , we get a linearized equation @xmath132 the equation ( [ eqn_hn10 ] ) represents lotka s integral equation , where the basic reproduction number , @xmath2 , is given by @xmath133 thus , the epidemic will grow if @xmath134 and decline to extinction if @xmath135 .
the above model can yield the same final size equation as seen in models governed by odes @xcite . +
assuming that the infection - age distribution is stable , we get a simplified renewal equation @xmath136 where @xmath137 is the product of @xmath138 and @xmath5 .
moreover , assuming that we observe an exponential growth of incidence during the initial phase ( i.e. @xmath139 where @xmath48 and @xmath140 are , respectively , a constant ( @xmath141 ) and the intrinsic growth rate ) , the following relationship must be met @xmath142 replacing @xmath143 in the right hand side of ( [ eqn_hn12 ] ) by ( [ eqn_hn13 ] ) , we get @xmath144 removing @xmath126 from both sides of ( [ eqn_hn14 ] ) , we get the lotka - euler characteristic equation:@xmath145 further , we consider a probability density of the * generation time * ( _ i.e. _ the time from infection of an individual to the infection of a secondary case by that individual @xcite ) , denoted by @xmath146 : @xmath147 using ( [ eqn_hn16 ] ) , the equation ( [ eqn_hn15 ] ) can be replaced by @xmath148 the equations ( [ eqn_hn13])-([eqn_hn17 ] ) are what wallinga and lipsitch discussed in a recent study @xcite , reasonably suggesting the relationship between the generation time and @xmath2 .
accordingly , the estimator of @xmath2 using the intrinsic growth rate is given by : @xmath149 where @xmath150 is the moment generating function of the generation time distribution @xmath146 , given the intrinsic growth rate @xmath140 @xcite .
is not used for equation ( [ eqn_hn18 ] ) and rather document ( [ eqn_hn18 ] ) as the estimator of @xmath16 .
most likely , there are two reasons for this .
first , we can not assure if all individuals are susceptible to pandemic influenza before the start of epidemic ( as discussed in section 2 ) .
second , we can not assume that infection - age distribution is stable during the initial growth phase , which is highlighted in ( [ eqn_hn4 ] ) .
thus , it should be remembered that the above discussion is mathematically tight in theory , but there are certain number of assumptions to apply the concept to observed data . since writing @xmath16 alone is always confusing ( as it is unclear if @xmath16 is concerned with time or immunity status ) , here we use @xmath2 instead . ]
equation ( [ eqn_hn18 ] ) significantly improved the issue of estimating @xmath2 using the intrinsic growth rate alone , because ( [ eqn_hn18 ] ) permits validating estimates of @xmath2 by various different distributional assumptions for @xmath146 .
the issue of assuming explicit distributions for latent and infectious periods has been highlighted in recent studies @xcite@xcite@xcite@xcite,@xcite and indeed , this point is in part addressed by ( [ eqn_hn18 ] ) , because the convolution of latent and infectious periods yields @xmath146 .
moreover , since the assumed lengths of generation time most likely yielded different estimates of @xmath2 for spanish influenza by different studies @xcite , equation ( [ eqn_hn18 ] ) highlights a critical need to clarify the generation time distribution using observed data .
+ here we briefly show a numerical example .
figure [ hn_fig1 ] shows the daily number of influenza deaths during spanish influenza pandemic in a suburb of zurich , 1918 @xcite . since the non - linear phase is difficult to analyze , our interest to estimate @xmath2 with this dataset is limited to the initial growth phase only ( right panel in fig [ hn_fig1 ] ) .
even though the data represent deaths over time ( _ i.e. _ not infection events with time ) , we can directly extract the same intrinsic growth rate as practised with onset data , assuming that death data are a good proxy for morbidity data ( see our discussions in section 6 ) .
assuming exponential growth in deaths as shown in ( [ eqn_hn13 ] ) , the intrinsic growth rate @xmath140 is estimated to be 0.16 per day .
supposing that @xmath146 is arbitrarily assumed to follow a gamma distribution with mean @xmath151 and coefficient of variation , @xmath152 , @xmath2 is given by @xmath153 although there is no concensus regarding the generation time of spanish influenza , we assume it ranges from 2 - 5 days . assuming further that @xmath154 , @xmath2 is estimated to range from 1.36 ( for @xmath155 day ) to 2.07 ( for @xmath156 days ) . in the following ,
let us consider the non - linear phase of an epidemic .
derivation of @xmath2 given by ( [ eqn_hn18 ] ) assumes an exponential growth which is applicable only during the very initial phase of an epidemic ( or , when the transmission is stationary over time ) , and thus , it is of practical importance to widen the utility of above - described renewal equations in order to appropriately interpret the time - course of an influenza pandemic .
let us explicitly account for the depletion of susceptible individuals , as we deal with an estimation issue with time - inhomogeneous assumptions ( i.e. non - linear phase ) . adopting the * mass action * assumption , we get : @xmath157 where @xmath158 should be interpreted as the reproductive power at time @xmath6 and infection - age @xmath116 at which an infected individual generates secondary cases .
we refer to the latter part of equation ( [ eqn_hn182 ] ) as a non - autonomous renewal equation , where the number of new infection at time @xmath6 is proportional to the number of infectious individuals ( as assumed in the renewal equation in the initial phase ) .
+ using equation ( [ eqn_hn182 ] ) , the effective reproduction number , @xmath15 ( _ i.e. _ instantaneous reproduction number at calendar time @xmath6 ) is defined as : @xmath159 following ( [ eqn_hn19 ] ) , we can immediately see that @xmath15 with an autonomous assumption ( _ i.e. _ where contact and recovery rates do not vary with time ) is given by : @xmath160 which is shown in @xcite . in practical terms ,
equation ( [ eqn_hn20 ] ) suggests that time - varying decrease in transmission potential as well as decline in the epidemic reflects only depletion of susceptible individuals .
this corresponds to a classic assumption of the kermack and mckendrick model .
+ however , as we discussed in the beginning of this section , we postulate that human contact behaviors ( and other extrinsic factors ) modifies the dynamics of pandemic influenza , assuming that the decline in incidence does reflect not only depletion of susceptibles but also various extrinsic dynamics ( _ e.g. _ isolation , quarantine and closure of public buildings ) .
thus , instead of the assumption in ( [ eqn_hn182 ] ) , we shall assume time - inhomogeneous @xmath161 ; _ i.e. _ @xmath162 to describe @xmath158 .
+ to derive simple estimator of @xmath15 , it is convenient to assume separation of variables for @xmath158 ( implicitly assuming that the relative infectiousness to infection - age is independent of calendar time ) @xcite . under this assumption ,
@xmath158 is rewritten as the product of two functions @xmath163 and @xmath164 : @xmath165 arbitrarily assuming a normalized density for @xmath164 , _
i.e. _ , @xmath166 then , it is easy to find that @xmath167 suggesting that the function @xmath163 is equivalent to the effective reproduction number @xmath15 .
another function @xmath164 represents the density of infection events as a function of infection - age @xmath116 .
accordingly , we can immediately see that @xmath164 is exactly the same as @xmath146 , the generation time distribution .
that is , the above arguments suggest that @xmath158 ( _ i.e. _ the rate at which an infectious individual at calendar time @xmath6 and infection - age @xmath116 produces secondary transmission ) is decomposed as : @xmath168 inserting ( [ eqn_hn25 ] ) into ( [ eqn_hn21 ] ) yields an estimator of @xmath15 @xcite : @xmath169 the above equation ( [ eqn_hn26 ] ) is exactly what was proposed in applications to sars @xcite and foot and mouth disease @xcite ; _ i.e. _ discretizing ( [ eqn_hn26 ] ) to apply it to the daily incidence data ( _ i.e. _ using @xmath170 incident cases infected between time @xmath171 and time @xmath172 and descretized generation time distribution @xmath173 ) , @xmath174 was used as the estimator .
however , it should be noted that the study in sars implicitly assumed that onset data @xmath175 at time @xmath6 reflects the above discussed infection event @xmath126 .
that is , supposing that we observed @xmath176 onset cases reported between @xmath171 and @xmath172 , @xmath15 was calculated as @xmath177 where @xmath178 is the discretized * serial interval * which is defined as the time from onset of a primary case to onset of the secondary cases @xcite@xcite .
the method permits reasonable transformation of an epidemic curve ( _ i.e. _ temporal distribution of case onset ) to the estimates of time - inhomogeneous reproduction number @xmath15 . employing the relative likelihood of case @xmath48 infected by case @xmath179 using the density function of serial interval @xmath180 ; _
i.e. _ , @xmath181 using ( [ eqn_hn29 ] ) , expected value and variance of @xmath182 are given by the following @xmath183 where @xmath184 is the total number of reported case onsets at time @xmath6 @xcite . + in the present day , only by using the above described methods ( or similar concepts with similar assumptions ) , we can transform epidemic curves into @xmath15 and roughly assess the impact of control measures on an epidemic .
however , whereas the equations ( [ eqn_hn27 ] ) and ( [ eqn_hn28 ] ) are similar in theory , we need to explicitly account for the difference between onset and infection event .
in fact , when there are many asymptomatic infections and asymptomatic secondary transmissions , serial interval is not equivalent to the generation time , and thus , directly adopting the above methods would be inappropriate . since this point is particularly important in analyzing influenza data
, we discuss this issue in section 6 .
in the previous sections , we discussed several different methods to estimate @xmath2 either by ( i ) employing detailed curve fitting method assuming a structured epidemic model or ( ii ) using the intrinsic growth rate ( or doubling time @xcite@xcite ) . summarizing the above discussions
, we believe that the readers should benefit from memorizing @xmath185 for the use of the intrinsic growth rate @xmath140 in estimating @xmath2 @xcite and remembering the final size equation @xmath186 suggesting the severity of an epidemic as the theoretical concept .
indeed , estimator using either the intrinsic growth rate or final size has still continued to play an important role in discussing @xmath2 of pandemic influenza @xcite .
+ however , it should be noted that deterministic models do not permit incorporating stochasticity explicitly ( _ e.g. _ standard error for @xmath2 is determined by measurement of errors alone ) , as the models argue only _ average number of secondary transmissions _ within the assumed transmission dynamics . that is , our arguments given above explore only the time - evolution of influenza spread in
the * mean field*. to address the issue of variation in secondary transmissions , full stochastic models are called for @xcite . + from a viewpoint of data science , the discrete - time branching process , which is also referred to as galton - watson process , can reasonably assess individual heterogeneity in secondary transmissions @xcite@xcite .
as we discussed the initial growth phase of the deterministic model , let us consider the same epidemic phase where we observe a geometric increase in the number of cases by generation @xcite .
we denote the initial number of infected individuals by @xmath187 in generation 0 .
then , during the first generation , @xmath188 cases are produced by secondary transmissions of @xmath187 .
similarly , let @xmath189 be the number of infections in generation @xmath88 .
the branching process of this type assumes that every infected individual has an independently and identically distributed stochastic random variable @xmath190 representing the number of secondary cases produced by case @xmath191 in generation ( @xmath88 ) , and that environmental stochasticity and immigration / emigration can be ignored .
supposing that the pattern of secondary transmission follows a discrete probability distribution @xmath192 with @xmath48 secondary transmission(s ) ; _ i.e. _ , @xmath193 then , the expected number of secondary transmissions and the variance are given by @xmath194 in other words , the concept of probability distribution @xmath192 reflects * offspring distribution * in population ecology , and this permits explicit modeling of variations in secondary transmissions in infectious diseases @xcite@xcite .
this approach is particularly important during the initial phase of an epidemic , because the number of infectious individuals is small in this stage , and thus , it is deemed essential to take into account demographic stochasticity , _
i.e. _ , variation in the numbers of secondary transmissions by chance .
indeed , the model has been applied to observed outbreak data where we observed the extinction before growing to a major epidemic @xcite@xcite .
+ let us briefly discuss the variation in secondary transmissions and an estimation method of @xmath2 using the discrete - time branching process , deriving analytical expressions of the expected number of infected individuals in generation @xmath88 , @xmath195 and the variance @xmath196 .
it is impossible to avoid using the probability generating function ( pgf ) to discuss the branching process .
the above described @xmath190 characterize _ positive _ and _ discrete _ number of secondary transmissions , and thus , is a non - zero discrete random variable .
the pgf of @xmath76 , @xmath197 is given by @xmath198 there are two basic properties concerning @xmath199 in relation to the epidemic process . first , @xmath2 is by definition the mean value of secondary transmissions ( equation ( [ eqn_hn203 ] ) ) and , thus given by @xmath200 .
second , the probability that an infected individual does not cause any secondary transmissions @xmath201 pr(@xmath76=0 ) is given by @xmath202 , which is useful for discussing threshold phenomena and extinction @xcite . if we note that @xmath203 ( i.e. only one index case ) , the galton - watson process has the following pgf identity : @xmath204 even when there are @xmath205 cases in generation 0 ( where @xmath206 ) , we just have to assume that there are @xmath207 different independent infection - trees and thus @xmath208 from the above discussions , the expected number of cases in generation @xmath88 , @xmath209 , and the variance @xmath210 is @xmath211 the process grows geometrically if @xmath134 , stays constant if @xmath212 , and decays geometrically if @xmath135 .
these three cases are referred to as * supercritical * , * critical * , and * subcritical * , respectively . however , unlike the deterministic model , it should be remembered that critical process does not permit continued transmissions , and rather , the process becomes extinct almost surely ( i.e. probability of extinction given @xmath212 is one ) @xcite .
+ mathematically , demographic stochasticity in transmission is represented by a poisson process , which has been practiced in the application of branching processes to epidemics @xcite . assuming that mean value of secondary transmissions is a constant @xmath2 , the conditional distribution of observing @xmath213 cases , given @xmath214 cases , follows a poisson distribution : @xmath215\ ] ] supposing that we analyze influenza data documenting the generations of cases from 0 to @xmath88 in which we observed geometric growth , the likelihood of estimating @xmath2 is proportional to @xmath216 here we apply the above model to the spanish influenza data in zurich ( figure [ hn_fig1 ] ) .
assuming that the generation time of length @xmath116 , @xmath146 , is given by the following delta function with the mean length 3 days , @xmath217 then the observed series of data can be grouped by generation ( @xmath187 , @xmath188 , @xmath218 , ... ):
@xmath219 since we assumed exponential growth during the initial 16 days in the previous section , here we similarly assume a geometric increase up to the 6th generation . applying ( [ eqn_hn210 ] ) to the above data , maximum likelihood estimate of @xmath2 ( and the corresponding 95 percent confidence intervals ) is 1.51 ( 1.24 , 1.81 ) .
the model is simple enough to estimate @xmath2 , and indeed , a slight extension of the discrete - time branching process has been employed to estimate @xmath2 as well as the proportion of undiagnosed cases in the analysis of sars outbreak data @xcite .
+ it should be noted that the discrete - time branching process assumes homogeneous pattern of spread .
a technical issue has arisen on this subject during the sars outbreak . usually , we observe some cases who produce an extraordinary number of secondary cases compared with other infected individuals , which are referred to as * superspreaders*. because of this , observed offspring distributions for directly transmitted diseases tend to be extremely skewed to the right .
empirically , it has been suggested that poisson offspring distribution is sometimes insufficient to highlight the presence of superspreaders in epidemic modeling @xcite .
for example , if non - zero discrete distribution of secondary cases follows a geometric distribution with mean @xmath2 , the pgf is given by a geometric distribution with mean @xmath2 : @xmath220 moreover , if the offspring distribution follows gamma distribution with mean @xmath2 and dispersion parameter @xmath48 , the pgf @xmath199 follows negative binomial distribution @xcite : @xmath221 we still do not know if pandemic influenza is also the case to warrant the skewed offspring distributions . to explicitly test
if superspreading events frequently exist in influenza transmission , it is necessary to accumulate contact tracing data for this difficult disease , the cases of which often show flu - like symptoms only ( as discussed in section 1 ) . in addition
, it should be noted that we can not attribute the skewed offspring distribution to the underlying contact network only . to date , there are two major reasons which can generate superspreaders : ( i ) those who experience very frequent contacts ( social superspreader ) or ( ii ) those who are suffering from high pathogen loads or those who can scatter the pathogen through the air such as the use of nebuliser in hospitals ( biological superspreader ) . from the offspring distribution
only , we can not distinguish these two mechanisms . with regard to the estimation of @xmath2 using final size
, we briefly discuss another method based on a stochastic process . as we discussed above , let @xmath4 , @xmath28 and @xmath114 be the numbers of susceptible , infectious and recovered individuals at time @xmath6 , respectively .
further , let @xmath29 and @xmath35 denote the transmission rate and the mean duration of the infectious period , respectively .
supposing that @xmath222 , the number of individuals who experienced infection between time 0 and time @xmath6 , is given by @xmath223 , the two processes @xmath222 and @xmath114 are increasing counting processes where the general epidemic is explained by : @xmath224 where @xmath225 denotes the @xmath226-algebra generated by the history of the epidemic @xmath227 and @xmath228 ( where @xmath88 is the size of the susceptible population at time 0 ) .
the latter is equivalent to assuming density - independent transmission ( _ i.e. _ also referred to as _ true mass action _ or frequency dependent assumption @xcite ) .
based on equation ( [ eqn_hn215 ] ) , two zero - mean martingales @xcite are defined by : @xmath229 from the martingale theory @xcite , a zero - mean martingale is given by @xmath230 thus , the estimator @xmath231 is given by @xmath232}{u(t ) } \\ & = & \dfrac{-ln(1-\tilde{p})}{u(t ) } \end{array}\ ] ] where @xmath233 is the observed final size ( @xmath234 ) at the end of the epidemic at time @xmath235 .
furthermore , the variance of the zero - martingale is given by @xmath236 from the martingale central limit theorem @xcite , the estimator @xmath237 is approximately normally distributed in a major outbreak in a large community .
the standard error is then consistently estimated by : @xmath238^{\dfrac{1}{2}}}{u(t ) } \\ & = & \dfrac{\left[\dfrac{n}{s(0)+\dfrac{1}{2 } } + \dfrac{n}{s(0)+\dfrac{1}{2}}-\hat{\theta}^2r(t ) \right]^{\dfrac{1}{2}}}{u(t ) } \end{array}\ ] ] consequently , the estimator and standard error of @xmath2 are given by : @xmath239 more detailed mathematical descriptions can be found elsewhere @xcite@xcite@xcite .
+ here we show a numerical example .
let us consider a large epidemic of equine influenza ( _ i.e. _ influenza in horses ) as our case study .
in 1971 , a nationwide epidemic of equine-2 influenza a ( h3n8 ) was observed in japan @xcite .
for example , in niigata racecourse , 580 influenza cases were diagnosed with influenza among a total of 640 susceptible horses .
the final size @xmath18 is thus 90.6 percent ( 95 percent ci : 88.4 , 92.9 ) . from this data ,
we calculate @xmath2 and its uncertainty bounds .
+ using @xmath240 and total number of infected @xmath241 in equation ( [ eqn_hn218 ] ) , @xmath242 is estimated as 0.00408 .
therefore , the estimate of @xmath243 is given by equation ( [ eqn_hn221 ] ) .
moreover , from equation ( [ eqn_hn220 ] ) where @xmath244 and @xmath245 ( we assume one case was already infected at time @xmath246 ) , we obtain @xmath247 . here , @xmath242 is assumed to follow normal distribution .
therefore , the 95 percent confidence interval for @xmath2 is given as @xmath248=[2.44,2.76]$ ] .
+ when applying the final size equation , it should be remembered that ( i ) we assume all individuals are initially susceptible ( in the above described model ) and ( ii ) we assume @xmath29 and @xmath30 are independent of time ( _ i.e. _ constant ) , and thus , that any extrinsic factors should not have influenced the course of the observed epidemic . in the above described models , we always assumed that the pattern of influenza transmission is homogeneous , which is clearly unrealistic to capture the transmission dynamics of influenza . since the last century
, it has already been understood that the transmission dynamics are not sufficiently modeled by assuming homogeneous mixing . however , because more detailed data are lacking ( _ e.g. _ epidemic records of pandemic influenza with time , age and space ) , what we could offer has been mainly to extract the intrinsic growth rate from the initial exponential growth , and estimate @xmath2 using the estimator based on a model with the homogeneous mixing assumption .
+ one line of addressing heterogeneous patterns of transmission using the observed data is separating household transmission from community transmission . in other words
, it is of practical importance to distinguish between individual and group @xmath2 @xcite .
from the beginning of explicit modeling of influenza @xcite@xcite , a method to separately estimate the transmission parameters has been proposed , which has been partly extended in a recent study @xcite or applied to further old data of pandemic influenza @xcite .
indeed , an important aspect of this issue was highlighted in a recent study which compared estimates of @xmath2 between those having casual and close contacts @xcite . to estimate key parameters of household and community transmissions of influenza , or to simulate realistic patterns of influenza spread , such a consideration is fruitful .
+ mathematically elaborating this concept , there are several publications which proposed the basis of analyzing household transmission data employing stochastic models @xcite@xcite@xcite .
moreover , a rigorous study has been made to estimate parameters determining the intrinsic dynamics ( _ e.g. _ infectious period ) using household transmission data with time @xcite .
+ future challenges on the estimation of @xmath2 include the application of such theories to the observed data with some extension .
for example , as we discussed above , knowing the generation time would be crucial to elucidate a robust estimate of @xmath2 @xcite@xcite@xcite@xcite . however , we do not know if the generation time varies between close and casual contacts ; this should be the case , because , as long as the generation time is given by covolution of latent and infectious periods , close contact should lead to shorter generation time than casual contact . in future studies , influenza models may better to highlight the increasing importance of considering household transmission to estimate the transmission potential using the temporal distribution of infection events .
except for our approach in section 3 , mathematical arguments given in this paper are not particularly special for influenza . in other words ,
we modelers have employed similarly structured models which describe the population dynamics of other directly transmitted diseases , and such models are applicable not only for influenza but also for many viral diseases including measles , smallpox , chickenpox , rubella and so on @xcite . however , influenza has many different epidemiologic characteristics compared to other childhood viral diseases .
for instance , following the previous efforts in influenza epidemiology @xcite@xcite@xcite and modeling @xcite@xcite , we should at least note the following : 1 .
detailed mechanisms of immunity have yet to be clarified .
since influenza virus has an wide antigenic diversity ( _ i.e. _ unlike other childhood viral diseases , antigenic stimulation is not monoclonal ) , this complicates our understanding in the fraction of immune individuals , cross - protection mechanisms and evolutionary dynamics @xcite@xcite@xcite@xcite .
flu - like symptoms are too common , and thus , we cannnot explicitly distinguish influenza from other common viral infections without expensive laboratory tests for each case .
because of this character , it is difficult to effectively implement usual public health measures ( _ e.g. _ contact tracing and isolation ) .
3 . although explicit estimates are limited @xcite@xcite , a certain fraction of infected individuals does not exhibit any symptoms ( following infection ) .
this complicates not only the eradication @xcite but also epidemiologic evaluations of vaccines and therapeutics @xcite .
4 . looking into the details of the intrinsic dynamics @xcite@xcite , it appears recently that the generation time and infectious period are much shorter than what were believed previously .
therefore , despite the relatively small @xmath2 estimate , the turn - over of a transmission cycle ( _ i.e. _ speed of growth ) is rather quick .
the incubation period of spanish influenza is as short as 1.5 days @xcite , complicating the implementation of quarantine measures @xcite .
thus , depending on the characteristics of observed data ( and the specific purpose of modeling ) , we have to highlight these factors referring to the best available evidence .
this is one of the most challenging issues in designing public health interventions against a potential future pandemic .
in addition to the above described issue , we , of course , must remember what the reported data is .
in many studies , the compartment @xmath28 or relevant class of infectious individuals of the sir ( or seir ) model was fitted to the observed data .
indeed , in the majority of previous classic studies , @xmath15 ( _ i.e. _ removed class ; denoted by @xmath114 in our discussion ) of kermack and mckendrick model was fitted to the data , assuming that the removed class highlights observed data as the reported cases no longer produce secondary cases .
however , the observed epidemiologic data is actually neither @xmath28 nor @xmath114 . always , what we get as the temporal distribution reflects _ case onset _ or _ deaths _ with time which is mostly accompanied by some reporting delay . + we believe this is one of the most challenging issues in epidemic modeling . except for rare examples in sexually transmitted diseases ,
infection event is not directly observable , and thus , we have to maximize the utility of reported ( observed ) data , explicitly understanding what the data represents .
+ in this case , back - calculation of the infection events is called for .
let @xmath175 denote the number of onsets at time @xmath6 , this should be modeled by using incidence @xmath126 and the density of the incubation period of length @xmath116 , @xmath249 : @xmath250 further , supposing that @xmath251 is the number of reported cases at time @xmath6 and the density of reporting delay of length @xmath226 is @xmath252 , observed data is modeled as : @xmath253 that is , only by using the observed data @xmath251 and known information of the reporting delay @xmath254 and incubation period distributions @xmath249 , we can translate the observed data into infection process @xmath126 . + in some cases , only death data with time ,
@xmath255 , is available @xcite .
similarly , this can be modeled using the backcalculation .
let @xmath75 denote the case fatality of influenza which is reasonably assumed time - independent , and further let @xmath256 be the relative frequency of time from onset to death , @xmath255 is given by : @xmath257 even when using _ onset data with delay _ or _ death _ data , it should be noted that the intrinsic growth rate is identical to that estimated from the infection event distribution . assuming that the incidence @xmath126 exhibits exponential growth during the initial phase of an epidemic , _
i.e. _ , @xmath139 , equations ( [ eqn_hn301 ] ) and ( [ eqn_hn303 ] ) can be rewritten as @xmath258 and @xmath259 thus , the growth terms @xmath260 ( _ i.e. _ which depends on time ) of @xmath251 and @xmath255 are still identical to that of incidence @xmath126 .
in other words , mathematically the equations ( [ eqn_hn304 ] ) and ( [ eqn_hn305 ] ) could be a justification to extract an estimate of the intrinsic growth rate from cases with reporting delay or deaths with time .
however , we should always remember that the infection - age distribution is not stable during the initial phase , and moreover , this method can not address individual variation in the secondary transmissions ( _ e.g. _ superspreaders , as we discussed in section 5 ) . in this way , it s not an easy task to clarify the infection events with time .
a similar application of the convolution equation has been intensively studied in modeling hiv / aids .
since aids has a long incubation period , and because aids diagnosis is certainly reported in the surveillance data ( at least , in industrialized countries ) , backcalculation of the number of hiv infections with time using the nubmer of aids diagnoses and the incubation period distribution has been an issue to capture the whole epidemiologic picture of hiv / aids @xcite@xcite@xcite@xcite . in the current modeling practice
using the temporal distribution of onset events , we are now faced with a need to apply this technique to diseases with much shorter incubation periods .
+ now , let s look back at a method to estimate @xmath15 , which was proposed by wallinga and teunis @xcite . whereas the method has a background of mathematical reasoning ( as shown in ( [ eqn_hn28 ] ) , section 4.2 ) , the estimator was derived implicitly assuming that _ observed data exactly reflects infection events_. if asymptomatic infection and transmission are rare , this assumption might be justifiable as the lengths of the serial interval and generation time become almost identical .
however , as long as we can not ignore asymptomatic transmissions , which is particularly the case for influenza , the assumption @xmath261 might be problematic @xcite .
+ since @xmath15 of this method was given by summing up the probability of causing secondary transmissions by an onset case _ at the onset time _ of this case @xmath6 , we should rewrite the assumption using a modified _ onset - based _ renewal equation as follows @xcite : @xmath262 for simplicity , we ignore reporting delay in the observed data , roughly assuming that the observed data reflects @xmath175 . translating equation ( [ eqn_hn306 ] ) in words , it is implicitly assumed that _
secondary transmission happens exactly at the time of onset _ , and based on this assumption , @xmath15 in the right hand side of ( [ eqn_hn306 ] ) can be backcalculated .
+ to understand the assumptions behind the above equation , let us assume that incidence @xmath126 is given by @xmath263 where @xmath264 is the transmission rate at * disease - age * @xmath226 ( _ i.e. _ the time since onset of infection @xcite ) and @xmath265 is the survivorship of cases following onset .
it should be noted that equation ( [ eqn_hn307 ] ) ignores secondary transmissions before onset of illness .
as we discussed above , @xmath175 is given by @xmath126 and the incubation period distribution @xmath249 , @xmath266 replacing @xmath175 in the right hand side of ( [ eqn_hn307 ] ) by ( [ eqn_hn308 ] ) , we get @xmath267 where @xmath268 represents infection - age ( _ i.e. _ time since infection ) , and @xmath269 is given by @xmath270 which represents generation time distribution . from equations ( [ eqn_hn309 ] ) and ( [ eqn_hn310 ] ) , we can find that @xmath15 is given by @xmath271 equation ( [ eqn_hn311 ] ) can be further reduced to @xmath272 which represents kermack and mckendrick s assumption .
replacing @xmath126 in the right hand side of ( [ eqn_hn308 ] ) by ( [ eqn_hn307 ] ) , we get @xmath273 where @xmath274 denotes the serial interval distribution of calender time @xmath6 and disease - age @xmath226 : @xmath275 equation ( [ eqn_hn313 ] ) is difficult to solve as it includes @xmath276 in the right hand side .
however , in the special case , _
e.g. _ , let s say when we can assume @xmath277 ( where @xmath48 is constant and @xmath278 is delta function ) , @xmath279 inserting ( [ eqn_hn314 ] ) back to ( [ eqn_hn312 ] ) , @xmath280 which is _ onset - based _ renewal equation which was presented in ( [ eqn_hn306 ] ) .
what to be learnt from ( [ eqn_hn315 ] ) is , the assumption that _ secondary transmission happens immediately after onset _ suggests that the _ incubation period distribution is identical to the serial interval distribution _ as shown above , which is a bit funny conclusion . maximizing the utility of observed data has still remained an open question .
+ in addition to modeling the temporal distribution , explicit modeling of asymptomatic infection is also called for @xcite .
provided that there are so many asymptomatic transmissions which are not in the negligible order , we need to shift our concept of transmissibility ; _
e.g. _ , rather than @xmath2 , a threshold quantity of symptomatic infection is required .
in such a case , application of type - reproduction number @xmath235 might be useful @xcite@xcite , and it has already been put into practice @xcite .
in this review , we focused on the use of the temporal distribution of influenza to estimate @xmath2 ( or @xmath15 ) and the relevant key parameters .
it must be remembered that our arguments , almost necessarily , employed homogeneous mixing assumption , as we can not extract information on heterogeneous patterns of infection from a single stream of temporal data alone .
presently , more information ( _ e.g. _ at least , spatio - temporal distribution ) is becoming available for influenza . in this section , we briefly sketch what can be ( and should be ) done in the future to quantify the transmission dynamics of pandemic influenza .
it s not a new issue that heterogeneous patterns of transmission could even destroy the mean field theory in infectious diseases .
for example , in a pioneering study of gonorrhea transmission dynamics by hethcote and yorke @xcite , an importance of contact heterogeneity was sufficiently highlighted . since a simple model assuming
homogeneous mixing did not reflect the patterns of gonorrhea transmission in the united states , hethcote and yorke divided the population in question into two ; those who are sexually very active and not , the former of which was referred to as * core * group .
compared with the temporal distribution of infection given by a model with homogeneous mixing assumption , the simple heterogeneous model with a core group revealed much quicker increase in epidemic size , showing rather different trajectory of an epidemic .
given that the variance of sexual partnership is extremely large ( _ i.e. _ if the distribution of the frequency of sexual intercourse is extremely skewed to the right with a very long right tail ) , the estimate of @xmath2 is shown to become considerably high .
the finding supports a vulnerability of our society to the invasion of sexually transmitted diseases . following this pioneering study
, considerable efforts have been made to approximately model the heterogeneous patterns of transmission using extended mean field equations @xcite@xcite@xcite@xcite . + in addition to such an approximation of heterogeneous transmission , recent progress in epidemic modeling with explicit contact network structures suggests that variance of the contact frequency plays a key role in determining the threshold quantity , and in some special cases , the concept of threshold phenomena could be confused @xcite@xcite@xcite@xcite . in section 4
, we defined the force of infection as @xmath281 in deteministic models given by simple odes ( which ignores infection - age ) , @xmath118 is equivalent to @xmath282 .
these are what classical mean field models suggest .
+ let us account for an epidemic on networks , whose node - connectivity distribution ( _ i.e. _ the distribution of probabilities that nodes have exactly @xmath48 neighbors ) follows some explicit distribution @xmath283 .
the force of infection @xmath284 , which yields @xmath212 , in a static contact network is given by @xmath285 here @xmath286 denotes the average connectivity of the nodes . assuming that @xmath283 follows a power law of the form @xmath287 ( where @xmath288 is constant ) , @xmath289 given that @xmath290 , @xmath291 , and in such a case , @xmath2 even becomes infinite .
this implies that the disease spread will continue for any mean estimate of @xmath2 .
such a network structure is referred to as * scale free * @xcite , complicating disease control efforts in public health @xcite@xcite .
the importance of the network structure would also be highlighted for @xmath292 .
+ for sexually transmitted infections , contact frequency is countable ( unlike airborne infection or transmission through droplets ) , and @xmath293 is estimated to be around 3 or a little larger @xcite .
following such a finding , many non - sexual directly transmitted diseases are also modeled in the present day assuming the scale - free network @xcite .
however , it should be noted that the pattern of contact does not necessarily follow scale - free for all directly transmitted diseases .
indeed , there is no empirical evidence which suggests that the contact structure of any droplet infections follows the power law ( _ i.e. _ we do not know if the above described contact heterogeneity is the case for diseases except for sexually transmitted diseases ) .
a typical example of confusion is seen in the superspreading events during the 2002 - 03 sars epidemics @xcite , where we can not explicitly attribute the phenomena either to contact network or biological factors ( as long as _ contact _ and infection event are not directly observable ) . we still do not know how we should account for the distribution @xmath283 for influenza and other viral respiratory diseases ( _ i.e. _ power law or not ) which remains to be clarified for each disease in future research .
methodoligical developments have been made to account for the network heterogeneity with data @xcite .
an approximate approach to address this issue is highlighted particularly in spatio - temporal modeling , an excellent account of which is reviewed by matt keeling @xcite . +
even though it s difficult to quantify the transmission dynamics with an explicit contact network with time , there are useful analytical approximations to capture the dynamics of influenza ( and other respiratory transmitted viral diseases ) and estimate the transmission potential . for example
, the force of infection with a power law approximation is reasonably given by : @xmath294 in ( [ eqn_hn404 ] ) , @xmath70 and @xmath295 characterize the epidemic dynamics ; _
e.g. _ initial growth ( _ i.e _ if @xmath70 is less than 0 , the modified form acts to dampen the exponential growth of incidence ) and endemic equilibrium ( _ i.e. _ when @xmath295 is greater than 0 , density - dependent damping is increased ) . a model of this type
was actually validated with measles data in england and wales , comparing the prediction with that of employing the mass action principle @xcite .
+ another approximation might be a pair - wise model @xcite , which can explicitly account for the correlation between connected pairs .
the model reasonably permits deriving the force of infection @xmath296 using the number of various connected paris , which implies wide applicability to the epidemiologic data of sexually transmitted infections . incorporating spatial heterogeneity in an approximate manner would shed light on further quantifications @xcite@xcite , and thus , simple and reasonably tractable models which permit spatio - temporal modeling of influenza are expected ( _ e.g. _ @xcite ) . summarizing the above discussions , we have presented modeling approaches that can quantify the transmission potential of pandemic influenza .
as we have shown , temporal case distributions have been analyzed in many instances , and previous efforts have come close to maximize the utility of temporal distributions ( _ i.e. _ epidemic curve ) .
however , at the same time , we have also learned that we can extract almost the intrinsic growth rate alone from a single time - evolution data .
accordingly , we are now faced with a need to clarify heterogeneous patterns of transmission and more detailed intrinsic dynamics of influenza @xcite@xcite@xcite . with regard to the latter , primitive epidemiologic questions (
_ e.g. _ probability of clinical attack given infection ) remain to be answered for spanish , asian and hong kong influenza .
let s summarize what we need to clarify theoretically about pandemic influenza in list : 1
. acquired immunity 2 .
evolutionary dynamics 3 .
multi - host species transmission 4 .
asymptomatic transmission 5 . attack rate ( _ i.e. _ pr(onset@xmath297infection ) ) 6 .
case fatality ( _ i.e. _ pr(death@xmath297onset ) ) 7 .
generation time and serial interval 8 .
latent , incubation , infectious and symptomatic periods with further data 9 .
transmission potential with time , space and antigenic types 10 .
transmission potential with time and age these issues highlight an importance to quantify the transmission of influenza using not only cases ( _ i.e. _ those followed onset of symptoms ) but also some hint suggesting the infection event .
for example , majority of the above listed issues could be reasonably addressed by implementing serological surveys ( _ e.g. _ antibody titers of individuals and , preferably , time - delay delay distribution from infection to seroconversion ) .
since the proportion of those who do not experience symptomatic infection ( _ i.e. _ probability of asymptomatic infection ) is not small for influenza @xcite@xcite , case records can tell us little to address the above mentioned issues , and thus , historical data of spanish influenza may hardly offer further information . by maximizing the utility of observed data
, we have to clarify the dynamics of influenza further , and identify key information which characterize the specific mechanisms of spread .
ferguson , d.a.t .
cummings , s. cauchemez , c. fraser , s. riley , a. meeyai , s. iamsirithaworn and d.s .
burke , strategies for containing an emerging influenza pandemic in southeast asia .
_ nature _ * 437 * ( 2005 ) 209 - 214 .
murray , a.d .
lopez , b. chin , d. feehan and k.h .
hill , estimation of potential global pandemic influenza mortality on the basis of vital registry data from the 1918 - 20 pandemic : a quantitative analysis .
_ lancet _ * 368 * ( 2006 ) 2211 - 2218 .
h. markel , h.b .
lipman , j.a .
navarro , a. sloan , j.r .
michalsen , a.m. stern and m.s .
cetron , nonpharmaceutical interventions implemented by us cities during the 1918 - 1919 influenza pandemic .
_ jama _ * 298 * ( 2007 ) 644 - 654 .
h. nishiura , k. dietz and m. eichner , the earliest notes on the reproduction number in relation to herd immunity : theophil lotz and smallpox vaccination .
_ journal of theoretical biology _
* 241 * ( 2006 ) 964 - 967 .
o. diekmann , j.a.p .
heesterbeek and j.a.j .
metz , on the definition and the computation of the basic reproductive ratio @xmath2 in models for infectious diseases .
_ journal of mathematical biology _ * 35 * ( 1990 ) 503 - 522 . c. castillo - chavez , z. feng and w. huang , on the computation of @xmath2 and its role on global stability , in : mathematical approaches for emerging and reemerging infectious diseases : an introduction , i m a volume 125 ( springer - veralg , berlin , 2002 ) pp .
229 - 250 .
kendall , deterministic and stochastic epidemics in closed populations . in : third berkeley symposium on mathematical statistics and probability 4 ,
ed p. newman ( university of california press , new york , 1956 ) pp.149 - 165 .
m. lipsitch , t. cohen , b. cooper , j.m .
robins , s. ma , l. james , g. gopalakrishna , s.k .
chew , c.c .
tan , m.h .
samore , d. fisman and m. murray , transmission dynamics and control of severe acute respiratory syndrome .
_ science _ * 300 * ( 2003 ) 1966 - 1970 .
g. chowell , c.e .
ammon , n.w .
hengartner and j.m .
hyman , transmission dynamics of the great influenza pandemic of 1918 in geneva , switzerland : assessing the effects of hypothetical interventions
. _ journal of theoretical biology _
* 241 * ( 2006 ) 193 - 204 .
g. chowell , h. nishiura and l.m .
bettencourt , comparative estimation of the reproduction number for pandemic influenza from daily case notification data .
_ journal of the royal society interface _ * 4 * ( 2007 ) 155 - 166 .
g. chowell , c.e .
ammon , n.w .
hengartner and j.m .
hyman , estimating the reproduction number from the initial phase of the spanish flu pandemic waves in geneva , switzerland .
_ mathematical biosciences and engineering _ * 4 * ( 2007 ) 457 - 470 .
s. cauchemez , p.y .
boelle , g. thomas and a.j .
valleron , estimating in real time the efficacy of measures to control emerging communicable diseases . _ american journal of epidemiology _ * 164 * ( 2006 ) 591 - 597 .
bettencourt , r.m .
ribeiro , g. chowell , t. lant and c. castillo - chavez , towards real time epidemiology : data assimilation , modeling and anomaly detection of health surveillance data streams . in :
intelligence and security informatics : biosurveillance .
proceedings of the 2nd nsf workshop , biosurveillance , 2007 .
lecture notes in computer science .
eds f. zeng et al .
( springer - verlag , berlin , 2007 ) pp .
79 - 90 .
h. nishiura , m. schwehm , m. kakehashi and m. eichner , transmission potential of primary pneumonic plague : time inhomogeneous evaluation based on historical documents of the transmission network .
_ journal of epidemiology and community health _ * 60 * ( 2006 ) 640 - 645 .
d. schenzle , k. dietz and g.g .
frosner , antibody against hepatitis a in seven european countries .
ii . statistical analysis of cross - sectional surveys .
_ american journal of epidemiology _ * 110 * ( 1979 ) 70 - 76 . c. viboud , t. tam , d. fleming , a. handel , m.a . miller and l. simonsen , transmissibility and mortality impact of epidemic and pandemic influenza , with emphasis on the unusually deadly 1951 epidemic .
_ vaccine _ * 24 * ( 2006 ) 6701 - 6707 .
g. sertsou , n. wilson , m. baker , p. nelson and m.g .
roberts , key transmission parameters of an institutional outbreak during the 1918 influenza pandemic estimated by mathematical modelling .
_ theoretical biology and medical modelling _ * 3 * ( 2006 ) 38 .
v. andreasen , c. viboud and l. simonsen , epidemiologic characterization of the summer wave of the 1918 influenza pandemic in copenhagen : implications for pandemic control strategies .
_ journal of infectious diseases _ * * ( 2008 ) in press .
mathews , c.t .
mccaw , j. mcvernon , e.s .
mcbryde and j.m .
mccaw , a biological model for influenza transmission : pandemic planning implications of asymptomatic infection and immunity .
_ plos one _ * 2 * ( 2007 ) e1220 .
g. chowell , l.m.a .
bettencourt , n. johnson , w.j .
alonso and c. viboud , the 1918 - 1919 influenza pandemic in england and wales : spatial patterns in transmissibility and mortality impact _ proceedings of the royal society b _ * * ( 2008 ) in press .
k. dietz , mathematical models for transmission and control of malaria . in : malaria , principles and practice of malariology ,
wernsdorfer and i. mcgregor ( churchill livingstone , edinburgh , 1988 ) pp.1091 - 1133 .
kermack and a.g .
mckendrick , contributions to the mathematical theory of epidemics - i. _ proceedings of the royal society series a _ * 115 * ( 1927 ) 700 - 721 ( reprinted in _ bulletin of mathematical biology _ * 53 * ( 1991 ) 33 - 55 ) .
h. nishiura , t. kuratsuji , t. quy , n.c .
phi , v. van ban , l.e .
long , h. yanai , n. keicho , t. kirikae , t. sasazuki and r.m .
anderson , rapid awareness and transmission of severe acute respiratory syndrome in hanoi french hospital , vietnam .
_ american journal of tropical medicine and hygiene _ * 73 * ( 2005 ) 17 - 25 .
haydon , m. chase - topping , d.j .
shaw , l. matthews , j.k .
friar , j. wilesmith and m.e .
woolhouse , the construction and analysis of epidemic trees with reference to the 2001 uk foot - and - mouth outbreak .
_ proceedings of the royal society of london series b _ * 270 * ( 2003 ) 121 - 127 .
s. cauchemez , p.y .
boelle , c.a .
donnelly , n.m .
ferguson , g. thomas , g.m .
leung , a.j .
hedley , r.m .
anderson and a.j .
real - time estimates in early detection of sars . _
emerging infectious diseases _ * 12 * ( 2006 ) 110 - 113 .
metz , the epidemic in a closed population with all susceptibles equally vulnerable ; some results for large susceptible populations and small initial infections .
_ acta biotheoretica _ * 27 * ( 1978 ) 75 - 123 .
separate roles of the latent and infectious periods in shaping the relation between the basic reproduction number and the intrinsic growth rate of infectious disease outbreaks .
_ journal of theoretical biology _ * * ( 2008 ) in press .
a. imahorn , epidemiologische beobachtungen ueber die grippe - epidemie 1918 i m oberwallis .
inaugural - dissertation zur erlangung der doktorwuerde der medizinischen fakultaet der universitaet zuerich ( universitaet zuerich , zurich , 2000 ) ( in german ) .
h. nishiura , epidemiology of a primary pneumonic plague in kantoshu , manchuria , from 1910 to 1911 : statistical analysis of individual records collected by the japanese empire .
_ international journal of epidemiology _ * 35 * ( 2006 ) 1059 - 1065 .
cowling , l.m . ho and g.m .
leung , effectiveness of control measures during the sars epidemic in beijing : a comparison of the rt curve and the epidemic curve .
_ epidemiology and infection _ * * , ( 2007 ) in press .
marques , o.p .
forattini and e. massad , the basic reproduction number for dengue fever in sao paulo state , brazil : 1990 - 1991 epidemic .
_ transactions of the royal society of tropical medicine and hygiene _ * 88 * ( 1994 ) 58 - 59 .
galvani , x. lei and n.p .
jewell , severe acute respiratory syndrome : temporal stability and geographic variation in case - fatality rates and doubling times .
_ emerging infectious diseases _ * 9 * ( 2003 ) 991 - 994 .
de jong , o. diekmann and j.a.p .
heesterbeek , how does transmission of infection depend on population size ? in : epidemic models : their structure and relation to data .
d. mollison ( cambridge university press , cambridge , 1995 ) pp.84 - 94 .
h. nishiura and g. chowell , household and community transmission of the asian influenza a ( h2n2 ) and influenza b viruses in 1957 and 1961 .
_ southeast asian journal of tropical medicine and public health _ * 38 * ( 2007 ) in press .
s. cauchemez , f. carrat , c. viboud , a.j .
valleron and p.y .
boelle , a bayesian mcmc approach to study transmission of influenza : application to household longitudinal data . _
statistics in medicine _ * 23 * ( 2004 ) 3469 - 3487 .
nelson , l. simonsen , c. viboud , m.a .
miller , j. taylor , k.s .
george , s.b .
griesemer , e. ghedin , n.a .
sengamalay , d.j .
spiro , i. volkov , b.t .
grenfell , d.j .
lipman , j.k .
taubenberger and e.c .
holmes , stochastic processes are key determinants of short - term evolution in influenza a virus
. _ plos pathogens _ * 2 * ( 2006 ) e125 .
t. colton , t. johnson and d. machin d ( eds ) .
proceedings of the conference on quantitative methods for studying aids , held in blaubeuren , germany , june 14 - 18 , 1993 .
_ statistics in medicine _ * 13 * ( 1994 ) 1899 - 2188 .
h. inaba and h. nishiura , on the dynamical system formulation of the type reproduction number for infectious diseases and its application to the asymptomatic transmission model . _ mathematical biosciences _ * * ( 2007 ) submitted .
woolhouse , c. dye , j.f .
etard , t. smith , j.d .
charlwood , g.p .
garnett , p. hagan , j.l .
hii , p.d .
ndhlovu , r.j .
quinnell , c.h .
watts , s.k .
chandiwana and r.m .
anderson , heterogeneities in the transmission of infectious agents : implications for the design of control programs .
_ proceedings of the natinal academy of science u s a _ * 94 * ( 1997 ) 338 - 342 .
duerr , m. schwehm , c.c .
leary , s.j
. de vlas and m. eichner , epidemic size and probability in populations with heterogeneous infectivity and susceptibility .
_ epidemiology and infection _ * 135 * ( 2007 ) 1124 - 1132 .
keeling , s.p .
brooks and c.a .
gilligan , using conservation of pattern to estimate spatial parameters from a single snapshot . _ proceedigns of the national academy of science u s a _ * 101 * ( 2004 ) 9155 - 9160 .
duerr , s.o .
brockmann , i. piechotowski , m. schwehm and m. eichner , influenza pandemic intervention planning using influsim : pharmaceutical and non- pharmaceutical interventions .
_ bmc infectious diseases _ * 7 * ( 2007 ) 76 .
+ @xmath298sensitivity of the @xmath16 estimates to different assumptions for the serial interval was examined ; @xmath299three pandemic waves were simultaneously fitted ; @xmath300the epidemic was observed in a community with closed contact ( i.e. military camp ) .
[ tablewaves ] | this article reviews quantitative methods to estimate the basic reproduction number of pandemic influenza , a key threshold quantity to help determine the intensity of interventions required to control the disease .
although it is difficult to assess the transmission potential of a probable future pandemic , historical epidemiologic data is readily available from previous pandemics , and as a reference quantity for future pandemic planning , mathematical and statistical analyses of historical data are crucial . in particular , because many historical records tend to document only the temporal distribution of cases or deaths ( i.e. epidemic curve ) , our review focuses on methods to maximize the utility of time - evolution data and to clarify the detailed mechanisms of the spread of influenza . + first , we highlight structured epidemic models and their parameter estimation method which can quantify the detailed disease dynamics including those we can not observe directly .
duration - structured epidemic systems are subsequently presented , offering firm understanding of the definition of the basic and effective reproduction numbers . when the initial growth phase of an epidemic is investigated , the distribution of the generation time is key statistical information to appropriately estimate the transmission potential using the intrinsic growth rate .
applications of stochastic processes are also highlighted to estimate the transmission potential using similar data .
critically important characteristics of influenza data are subsequently summarized , followed by our conclusions to suggest potential future methodological improvements .
* pacs classifications : * viral diseases ( 87.19.xd ) ; population dynamics ( 87.23.cc ) ; stochastic models in biological physics ( 87.10.mn ) * keywords : * influenza ; pandemic ; epidemiology ; basic reproduction number ; model . |
all graphs considered are loopless and finite .
a _ cycle _ is a connected @xmath8-regular graph .
a _ quadrangulation _ on an oriented surface ( i.e a sphere with @xmath9 attached handles ) is a @xmath8-edge connected graph embedded on the surface such that every face boundary is a cycle of length four .
a vertex - coloring is said to be _ proper _ if any two adjacent vertices receive different colors . for standard graph - theoretic terms
not defined here , we refer to @xcite .
let @xmath0 be a quadrangulation on an orientable surface with a proper vertex - coloring @xmath10 .
then for each face of @xmath0 , two , three or four different colors appear on its boundary .
we consider only those faces of @xmath0 which have all four colors on the boundary and call them _ rainbow - faces_. a rainbow - face @xmath3 is said to be a _ @xmath4-face _
, where @xmath11 , if the four distinct colors @xmath12 appear on the boundary of @xmath3 in the clockwise order , where the clockwise orientation can be fixed in all faces of @xmath0 simultaneously , since the surface is orientable .
a @xmath4-face is also an @xmath13-face , where @xmath14 .
it was shown in @xcite that the number of @xmath4-faces and the number of @xmath7-faces have the same parity if @xmath0 is a quadrangulation on the sphere .
then it was proven in @xcite that these two numbers are even equal . in this paper , we show that this holds for all orientable surfaces by a very simple proof .
( for example , figure 1 shows a quadrangulation on the torus , where we identify the top and the bottom , and the right and left sides , respectively , in the rectangle . ) the idea for the proof comes from the paper @xcite , which points out the surprising phenomenon that a certain kind of quadrangulations on a nonorientable surface has no proper vertex-@xmath15-coloring .
[ main ] let @xmath0 be a quadrangulation on an orientable surface with a proper vertex-@xmath2-coloring @xmath10 .
then the number of @xmath4-faces in @xmath0 equals the number of @xmath7-faces , where @xmath11 .
* proof.*we may assume that @xmath16 otherwise we permute the four colors in a suitable manner .
we orient all edges of @xmath0 by giving a direction to each edge @xmath17 from @xmath13 to @xmath18 if and only if @xmath19 .
let @xmath3 be a face of @xmath0 and let @xmath20 ( resp . ,
@xmath21 ) denote the set of all boundary edges of @xmath3 which are oriented clockwise ( resp . , anti - clockwise ) in @xmath3 .
set @xmath22 .
let @xmath23 denote the set of all faces of @xmath0 and let @xmath24 denote the set of all @xmath25- and @xmath26-faces in @xmath0 .
then , @xmath27 since each @xmath28 is contained in precisely two faces , say @xmath3 and @xmath29 , sharing @xmath30 , in which if @xmath30 has the clockwise orientation in @xmath3 , then @xmath30 has the anti - clockwise orientation in @xmath29 , and vice versa . on the other hand
, we see that @xmath31 if and only if @xmath3 is neither a @xmath25-face nor a @xmath26-face in @xmath0 . therefore ,
since @xmath32 , we have @xmath33 observe that @xmath34 if @xmath3 is a @xmath25-face and @xmath29 is a @xmath26-face , and hence the number of @xmath25-faces equals the number of @xmath26-faces . @xmath35
[ ] let @xmath0 be a quadrangulation of an orientable surface .
then @xmath0 admits no proper vertex @xmath2-coloring of @xmath0 such that the number of rainbow faces is odd .
this work was supported by the austrian science fund ( fwf projects : p 20543 , p 26686 ) .
a. hoffmann - ostenhof .
thesis `` nowhere - zero flows and structures in cubic graphs '' ( 2011 ) .
quadrangulations and 5-critical graphs on the projective plane .
topics in discrete mathematics , algorithms and combinatorics 26 , springer 2006 , 565580 . | let @xmath0 be a quadrangulation on an orientable surface and let @xmath1 be a proper vertex-@xmath2-coloring of @xmath0 .
a face @xmath3 of @xmath0 is said to be a _ rainbow - face _ if all four distinct colors appear on its boundary .
a _
@xmath4-face _ in @xmath0 is a rainbow face with colors @xmath5 , @xmath6 on the boundary in clockwise order .
we show that the number of @xmath4-faces in @xmath0 equals the number of @xmath7-faces .
this implies in particular that the number of rainbow - faces of @xmath0 is even .
* keywords : * quadrangulation , coloring , surface |
* notes : * we have used the standard routine from the matlab optimization toolbox for the solution of the optimization problem @xcite . [ s1 ] fig .
s1 : * phase - space transformation . * ( a ) oscillon @xmath2 from ref.@xcite with twenty - six reference points @xmath21 , ( b ) original limit cycle of the hindmarsch - rose model with reference points @xmath22 ; ( c ) limit cycle after a frame transformation with superimposed reference points @xmath21 .
* electronic chaos prehistory : * * @xmath23 first encounter with electronic chaos : `` often an irregular noise is heard ... [ that ] strongly reminds one of the tunes of a bagpipe '' [ b. van der pole and j. van der mark , nature * 120 * , 363 ( 1927 ) ] ; * @xcite ; + * first encounter with dynamical chaos in numerical studies ( ? ) : two - disc dynamo model by tsuneji rikitake : `` ... the system in question performs an extremely complicated oscillation . '' [ t. rikitake , proc .
cambridge philos .
soc . * 54 * , 89 ( 1958 ) ] ( see also @xcite ) ; + * @xmath23 experiments with glow - lamp ring circuits : `` larger rings [ of glow lamps ] , on the other hand , perform erratically and are always hard to adjust for a specific firing order . ''
[ r. l. ives , electronics * 31 * , 108 ( 1958 ) ] ; + * * e. lorenz found his attractor : the rise of chaology * @xcite ; * @xmath23 chaotic signal from an overloaded traveling wave tube amplifier : `` high power noise source employing a feedback path around a traveling wave tube '' [ c. a. reis & j. e , zellens , pat .
usa number : 3,176,655 filed nov .
16 , 1961 ] ; * while experimenting with vacuum tube circuits , yoshishuke ueda noticed `` randomly transitional phenomena '' ( though his finding was not published until 1970 @xcite ) .
* @xmath23 an overloaded amplifier with a feedback element generates aperiodic broad band oscillations [ e. a. koptyrev & l. e. pliss , radiotekhnika i elektronika * 10 * , 1828 ( 1965 ) ( in russian ) ] ; * .................... * experiments with a self - excited circuit : `` ... the results of a theoretical , numerical and experimental investigation of one of the simplest self - excited noise generators . ...
the statistical properties of the signal are determined ... by the intrinsic dynamics of the system rather than by .. the noise '' [ s. v. kiyashko , a. s. pikovskii and m. i. rabinovich , radio eng . and electron
* 25 * , 74 ( 1980 ) ] ; + * * chua s circuit * : incarnation of chaotic attractor regimes with a specially designed circuit @xcite .
@xmath23 all these researchers did not attribute the observed phenomena to the underlying chaotic dynamics of the employed systems .
they interpreted chaotic signals as the result of amplification of inevitable " internal fluctuations .
strickly speaking , we should not exclude the possibility that , by feeding the vertical and horizontal control inputs of an oscilloscope with specially devised quasiperiodic signals , one could get something similar to the oscillons shown on figs .
a , b. j. briggs , _ fractals : the patterns of chaos : discovering a new aesthetic of art , science , and nature _
( simon @xmath24 schuster , n.y . , 1992 ) ; r. chapman and j. c. sprott , _ images of a complex world : the art and poetry and chaos _ ( world scientific , singapore , 2005 ) .
y. ueda , c. hayashi , n. akamatsu , and h. itakura , electronics @xmath25 communication in japan * 53 * , 31 ( 1970)(in japanese ) ; english translation in : y. ueda , _ the road to chaos _
( aerial press , santa cruz , 1992 ) . | we discuss the works of one of electronic art pioneers , ben f. laposky ( 1914 - 2000 ) , and argue that he might have been the first to create a family of essentially nonlinear analog circuits that allowed him to observe chaotic attractors .
* as any branch of science , dynamical chaos theory has no definite birthdate or starting point , but a certain number of milestones .
edward n. lorenz recognized the chaotic attractor , which now bears his name , in numerical simulations somewhere around 1961 @xcite , while leon chua incarnated chaotic attractors with a specially designed electronic circuit in 1983 @xcite . here
we present evidences that ben f. laposky was able to create a family of analog circuits that allowed him to observe chaotic attractors and other trademarks of nonlinear science as early as 1953 .
* although laposky , a draftsman by profession , had received a proper recognition as a pioneer of electronic art , at no time his name has emerged in the context of dynamical chaos theory .
the circuits he had implemented for generation of `` oscillons '' on the screen of a cathode ray tube oscilloscope , remain a mystery @xcite .
it is known that some of his thirty - seven circuits @xcite had `` as many as 70 different setting of controls''@xcite and that ac - voltage has been used for the circuit feeding .
our analysis is based on the vanity press booklet with the still photos of the fifty - six oscillons , which were exhibited at the sanford museum ( cherokee , iowa ) in 1953 @xcite .
there are three oscillons that captured our attention ( figs . 1 , a - c ) .
the first oscillon ( fig .
1a , @xmath0 in the booklet ) looks very similar to the celebrated rssler attractor @xcite ( fig .
1d ) , while the second one ( fig .
2b , @xmath1 ) looks like a multi ( in this case , three)- scroll attractor @xcite ( fig .
1e ) , a typical output of chaotic circuits which include nonlinear elements with nonmonotonous input / output characteristics .
; ( b ) @xmath1 ; and ( c ) @xmath2 .
rssler attractor , with the parameters @xmath3 , @xmath4 , and @xmath5 ( d ) , three - scroll attractor , @xmath6 , @xmath7 , @xmath8 @xcite ( e ) , and limit - cycle regime of the hindmarsch - rose model ( f ) . while the last attractor was modified by a nonlinear frame transformation , the first two
were obtained by a simple adjusting rotation of the original frame @xmath9 and subsequent projection on the @xmath10-plane .
( g ) the periodic bursting pattern , @xmath11 , generated by the limit cycle from fig .
1c , bottom row.,scaledwidth=48.0% ] the words `` looks similar '' stand here for the similarity of the attractor shapes only .
obviously , this observation can not be taken as a rigorous proof of the statement that laposky observed chaotic attractors , yet it certainly confirms the fact that the electronic systems he used were essentially nonlinear .
the third oscillon ( fig .
1c , @xmath2 ) provides more substantial evidence of the last statement .
it represents a peculiar limit cycle , which , we believe , demonstrates the bursting phenomenon @xcite , a dynamical regime of neuronal activity where a neuron periodically fires discrete series of spikes , and which can be reproduced with nonlinear models only . in order to substantiate this claim we employed the hindmarsh - rose model @xcite , with the parameters @xmath12 , @xmath13 , and @xmath14 . using the shape of the model attractor in a three - dimensional space @xmath9
, we tried to find a frame transformation , @xmath15 , which maps the original frame space onto a new one , @xmath16 , so that the new coordinates of reference attractor points , @xmath17 , would maximally approach the corresponding reference points of the oscillon , @xmath18 ( see appendix a ) .
namely , we first digitized the image of the oscillon and distributed twenty - six reference points over it .
since allowable frame transformations should be smooth in order to preserve the topology of the attractor , we assumed the functions @xmath19 to be forth - order polynomials , and then minimized the quantity @xmath20 in a standard least - squares fashion , by varying the polynomial coefficients .
the resulting limit cycle ( fig .
1f ) has the shape of the original oscillon , and produces a typical periodic bursting pattern ( fig .
1 g ) .
thus the oscillon @xmath2 might be a fingerprint of an analog model of neuron , which was created even before the very first mathematical model was presented @xcite .
finally , we would like to underline that the trajectory shown on fig .
1c exhibits a specific type of instability @xcite which is irreproducible with simple linear setups ( for example , by crumpling a lissajous curve ) .
our results give substantial evidences that ben f. laposky had all ingredients needed to encounter chaotic regimes of analog electronic systems and it is quite _ probable _ that he had witnessed these regimes while tuning the parameters of his mysterious circuits @xcite . yet `` science starts from problems , and not from observations '' @xcite , so that , even if our hypothesis is correct , laposky can not be nominated for the discovery of chaotic attractors ( although he certainly deserved to be mentioned in the curriculum vitae of electronic chaos ; see appendix b ) .
but what is most exciting about the oscillon story is that the use of pure aesthetic criteria , which guide artist s preferences , had led to the selection of several aperiodic , chaotic attractor - like structures from more than 6,000 images @xcite decades before scientists started to talk about the `` beauty of chaos '' @xcite . |
a recent hst imaging project of the galactic globular cluster m4 ( go 8679 ) has so far investigated both the low - mass end of the hydrogen burning main sequence @xcite and the termination of the white dwarf cooling sequence @xcite .
the data for the cluster members was isolated from background / foreground contaminants by using the proper motion of the cluster with respect to the field over a 6 year baseline ( previous epoch data obtained with hst in cycle 4 , 1995 @xcite ) .
richer et al . ( 2002 ) found the m4 cluster mass function to be slowly rising to low masses . in hansen
et al . ( 2002 ) , the cluster white dwarfs were used as chronometers to provide an age estimate of 12.70 @xmath6 0.35 gyrs ( 1-@xmath14 ) largely independent of the turn - off for m4 . in this contribution
, we shift our focus to the background contamination itself ( the spheroid and galaxies ) and isolate these populations for investigation .
a separate contribution will address the main - sequence stars , luminosity and mass functions of the spheroid , and foreground thin / thick disk components .
stars , gas and dust in the disk , bulge and spheroid of our galaxy only account for @xmath1510% of the total mass within @xmath16 = 50 kpc @xcite .
the remaining mass is believed to reside in the dark halo of our galaxy ( dark matter ) , and determining the nature of this mass is a critical issue today in astrophysics .
recently , the macho project has analyzed data from microlensing events in the direction of the lmc and determined that the mean mass of the lenses is 0.5 @xmath6 0.3 @xmath17 , with a @xmath1520% macho fraction @xcite .
this naturally suggests faint white dwarfs as the source of these lenses .
even if the machos can not account for the total dark matter contribution , determining the properties and number density of faint halo white dwarfs is important in several areas of astrophysics .
these include studying the imf of population iii stars ( through their remnants ) and constraining the star - formation history of our galaxy .
studying white dwarfs in the galactic bulge or spheroid from the ground is difficult for two reasons .
first , the end of the white dwarf cooling sequence for 0.5 @xmath17 objects in a population 12 gyrs old is @xmath18 17 @xcite , which , at the center of the galaxy ( 8 kpc ) , is @xmath19 31.5 .
the depth of the m4 study ( @xmath12 30 ) , however , is faint enough to detect the brighter end of this cooling sequence .
secondly , piercing through the disk of our galaxy picks up many foreground metal - rich disk stars which contaminate the sample .
although observing at higher latitudes above the center helps avoid the thick disk , the resulting spheroid number density also drops off rapidly , @xmath20 @xmath21 @xcite .
with hst we can achieve the depth required to measure these faint stars , but there are several disadvantages .
first , we are dealing with small - number statistics due to the limited field of view ( for instance , wfpc2 has a field area of 5.7 arcmin@xmath22 ) .
more importantly , long exposures pick up many galaxies , such as seen in the hdf , which can mimic stars in faint photometry .
given the caveats listed above , we attempt in this paper to isolate potential white dwarfs in the field by separating out populations using both their morphology and their kinematics .
after briefly presenting the data in [ data ] , we begin by finding galaxies and establishing our extragalactic zero - motion frame of reference .
this then allows us to determine two important quantitities directly measured from a sample of extragalactic objects : the velocity of the local standard of rest ( [ circ ] ) and the absolute proper motion of m4 ( [ absolute ] ) . in [ cmd ]
we examine the different populations which make up the corrected proper - motion and color - magnitude diagrams and discuss the spheroid population .
next , we analyze faint , blue stars in this population and identify our best candidates for disk and spheroid white dwarfs ( [ analysis ] ) .
this includes comparing our results to the expected numbers of white dwarfs given the searchable volume in our data and the various population - density distributions along the line of sight . in
[ discussion ] , we discuss the current status of the search for dark halo white dwarfs as well as some of the different views which have been presented in the recent literature .
the current results are placed in the context of these independent efforts .
the overall study is concluded in [ conclusions ] .
the data ( 35 hours in f606w , 55 hours in f814w ) for the present and previous observations ( 1995 - cycle 4 @xcite ) were reduced using the daophot / allstar @xcite and allframe @xcite photometry packages .
individual frames were registered and co - added using the daomaster and montage2 programs . for this we transformed all images for a given ccd to a common coordinate system for each epoch and filter and combined the images using standard iraf tasks . as the fields are dominated by m4 cluster members , smaller matching radii rejected non cluster members .
a plot showing @xmath23 vs @xmath24 between the older and newer epochs therefore shows the cluster stars grouped at the origin and a secondary group of stars offset from this .
the transformations were then iterated using a list of stars likely to be cluster members as selected from this clump near the origin .
the cmd of the resulting stars verified their cluster origin .
the list of these stars were then transformed back onto each of the individual frames which were reduced using allstar , with the recentering option enabled .
the output allstar files , which contain only cluster members , were matched with daomaster using a 20-term transformation equation .
the 20-term transformation was found to reduce systematic errors significantly from the 6-term transformation .
we produced histograms of @xmath23 and @xmath24 for the cluster stars and minimized streaming motions near the edges of the ccd caused by distortions by reducing the matching radius until it was similar in magnitude to the rms error in the transformation .
the residual systematic bias in the proper motion diagram is evaluated below .
finally , montage2 was used to expand the frames by a factor of 3 before registering them as noted above .
this procedure is similar to a drizzling routine and effectively decreased the cluster proper - motion dispersion in the proper - motion diagram .
the final photometry list includes only those stars which were measured in both filters and at both epochs , and which passed a visual inspection .
further details of the reduction and calibration of the data set used in this analysis will be described in richer et al .
( 2003 ) ( see also 2 of richer et al . ( 2002 ) ) .
we summarize the key parameters for both the cluster and field population and present the observational log in table 1 . in figure 1
, we present the proper - motion diagram for our data .
the motion of the cluster with respect to the field ( 6 year baseline ) , can be clearly seen as the tight clump toward negative proper motions . for the units , we first translated the @xmath25 and @xmath26 pixel motions into galactic coordinates ( @xmath0 , @xmath1 ) by using a rotation angle which aligned the @xmath26 axis of the ccd to the north galactic pole . to express these proper motions
, we adopt the convention that the vector proper motion is given by @xmath27 = @xmath28 , where @xmath29 and @xmath30 are unit vectors in the corresponding directions .
we then converted the hst wfpc2 pixels into arcseconds using the plate scale ( 01/pixel , biretta et al .
2002 ) of the ccd .
the motions are then divided by the baseline of 6 years and converted to mas / yr . zero motion in the diagram , ( @xmath31@xmath29 , @xmath32@xmath30 ) = ( 0 , 0 ) , is centered on well measured , bright galaxies ( larger red dots ) and is described in detail in [ centering ] .
a small systematic bias , seen as a stretching of the proper - motion diagram in the @xmath0 direction , is evaluated to be 0.26 mas / yr . this was measured by producing histograms of the proper - motion diagram for different magnitude cuts . comparing the dispersions in the @xmath0 and @xmath1 directions for the brightest stars
directly gives the error .
this systematic bias has been factored into the error budget for all measurements .
isolating galaxies in the current study is crucial as they can mimic faint blue white dwarfs in the data .
additionally , since the galaxies are not moving , they represent a fixed zero - motion position in the proper - motion diagram from which we can measure absolute motions .
visually , the images of m4 only show a small number of obvious galaxies .
were it not for the greatly larger number of foreground stellar objects , the 1.3 magnitudes of foreground visual extinction and the higher background produced by scattered light and zodiacal light , this image would likely look similar to the hdf @xcite which shows 1781 galaxies in the range 26 @xmath33 @xmath34 29.5 . to measure galaxies in our field , we used the image morphology classification tool , sextractor @xcite .
sextractor assigns a stellarity index to each object which can be used to distinguish between stars and galaxies .
the stellarity is determined through a neural network , which learns based on other , high signal - to - noise , stars .
although sextractor was not able to recover statistics for every object that allstar measured in our data , the classifications are @xmath1575% complete through @xmath35 = 2429 .
fainter than @xmath35 = 29 , sextractor was only able to measure @xmath1540% of the allstar sources and struggled to classify objects as stars or galaxies .
altogether , just over 50 galaxies were identified and measured using a stellarity @xmath36 0.2 cut .
figure 2 ( top ) shows the sextractor classifications as a function of magnitude for all objects .
the distribution of points on this diagram can be broken into two classes which are separated by the horizontal lines .
the stars are found predominantly in a clump near stellarity = 1 down to a magnitude of @xmath12 28 .
fainter than this , the classifications are not as good .
however a clump of stars can be seen extending to @xmath12 29 with a sharp cutoff in stellarity at 0.7 .
the galaxies are found near the bottom of the figure , at stellarity = 0 .
here we have chosen a conservative stellarity @xmath36 0.2 cut to include possible outliers above the predominant galaxy sequence .
very few objects ( @xmath154% for v @xmath36 29 ) are seen between the stars and galaxies indicating that the classifications are reliable . at the faintest magnitudes , @xmath12 29 , a clump of objects
can be seen at stellarity = 0.5 .
this is expected as the program will , on average , choose stellarity = 0.5 for an object which it can not classify .
shape parameter comparisons from artificial star tests also confirm that the stellarity is measured accurately down to this limit .
therefore @xmath12 29 represents the limit at which we can accurately separate stars from galaxies . in figure 2 ( bottom )
we use the stellarity to illustrate the separate populations in our data . here
, the stellarity of all objects is plotted against the total absolute proper motion , @xmath37 = @xmath28 as determined with respect to the center of the large red dots in figure 1 .
we will justify this location in the next section .
first , we note two clear clumps for stars ( at constant stellarity @xmath15 1 ) representing the cluster ( @xmath38 22.5 mas / yr , with a small dispersion ) and field populations ( @xmath38 5 mas / yr , with a much larger dispersion ) .
this bi - modality continues for lower stellarities down to stellarity @xmath15 0.7 .
the galaxies ( stellarity @xmath15 0 ) are obviously found at low @xmath37 .
however , a tail is seen to higher proper - motion displacements ( these are the small red dots in figure 1 ) . as we will see in [ centering ] , this tail represents the faintest galaxies and indicates that our ability to measure the proper motions has degraded due to reduced signal - to - noise . in figure 1
, we showed the locations of bright , well - measured galaxies ( large red dots ) which represent an absolute background reference frame .
as these galaxies clumped together , we first estimated their position and measure the absolute proper motion of all objects with respect to this point .
figure 3 shows the total proper - motion displacement as a function of magnitude for both stars ( top ) and galaxies ( bottom ) .
the distributions are found to be very different .
the stars behave as expected and are confined within a constant envelope in @xmath37 across all but the faintest magnitudes .
note that the spread in the distribution of m4 cluster members ( @xmath38 22.5 mas / yr ) for @xmath39 28 is not due to the main - sequence stars . at @xmath40 28 ,
the main sequence of m4 contains very few stars @xcite and so the scatter is due to the faintest white dwarfs in the cluster . on the other hand ,
the total proper - motion displacement of the galaxies ( bottom ) is found to completely degrade for @xmath40 27 . considering that we know from figure 2 that the stellarities for the galaxies are measured accurately to well beyond this limit
, this flaring must be entirely due to difficulties in the astrometric centering on these faint extended sources .
therefore the galaxies with @xmath41 27 should dictate the location of the zero - motion frame as some of their fainter counterparts have been poorly measured .
the centroid of the 12 galaxies which satisfy this criteria represents the zero - motion frame of reference shown on all proper - motion diagrams ( see large red dots in figure 1 ) . the one - dimensional error in this location
is calculated as @xmath14/@xmath42 mas / yr , where @xmath14 is the proper - motion dispersion of the galaxy sample ( @xmath43 mas / yr ) .
we now use the extragalactic stationary frame of reference established above to measure two important quantities .
the first , the circular - speed curve , is a plot of the velocity of a test particle ( @xmath8 ) moving in a circular orbit in the galactic plane and around the galactic center , vs the distance @xmath16 at which it is located relative to the center .
the ratio of these quantities at the solar radius ( @xmath4 = @xmath5 ) represents one of the more difficult problems in galactic structure and directly provides the mass interior to @xmath3 .
this constant is fundamentally important in understanding the dynamics of the galactic halo and the local group . for the motion in the galactic plane
, we measure an angular proper motion from the center of the galaxy clump to the center of the spheroid clump of @xmath31@xmath29 = @xmath105.53 @xmath6 0.56 mas / yr . the spheroid clump was chosen by first eliminating foreground disk stars by overplotting a solar metallicity disk dwarf fiducial on our cmd ( presented in [ cmd ] ) .
these disk stars represent the redder , brighter stars located above m4 .
we then removed both m4 and galaxies , and iteratively determined the center of the inner halo based on @xmath14-clipping to remove outliers .
the resulting population is shown within the large circle in figure 1 . assuming the spheroid is not rotating ( see below ) , the vector motion of @xmath31@xmath29 = @xmath105.53 @xmath6 0.56 mas / yr represents the reflex motion of the sun , a combination of the lsr circular orbit and the deviation of the sun from that circular orbit ( solar motion ) .
the uncertainty in this number is derived from the quadrature sum of the dispersions in each of the galaxy and spheroid clumps .
we can correct this observed proper motion for the known solar motion in the galactic plane , @xmath44 = 5.25 @xmath6 0.62 km / s in the direction of galactic rotation @xcite .
orienting this solar motion onto the @xmath0 direction , we find a correction of @xmath45@xmath31@xmath29 = 0.19 @xmath6 0.02 mas / yr . therefore , @xmath4 = @xmath8/@xmath3 = @xmath464.74@xmath31@xmath29 = 25.3 @xmath6 2.6 km / s / kpc .
the angular velocity of the circular rotation of the sun can be directly compared to the oort constants , @xmath47 and @xmath48 @xcite , which measure the shear and vorticity of the disk .
recent analysis based on _ hipparcos _ measurements of 220 galactic cepheids gives @xmath4 = @xmath49 = 27.19 @xmath6 0.87 km / s / kpc @xcite , larger than our value but consistent within the uncertainty .
other measurements , such as the sgr a@xmath50 proper - motion study @xcite , also find a higher value than ours , @xmath4 = 27.2 @xmath6 1.7 km / s / kpc . at @xmath3 = 8.0
@xmath6 0.5 kpc , these two measurements give a solar reflex velocity of @xmath7 = @xmath8 @xmath15 218 km / s , in good agreement with the independent iau adopted value of 220 @xmath6 15 km / s @xcite based on kinematics of high velocity stars .
the most recent study by bedin et al .
( 2003 ) uses a single qso in a different field of m4 than studied here and finds @xmath4 = 27.6 @xmath6 1.7 km / s / kpc .
however , other studies , such as kuijken & tremaine ( 1994 ) find much smaller values .
based on a set of self - consistent solutions for different galactic parameters , they find @xmath7 = 180 km / s .
merrifield ( 1992 ) also calculates a lower local standard of rest velocity , @xmath7 = 200 @xmath6 10 km / s , using rotation of h i layers . for @xmath3 = 8.0 @xmath6 0.5 kpc ,
our result gives a similar value to this latter study , @xmath7 = 202.7 @xmath6 24.7 km / s .
only 15% of the uncertainty in our result is contributed by the 0.5 kpc uncertainty in the galactocentric distance with the remaining error almost entirely due to the dispersion in the galaxy sample .
the dispersion in the spheroid sample is calculated from those stars within the circle in figure 1 and has also been folded into this uncertainty .
although our value for @xmath4 is lower than some previous studies of the circular speed curve , it is consistent with those studies within the 1-@xmath14 uncertainty .
one consideration that we are currently investigating is possible rotation by the spheroid itself , which would cause an over / underestimate of the solar reflex motion depending on the degree to which stars in front of the tangent point dominate over stars behind it .
given the small field of view , we could also potentially be affected by debris tails or small groups of stars with peculiar orbits . despite these caveats ,
we stress the importance of this measurement .
kuijken & tremaine ( 1994 ) estimate that if they are correct and a reduction of @xmath8 from 220 km / s to 180 km / s is needed , the line - of - sight velocity of m31 relative to the galactic center would correspondingly increase by 25% .
this , in turn , would increase the mass of the local group from timing arguments by 35% . although additional methods ( see kuijken & tremaine 1994 for a summary ) to constrain the circular - speed curve are being used ( e.g. , using local group kinematics , motions of globular clusters or halo stars , tangent - point measurements of the inner rotation curve and outer - rotation curve - measurements using ob stars and h ii regions , etc . ) , they are not superior to measurements made with extragalactic reference frames .
these global measurements , such as the present work and the sgr a@xmath50 proper motions @xcite , avoid galactic variations in the sample of objects and therefore reduce systematic uncertainties @xcite .
the apparent motion of the spheroid stars in the direction perpendicular to the plane is very small , @xmath32@xmath30 = @xmath100.32 @xmath6 0.56 mas / yr ( see figure 1 ) .
the correction for the component of the known solar motion , @xmath51 = 7.17 @xmath6 0.38 km / s @xcite in this direction is @xmath45@xmath32@xmath30 = @xmath100.12 @xmath6 0.01 mas / yr . therefore , the residual velocity for @xmath3 = 8.0 @xmath6 0.5 kpc is found to be 7.4 @xmath6 21.2 km / s , consistent with a stationary population .
the space motions of galactic globular clusters are of interest in order to constrain and reconstruct their orbits around the galaxy .
this yields important information on cluster origins and destruction processes , as well as galactic dynamics .
formation scenarios of the galaxy can be better understood by coupling the kinematics of the clusters with other properties , such as the metallicities and ages @xcite .
three primary methods exist to determine the space motions of globular clusters and all involve measuring the absolute proper motion of the stars in the cluster .
the indirect methods involve either using secular parallaxes ( which require a knowledge of galactic rotation ) or using the new ` bulge - relative ' method ( which requires the distance and proper motion of the bulge : terndrup et al .
the preferred method involves using extragalactic objects , galaxies or qsos , and directly measuring the motion of the cluster .
only three previous estimates of the m4 absolute proper motion exist in the literature .
cudworth & rees ( 1990 ) used bright field stars as a reference to the brighter cluster stars to estimate @xmath9= @xmath1011.6 @xmath6 0.7 mas / yr , @xmath11= @xmath1016.3 @xmath6 0.9 mas / yr . dinescu et al . ( 1999 ) used _ hipparcos _ field stars as a reference to carefully measure the motion of the cluster after accounting for plate - transformation systematics : @xmath9 = @xmath1012.50 @xmath6 0.36 mas / yr , @xmath11= @xmath1019.93 @xmath6 0.49 mas / yr . the bedin et al .
( 2003 ) study finds @xmath9= @xmath1013.21 @xmath6 0.35 mas / yr , @xmath11= @xmath1019.28 @xmath6 0.35 mas / yr .
here we present an absolute proper - motion measurement of the cluster measured with respect to a sample of galaxies , an important measurement considering that m4 is a very well studied cluster .
as discussed earlier , our coordinate differences were measured by centering on the cluster stars .
we then found the center of the extragalactic distribution and directly measured the difference between this and the center of the m4 clump . converting galactic coordinates to the equatorial system
, we find @xmath9= @xmath1012.26 mas / yr , @xmath11= @xmath1018.95 mas / yr , with a one - dimensional uncertainty of 0.54 mas / yr . the error in the mean for the absolute proper motion is the quadrature sum of the total errors for each of the galaxy and cluster star distributions .
we ve also factored in the 0.26 mas / yr systematic bias discussed in [ data ] .
the final errors in our absolute proper motion are completely dominated by the dispersion in the galaxy sample . comparing these results directly to previous studies
, we find that our total proper motion ( @xmath37 = @xmath52 ) is @xmath154% smaller than the dinescu et al . ( 1999 ) value , @xmath1513% larger than the cudworth & rees ( 1990 ) value and @xmath153% smaller than the bedin et al .
( 2003 ) study . for completeness
, we can also convert our cluster proper motions into absolute space motions .
the solar motion is ( @xmath53 , @xmath54 , @xmath55 ) = ( + 10.0 , + 5.2 , + 7.2 ) km / s @xcite , where @xmath53 is positive towards the galactic center , @xmath54 is in the direction of galactic rotation , and @xmath55 is out of the plane .
therefore , for m4 we derive an absolute space motion of ( @xmath56 , @xmath35 , @xmath57)@xmath58 = ( 57.0 @xmath6 6.3 , @xmath46186.1 @xmath6 6.3 , @xmath462.4 @xmath6 4.5 ) km / s with respect to the lsr , for a distance of @xmath59 = 1720 pc and a cluster radial velocity of @xmath60 = 70.9 @xmath6 0.6 km / s ( peterson , rees & cudworth 1995 ) .
the uncertainties calculated for the space motions above do not include any distance uncertainty .
in figure 4 we again present the final proper - motion diagram and also the corresponding color - magnitude diagram for all objects within the m4 field . as we mentioned earlier , the tighter clump of dots in the proper - motion diagrams represent m4 ( proper - motion dispersion , @xmath61 = 1.9 mas / yr ) whereas the more diffuse clump represents the spheroid population plus other stars along the line of sight ( proper - motion dispersion @xmath61 = 3.7 mas / yr ) .
the dispersions in these cases actually measure a combination of each population s intrinsic velocity dispersion as well as scatter produced from instrumental errors .
the latter is clearly evident as we get smaller dispersions for brighter magnitude cuts which eliminate the lowest signal - to - noise ratio objects ( the above numbers however do reflect the dispersions measured from using the entire data set ) .
a detailed analysis of the intrinsic dispersion , after correcting for instrumental scatter , will be provided in saviane et al .
( 2003 ) . the extragalactic objects in our data , which have not moved during the 6 year baseline
, are also convolved within the field clump identified above .
these are identified with red dots in the top - right panel .
the cmd ( left panel ) shows the remarkably tight m4 main - sequence extending down to at least @xmath12 27 on top of the foreground / background stars .
a rich cluster white - dwarf cooling sequence is also seen stretching from 23 @xmath62 29.5 .
as mentioned earlier , the line of sight through m4 ( @xmath59 = 1.7 kpc ) also intersects the spheroid of our galaxy at a projected distance of 2.2 kpc above the galactic nucleus ( for a tangent point distance of 7.6 kpc ) .
this population is more easily seen when the m4 main - sequence stars are removed by proper - motion selection ( right panel ) .
the field population clearly extends to @xmath12 30 and shows some evidence for potential white dwarfs in the faint - blue region of the cmd
. these can be clearly confused with the galaxies in the @xmath35 , @xmath63 plane , thereby illustrating the importance of stellarity measurements .
the main sequence of the field population in the range @xmath35 = 2430 is too broad to be consistent with the color distribution expected from a population with no metallicity spread , at the distance of the spheroid .
however , we can rule out a contribution of stars in these data from the galactic bulge given that the line of sight passing through m4 lies well outside the infrared bulge as imaged by _ cobe _ @xcite .
the field also misses the highest surface - brightness isophotes in binney , gerhard & spergel ( 1997 ) and clearly avoids the classical metal - rich bulge population ( mcwilliam & rich 1994 ; frogel , tiede & kuchinski 1999 ) .
similarly , the redder stars in the field can not be stars in the tri - axial bar near the galactic center .
the orientation of the bar is such that the near side of the semi - major axis ( @xmath152 kpc ) lies in the first quadrant .
the angle between the sun - center line and the bar s major axis in the plane of the disk is @xmath64 20@xmath2 @xcite .
the line of sight through m4 to the center of the galaxy , ( @xmath0 , @xmath1 ) = ( 350.97@xmath2 , 15.97@xmath2 ) also falls in this quadrant .
given the thickness of the bar and axis ratio 1.0:0.6:0.4 ( binney , gerhard & spergel 1997 ) , an azimuthal angle of 20@xmath2 would place the near side of the bar major - axis only @xmath156.5 kpc from us ( assuming an 8 kpc galactocentric distance ) .
however , with a latitude of @xmath1516@xmath2 , the line of sight through m4 is already 1.9 kpc above the plane at a distance of 6.5 kpc and therefore well above the thickness of the bar .
other bar models , such as that in cole & weinberg ( 2002 ) , would also not intersect our line of sight .
a better explanation for the thickness of the main sequence is that a small admixture of metal - rich thick disk stars along the line of sight make up the redder population .
this is expected in these data given that the scale height of the thick disk is 1 - 1.5 kpc ( kuijken & gilmore 1991 ; bienaym , robin & crz 1987 ) .
these would be co - rotating , perhaps with a small lag , and therefore reside closer to the location of the galaxies .
however , the proper - motion diagram does not show any obvious evidence of this population even when isolating only the reddest , brightest possible thick disk stars .
the exact interpretation of the thickness of the observed main sequence is therefore still somewhat unclear and currently under investigation .
with galaxies clearly separated from stars in these data , finding spheroid / dark halo white dwarfs should be relatively easy considering they must be faint , blue and have a proper motion consistent with the spheroid clump .
since we also may have contamination from the thin and thick disk along our line of sight , the best way to isolate the inner - halo white dwarfs is to select them from the reduced proper motion diagram ( rpmd ) .
this two dimensional plot ( @xmath65 vs @xmath63 , where @xmath65 = @xmath35 + 5log(@xmath37 ) + 5 , and @xmath37 is measured in @xmath66/yr ) removes the different distances of individual populations by using the size of the proper motion as an effective distance indicator to offset the apparent magnitude . therefore , similar populations , despite their distance , will occupy similar regions on this diagram based both on their stellar properties and kinematics .
figure 5 presents our stellar objects ( stellarity @xmath67 0.7 ) for the large field clump ( see figure 4 ) on this plane together with the m4 stars ( red dots ) .
the rpmd clearly shows that the majority of our spheroid main - sequence sample ( small dark dots ) is similar to m4 ( red dots ) .
this is expected considering the mean properties , such as the age and metallicity , are similar for both m4 and the spheroid . for a fixed magnitude ,
we see objects both above and below the m4 main - sequence stars on this plane .
the objects above represent those with smaller proper motions ( pop i ) and those below have greater proper motions ( pop ii ) .
the m4 white - dwarf cooling sequence is seen towards the blue end of the diagram and represents a locus for potential halo white dwarfs .
several candidate white dwarfs from the field population ( larger dark dots ) are clearly seen sprinkled through the range 14 @xmath68 25 , near @xmath69 1 .
most of these objects are found above the m4 white dwarfs , suggesting they have smaller proper motions ( disk objects ) .
three objects are found below the m4 white dwarfs suggesting possible spheroid stars . also shown
are the locations of the thick - disk and putative dark halo - white dwarfs from the mendez ( 2002 ) study ( green squares ) and those from the nelson et al .
( 2002 ) study ( blue triangles ) . the faintest mendez object as well as the faintest two nelson objects agree very well in location with the m4 white dwarfs on the rpmd ( these are the claimed dark halo detections , see [ discussion ] for further discussion ) .
the remaining mendez and nelson objects are found above the m4 stars , suggesting a possible thick disk origin .
these points have all been corrected to match the reddening and extinction along this line of sight .
we also note that we do not select white dwarfs based on their location in the color - magnitude diagram for a reason .
one would naively expect that the spheroid white dwarfs should be consistent with the location of the m4 white dwarf cooling sequence shifted down to the tangent point ( 3.2 magnitudes fainter ) .
however we need to consider that the spheroid white dwarfs will occupy some depth in distance around the tangent point so there are both closer and farther objects .
since the luminosity function @xcite of white dwarfs at the tangent point rises sharply at a point fainter than our limiting magnitude , we can expect to be systematically biased to seeing some fraction of the fainter white dwarfs within the region of the rising luminosity function slope if they reside closer to us than the tangent point .
since these fainter white dwarfs are redder than their brighter counterparts , the locus of these points will shift the + 3.2-magnitude fiducial up to brighter magnitudes , on a parallel slope to the white dwarf cooling sequence .
therefore , we can not use a color cut on the cmd to select these objects . the approach used below
does not suffer from these effects . in order to assign our candidate white dwarfs to particular galactic components ,
we need to model the underlying white dwarf populations .
we model four such components ; the thin disk ( a double exponential with radial scale length 3 kpc and vertical scale height 0.3 kpc ) , thick disk ( a double exponential with radial scale length 3 kpc and vertical scale height 1 kpc ) , spheroid / stellar halo ( a power law with @xmath70 , where @xmath71 is the spherical galactocentric radius ) and a putative ` dark halo ' ( a power law with @xmath72 ) . the solar - neighborhood ( white dwarf only ) normalizations for the various components
are taken to be @xmath73 ( holberg , oswalt & sion 2002 ) for the thin disk , @xmath74 for the thick disk ( oppenheimer et al .
2001 ; reid , sahu & hawley 2001 ) , @xmath75 for the spheroid ( gould , flynn & bahcall 1995 ) and @xmath76 for the dark halo ( corresponding to a fraction 20% of the dark matter density , as indicated by the macho group s results , alcock et al .
2000 ) .
the above density laws are projected onto the line of sight , taking into account the increasing volume at larger distance ( for fixed solid angle ) .
using the same white dwarf models as in hansen et al .
( 2002 ) , we derive monte - carlo realizations of the field white dwarf populations distributed along our line of sight . the thin ,
thick and spheroid populations are drawn from a population with progenitor imf @xmath77 .
the dark halo population is drawn instead from a population with a chabrier imf ( chabrier 1999 ; imf2 in the notation of that paper ) .
the models of hurley , pols & tout ( 2000 ) were used to provide main - sequence lifetimes .
the thin - disk population was assumed to be forming stars at a continuous rate ( over the past 10 gyr ) while all the others were assumed to be 12 gyr old bursts .
the monte - carlo calculations include the detection probability using the incompleteness corrections from richer et al .
( 2002 ) and hansen et al .
( 2002 ) .
the final results indicate that , on average , we expect 7.9 thin - disk , 6.3 thick - disk and 2.2 spheroid white dwarfs .
the dark halo will contribute 2.5 white dwarfs if 20% of the dark halo is made up of da white dwarfs .
the first three ( i.e. , the standard ) populations are dominated by bright white dwarfs located near the tangent point ( even for the disk populations , as the radial exponential largely compensates for the decrease due to the vertical exponential , at least for this line of sight ) . the dark - halo population , on the other hand ,
is dominated by objects at distances @xmath152 - 3 kpc , a consequence of the bias inherent in the chabrier imf towards higher progenitor masses and thus shorter main - sequence lifetimes ( corresponding to longer white dwarf incarnations ) .
this results in an anticipated systematic color difference between any dark - halo white dwarfs ( relatively red ) and those from standard populations ( relatively blue ) .
we stress this applies particularly to our data set and does not necessarily extend to wider - field surveys . in figure 6 , we display the location of the white dwarfs in the rpmd .
we also overlay the positions of the simulated sample of white dwarfs ( open circles ) on this diagram ( for a much larger area than that observed ) .
the thin - disk ( top - left ) and thick - disk ( top - right ) stars are generally found brighter than the spheroid ( bottom - left ) stars , with a limit at @xmath78 25 beyond which there are very few disk white dwarfs .
a similar threshold was used by flynn et al .
( 2001 ) to separate disk and halo white dwarfs .
they found that most dark halo white dwarfs would lie at 24.3 @xmath79 26.3 and peak at @xmath78 25.3 ( corrected to the extinction along this line of sight ) .
the _ relative _ numbers of simulated thin - disk , thick - disk and spheroid objects plotted in figure 6 have been scaled to match the expectations in our data .
the @xmath65 @xmath15 14 cutoff at the top of the diagram indicates the saturation limit of our data .
the simulated dark - halo white dwarfs ( open circles , bottom - right ) are clearly distinct from these conventional populations and are found at redder colors ( for an explanation see [ simulations ] ) .
also included on this panel are those objects which did not satisfy the stellarity or proper motion cut ( small crosses ) to show that the data do extend fainter than @xmath65 = 25 and well into the dark halo regime .
although it is difficult to separate thin-/thick - disk white dwarfs , we can conclude that nine of our objects are consistent with disk white dwarfs ( @xmath80 23 , @xmath81 1.5 ) .
three objects in our data ( @xmath65 = 23.15 , @xmath63 = 0.48 , @xmath65 = 23.78 , @xmath63 = 0.93 , and @xmath65 = 24.09 , @xmath63 = 0.93 ) which are fainter than this cut are found within the tail of the thick disk simulated sample .
these objects are however in excellent agreement with the simulated spheroid white dwarfs ( bottom - left ) .
these are also the three objects which fall below the m4 white dwarfs in figure 5 .
although the predicted number of stars allows for incompleteness in the data set , we must still correct these observed numbers of stars for the incompleteness caused by mismatches in the sextractor vs allstar classifications . in [ expectedgalaxies ]
we estimated 25% of the allstar sources were missed for these magnitude bins .
our final corrected numbers of detected objects are therefore 12 disk white dwarfs ( 14 expected ) and 4 spheroid white dwarfs ( 2 expected ) and therefore in good agreement with the predictions considering that the observed spheroid number is an upper limit ( i.e. , two of the objects are consistent with the tail of the thick disk distribution ) .
a key result from figure 6 is that the simulated dark halo region is basically not populated with data points .
only one object which we classified as thick disk , @xmath65 = 22.25 , @xmath63 = 1.38 , marginally agrees with the edge of the simulated sample ( small open circles )
. less than 4% of our simulated dark - halo sample is found with a reduced proper motion less than this object , suggesting @xmath360.1 expected real objects .
also , the color of this star falls close to the blue edge of the simulated sample and is only marginally consistent with the expected location of dark - halo white dwarfs .
a possible explanation for the other extremely red object , as well as possibly this object , is that these may be binary thin-/thick - disk white dwarfs with red main - sequence companions .
we can also rule out the spheroid candidate objects as potential dark - halo white dwarfs given that they have much bluer colors . we ve added photometric and astrometric error bars to all candidates in the bottom - right panel of figure 6 .
the errors are 1-@xmath14 values from the results of artificial star tests .
for example , the magnitude and color errors are the input vs output photometric values and the positional errors are the input vs output coordinates in the tests .
the @xmath65 error combines the contribution from the magnitude and proper - motion terms .
we have left these error bars off from other panels to prevent cluttering in the diagrams .
we also note that just relaxing the stellarity cut to 0.5 would not allow any of the spurious sources ( small crosses ) to qualify as candidate dark halo white dwarfs given their inconsistent proper motions .
finally , we indicate the limit at @xmath78 27 above which we would have expected dark - halo white dwarfs given our simulations and detection efficiency .
table 2 presents our final white dwarf candidate list , both for the disk and spheroid .
contrary to our result , recent claims by oppenheimer et al .
( 2001 ) , koopmans & blandford ( 2001 ) , nelson et al .
( 2002 ) and mendez ( 2002 ) suggest that a halo white - dwarf population exists and that the number of white dwarfs is in excess of that expected from the stellar spheroid alone .
others ( reid , sahu & hawley 2001 ; hansen 2001 ; richer 2001 ) have argued that this small population is a component of the thick disk .
gibson & mould ( 1997 ) used simple chemical evolutionary arguments to show that it is difficult to reconcile the observed spheroid main - sequence chemical abundance with that expected from an imf which would give the speculated halo white - dwarf density .
we briefly summarize these independent studies and place them in the context of the present observations .
oppenheimer et al .
( 2001 ) sampled 4165 sq .
deg . towards the southern galactic cap and found 38 cool white dwarfs , the majority of which they assign to a dark - halo population .
the data from this study were made available and consequently several authors have re - analyzed the result .
koopmans & blandford ( 2001 ) used a dynamical model and performed a maximum - likelihood analysis on the oppenheimer sample of white dwarfs and found two kinematically distinct populations at a greater than 99% confidence limit ( a thick disk and a flattened halo ) . however , reid , sahu & hawley ( 2001 ) compared the kinematics of the oppenheimer sample to the local m - dwarf sample and suggested that the majority of this population resides in the thick disk . richer ( 2001 ) and
hansen ( 2001 ) examined the temperature and age distribution of the oppenheimer white dwarf sample and found that the star - formation history for most of these stars is more representative of the disk rather than the halo .
the nelson et al .
( 2002 ) study looked for high - proper - motion objects in the groth - westphal strip .
they measured proper motions over a 7-year baseline using wfpc2 observations of a 74.8 arcmin@xmath22 region .
of the 24 high - proper - motion objects detected , five are believed to be strong white dwarf candidates ( two of these are dark halo , three are disk ) and a further two are classified as marginal candidates .
the two halo objects are the faintest two blue triangles in figure 5 and the three disk objects are the brightest three blue triangles in figure 5 . by creating a model of the stellar components of our galaxy , nelson et al .
( 2002 ) demonstrate that the observations are clearly in excess of the expected number of white dwarfs from the galactic stellar components .
although systematic errors due to uncertainties in the model parameters are large , the observations suggest a 7% white - dwarf dark halo .
most recently , mendez ( 2002 ) has presented new evidence for a bi - modal kinematic population of old white dwarfs and suggests that most of the required dark matter in the solar vicinity can be accounted for by these populations .
this analysis is based on a proper - motion membership study of the galactic globular cluster ngc6397 @xcite .
mendez ( 2002 ) used both the reduced proper motion and the colors of field objects to assign six white - dwarf candidates to the thick disk and one to the dark halo ( the faintest green square shown in figure 5 ) .
the mendez ( 2002 ) study gives no information on the morphology of the sources and possible extragalactic contamination .
there is clearly no convincing answer yet as to whether white dwarfs make up an important component of the dark matter in our galaxy .
the controversy spurred from these studies can only be resolved by obtaining deeper , larger images of halo fields which could contain many white dwarfs .
the current data set represents an important constraint in the context of searching for these objects as it is sensitive to white dwarfs from all three ( disk , spheroid and dark - halo ) components and it is only the dark - halo ones that are not found .
furthermore , our images represent the deepest - ever probe into the inner halo of our galaxy and we do see the stellar halo main - sequence progenitors of the white dwarfs . by using only directly observable quantities , such as the proper motions , magnitudes and colors
we have demonstrated that these data do not support a white - dwarf component to the dark halo .
a 6-year baseline of imaging of the globular star cluster m4 has allowed us to separate out the cluster stars from the background spheroid population . over this time ,
distant extragalactic sources have not moved and are identified using image morphology criteria .
we establish a zero - motion frame of reference using these galaxies and search for halo white dwarfs based on their morphology and kinematics .
we find that distinguishing faint galaxies from field stars is impossible when based solely on proper motions , and that an index of stellarity is crucial in separating the two classes .
based on the reduced proper motion diagram , we identify nine thin / thick disk and three spheroid white dwarfs in these data .
these numbers are consistent with the expected contribution of the conventional populations along the line of sight .
additionally , 2.5 dark halo white dwarfs are expected in these data based on a 20% white dwarf dark halo and are not found .
therefore , we do not need to invoke a dark halo population of white dwarfs to explain the present data .
the extragalactic reference frame also allows us to perform two other important measurements .
first , this fixed frame gives us an independent measurement of the fundamental galactic constant , @xmath4 = @xmath5 = 25.3 @xmath6 2.6 km / s / kpc .
this provides a velocity of the local standard of rest @xmath7 = @xmath8 = 202 @xmath6 25 km / s at @xmath3 = 8.0 @xmath6 0.5 kpc , in agreement with independent studies .
secondly , the galaxies give us a direct measurement of m4 s absolute proper motion , @xmath9= @xmath1012.26 @xmath6 0.54 mas / yr , @xmath11= @xmath1018.95 @xmath6 0.54 mas / yr , also in good agreement with the latest studies .
the author wishes to thank c.a .
nelson et al . for making available their simulated white dwarfs and d.i .
dinescu for useful information about coordinate conversions .
we also acknowledge i. king & s. courteau for useful discussions regarding the circular speed curve and p. guhathakurta for pointing out an important oversight in a preliminary version of this paper .
jsk received financial support during this work through an nserc pgs - b graduate student research grant .
hbr is supported in part by the natural sciences and engineering research council of canada .
hbr also extends his appreciation to the killam foundation and the canada council for the award of a canada council killam fellowship .
rmr and is acknowledges support from proposal go-8679 and bmh from a hubble fellowship hf-01120.01 both of which were provided by nasa through a grant from the space telescope science institute which is operated by aura under nasa contract nas5 - 26555 .
bkg acknowledges the support of the australian research council through its large research grant program a00105171 .
alcock , c. , allsman , r. a. , alves , d. r. , axelrod , t. s. , becker , a. c. , bennett , d. p. , cook , k. h. , dalal , n. , drake , a. j. , freeman , k. c. , geha , m. , griest , k. , lehner , m. j. , marshall , s. l. , minniti , d. , nelson , c. a. , peterson , b. a. , popowski , p. , pratt , m. r. , quinn , p. j. , stubbs , c. w. , sutherland , w. , tomaney , a. b. , vandehei , t. , & welch , d. 2000 , , 542 , 281 arendt , r. g. , odegard , n. , weiland , j. l. , sodroski , t. j. , hauser , m. g. , dwek , e. , kelsall , t. , moseley , s. h. , silverberg , r. f. , leisawitz , d. , mitchell , k. , reach , w. t. , & wright , e. l. 1998 , , 508 , 74 richer , h. b. , fahlman , g. g. , ibata , r. a. , pryor , c. , bell , r. a. , bolte , m. , bond , h. e. , harris , w. e. , hesser , j. e. , holland , s. , ivanans , n. , mandushev , g. , stetson , p. b. , & wood , m. a. 1997 , , 484 , 741 williams , r. e. , blacker , b. , dickinson , m. , dixon , w. van dyke , ferguson , h. c. , fruchter , a. s. , giavalisco , m. , gilliland , r. l. , heyer , i. , katsanis , r. , levay , z. , lucas , ray a. , mcelroy , d. b. , petro , l. , postman , m. , adorf , hans - martin . , & hook , r. 1996 , , 112 , 1335 lll field location : & + @xmath83 & ra & = 16@[email protected]@xmath86 + @xmath87 & declination & = @xmath1026@[email protected]@xmath66 + l@xmath89 & galactic longitude & = 351.01@xmath2 + b@xmath89 & galactic latitude & = 15.91@xmath2 + + cluster distance & reddening : & + @xmath90 & apparent distance modulus & = 12.51 + @xmath91 & reddening & = 0.35 + @xmath92 & visual extinction & = 1.33 + @xmath93 & true distance modulus & = 11.18 + @xmath94 & distance from sun & = 1.72 kpc + + background field ( spheroid ) distance & reddening : & + @xmath90 & apparent distance modulus & = 15.71 + @xmath91 & reddening ( all in front of m4 ) & = 0.35 + @xmath92 & visual extinction & = 1.33 + @xmath93 & true distance modulus & = 14.39 + @xmath94 & tangent point distance & = 7.6 kpc + @xmath95 & projected distance from nucleus & = 2.2 kpc + + metallicity ( cluster & spheroid : & + @xmath96 $ ] & heavy metal abundance & = @xmath101.3 + + data ( go 5461 - cycle 4 & go 8679 - cycle 9 ) : & + no . of images f555w - 1995 & 15@xmath972600 seconds + no . of images f606w - 2001 & 98@xmath971300 seconds + no . of images
f814w - 1995 & 9@xmath97800 seconds + no . of images f814w - 2001 & 148@xmath971300 seconds + + limiting magnitude ( based on sextractor classifications ) : & + @xmath35 & 29 + @xmath98 & 27.5 + + wd 1 & disk & 16:23:52.76 & @xmath1026:33:13.07 & 24.60 & 0.30 & 0.77 & 0.77 & @xmath100.05 + wd 2 & disk & 16:23:54.49 & @xmath1026:31:19.38 & 27.17 & 0.43 & 2.56 & 2.33 & 1.06 + wd 3 & disk & 16:23:55.54 & @xmath1026:32:36.64 & 24.18 & 0.59 & 7.47 & @xmath107.27 & @xmath101.71 + wd 4 & disk & 16:23:55.87 & @xmath1026:33:03.28 & 27.43 & 0.63 & 5.31 & @xmath105.22 & 0.96 + wd 5 & disk & 16:23:58.75 & @xmath1026:32:40.13 & 25.22 & 0.84 & 8.28 & @xmath108.26 & 0.51 + wd 6 & disk & 16:24:02.33 & @xmath1026:32:25.55 & 27.98 & 0.83 & 3.54 & 0.88 & 3.43 + wd 7 & disk & 16:24:02.81 & @xmath1026:32:42.61 & 27.82 & 1.01 & 5.70 & @xmath105.37 & @xmath101.91 + wd 8 & disk & 16:24:03.10 & @xmath1026:32:45.10 & 27.99 & 1.19 & 7.99 & @xmath107.41 & @xmath103.01 + wd 9 & spheroid & 16:23:55.28 & @xmath1026:32:47.47 & 27.02 & 0.48 & 16.85 & @xmath1015.48 & 6.66 + wd 10 & spheroid & 16:23:55.87 & @xmath1026:32:37.77 & 28.34 & 0.93 & 14.15 & @xmath1013.69 & 3.57 + wd 11 & spheroid & 16:23:58.12 & @xmath1026:32:10.13 & 27.82 & 0.93 & 15.58 & @xmath1015.31 & 2.89 + wd 12 & disk ? &
16:23:53.25 & @xmath1026:33:10.66 & 28.25 & 1.38 & 6.30 & @xmath104.73 & @xmath104.17 + wd 13 & disk ? & 16:23:59.35 & @xmath1026:31:28.52 & 27.49 & 1.93 & 13.99 & @xmath1013.89 & @xmath101.67 + | we establish an extragalactic , zero - motion frame of reference within the deepest optical image of a globular star cluster , a _ hubble space telescope ( hst ) _
123-orbit exposure of m4 ( go 8679 , cycle 9 ) . the line of sight beyond m4 ( @xmath0 , @xmath1 = 351@xmath2 , 16@xmath2 ) intersects the inner halo ( spheroid ) of our galaxy at a tangent - point distance of 7.6 kpc ( for @xmath3 = 8 kpc ) .
the main sequence of this population can be clearly seen on the color - magnitude diagram ( cmd ) below the m4 main sequence .
we isolate these spheroid stars from the cluster on the basis of their proper motions over the 6-year baseline between these observations and others made at a previous epoch with _ hst _ ( go 5461 , cycle 4 ) .
distant background galaxies are also found on the same sight line by using image - morphology techniques .
this fixed reference frame allows us to determine an independent measurement of the fundamental galactic constant , @xmath4 = @xmath5 = 25.3 @xmath6 2.6 km / s / kpc , thus providing a velocity of the local standard of rest @xmath7 = @xmath8 = 202.7 @xmath6 24.7 km / s for @xmath3 = 8.0 @xmath6 0.5 kpc .
secondly , the galaxies allow a direct measurement of m4 s absolute proper motion , @xmath9= @xmath1012.26 @xmath6 0.54 mas / yr , @xmath11 = @xmath1018.95 @xmath6 0.54 mas / yr , in excellent agreement with recent studies . the clear separation of galaxies from stars in these deep data also allow us to search for inner - halo white dwarfs .
we model the conventional galactic contributions of white dwarfs along our line of sight and predict 7.9 ( thin disk ) , 6.3 ( thick disk ) and 2.2 ( spheroid ) objects to the limiting magnitude at which we can clearly delineate stars from galaxies ( @xmath12 29 ) .
an additional 2.5 objects are expected from a 20% white dwarf dark halo consisting of 0.5 m@xmath13 objects , 70% of which are of the da type .
after considering the kinematics and morphology of the objects in our data set , we find the number of white dwarfs to be consistent with the predictions for each of the conventional populations .
however , we do not find any evidence for dark halo white dwarfs . |
extremal graph theory @xcite deals with graphs in which conflicting graph invariants are on the verge of contradiction .
a classic example due to mantel from 1907 shows that , among graphs of order @xmath0 , as the edge number increases beyond @xmath1 a graph can no longer be bipartite and must contain a triangle . generalizing slightly , the mantel problem is to determine those graphs with fixed edge density @xmath2 which minimize , and those which maximize , the possible values @xmath3 of triangle density . in this vein extremal graph theory
is concerned with qualitative features of graphs with invariants on the boundary @xmath4 of the space @xmath5 of possible values of some particular set of invariants which , for the mantel problem , are the edge and triangle densities , @xmath2 and @xmath3 .
( the set @xmath5 for the mantel problem was finally determined in @xcite , and the optimizing graphs in @xcite . ) in this paper we are concerned with a natural generalization of extremal graph theory to the interior of @xmath5 .
borrowing an idea from physics , it is possible that qualitative graph features which are forced in an absolute sense on a subset @xmath6 of the boundary of @xmath5 are still retained for typical graphs in some phase , a region of @xmath5 abutting @xmath6 .
( we define ` phase ' below and ` typical ' in the next section . ) for instance for the mantel problem there is evidence in @xcite that for edge density less than 1/2 there is a region of @xmath5 abutting the interval @xmath7\times\{0\}$ ] of @xmath4 , in which now a typical graph is nearly bipartite .
( the vertices are divided into two clusters of nearly equal size , with nearly all edges connecting vertices in one cluster to vertices in the other . )
one objective in such a study is ` phase transitions ' , boundaries between phases in which the competition between invariants which has traditionally been studied on @xmath8 is extended into the interior of @xmath5 , and now concerns typical graphs .
we study typical graphs using entropy and the graph limit formalism , which we sketch after the following summary of results .
consider the set @xmath9 of simple graphs @xmath10 with set @xmath11 of ( labeled ) vertices , edge set @xmath12 and triangle set @xmath13 , where the cardinality @xmath14 .
( ` simple ' means the edges are undirected and there are no multiple edges or loops . )
we will be concerned with the asymptotics of @xmath9 as @xmath0 diverges , specifically in the relative number of graphs as a function of the cardinalities @xmath15 and @xmath16 .
let @xmath17 be the number of graphs in @xmath9 such that the edge and triangle densities , @xmath18 and @xmath19 , satisfy : @xmath20 graphs @xmath21 in @xmath22 are known to have edge and triangle densities , @xmath23 , whose accumulation points form a compact subset @xmath24 of the @xmath25-plane bounded by three curves , @xmath26 , the line segment @xmath27 , and a certain scalloped curve @xmath28 , lying above the curve @xmath29 , and meeting it when @xmath30 ; see @xcite and references therein , and figure 1 .
( note the minor shift in emphasis from @xmath5 , as discussed earlier , to the accumulation points @xmath24 of @xmath5 . )
.3truein , outlined in solid lines , title="fig:",width=288 ] we are interested in the relative number of graphs with given numbers of edges and triangles , asymptotically in the number of vertices .
more precisely we will analyze the entropy density , the exponential rate of growth of @xmath31 as a function of @xmath0 .
first consider @xmath32 the limits defining the entropy density @xmath33 are proven to exist in @xcite .
the objects of interest for us are the qualitative features of @xmath33 in the interior of @xmath24 .
in particular , a phase is commonly defined as a maximal connected open subset in which the entropy density is analytic @xcite .
our main result is : [ thm0 ] in the interior of its domain @xmath24 the entropy density @xmath33 satisfies : @xmath34 for some @xmath35 .
therefore for fixed @xmath2 , @xmath33 attains its maximum at @xmath36 but is not differentiable there .
for @xmath37 we have the stronger inequality @xmath38 for some @xmath39 .
so the graph of @xmath33 has its maxima , varying @xmath3 for fixed @xmath2 , on a sharp crease at the curve @xmath40 , and is not concave for @xmath3 just below @xmath41 .
the importance of the result lies in the implication from ( [ diff ] ) of the lack of differentiability of @xmath33 on the crease , and thus the existence of a phase transition , and the implication from ( [ strong ] ) of a lack of concavity of @xmath33 , discussed below .
we begin with a quick review of the formalism of graph limits , as recently developed in @xcite ; see also the recent book @xcite .
the main value of this formalism here is that one can use large deviations on graphs with independent edges @xcite to give an optimization formula for @xmath33 @xcite .
consider the set @xmath42 of all symmetric , measurable functions @xmath43 ^ 2\to g(x , y)\in [ 0,1].\ ] ] think of each axis as a continuous set of vertices of a graph .
for a graph @xmath44 one associates @xmath45 where @xmath46 denotes the smallest integer greater than or equal to @xmath47 . for @xmath48 and simple graph @xmath49
we define @xmath50^\ell } \prod_{(i , j)\in e(h)}g(x_i , x_j)\,dx_1\cdots dx_\ell,\ ] ] where @xmath51 , and note that for a graph @xmath10 , @xmath52 is the density of graph homomorphisms @xmath53 : @xmath54 we define an equivalence relation on @xmath42 as follows : @xmath55 if and only if @xmath56 for every simple graph @xmath49 . elements of @xmath42 are called `` graphons '' , elements of the quotient space @xmath57 are called `` reduced graphons '' , and the class containing @xmath48 is denoted @xmath58 .
equivalent functions in @xmath42 differ by a change of variables in the following sense .
let @xmath59 be the space of measure - preserving maps @xmath60\to [ 0,1]$ ] , and for @xmath61 in @xmath42 and @xmath62 , let @xmath63 .
then @xmath64 if and only if there exist @xmath65 in @xmath59 such that @xmath66 almost everywhere ; see cor . 2.2 in @xcite .
the space @xmath42 is compact with respect to the ` cut metric ' defined as follows .
first , on @xmath67 define : @xmath68}\big| \int_{s\times t}[f(x , y)-g(x , y)]\ , dxdy\big|.\ ] ] then on @xmath57 define the cut metric by : @xmath69 we will use the fact , which follows easily from lemma 4.1 in @xcite , that the cut metric is equivalent to the metric @xmath70 where @xmath71 is a countable set of simple graphs , one from each graph - equivalence class . also note that if each vertex of a finite graph is split into the same number of ` twins ' , each connected to the same vertices , the result stays in the same equivalence class , so for a convergent sequence @xmath72 one may assume @xmath73 .
the following was proven in @xcite .
[ thm1 ] ( @xcite ) for any possible pair @xmath25 , @xmath74 $ ] , where the maximum is over all graphons @xmath21 with @xmath75 and @xmath76 , where @xmath77 ^ 2 } g(x , y ) \ , dxdy , \qquad t(g ) = \int_{[0,1]^3 } g(x , y ) g(y , z ) g(z , x ) \ , dxdydz\ ] ] and the rate function is @xmath78 ^ 2 } i_0(g(x , y ) ) \,dxdy , \hbox { \rm where } i_0(u)= \frac{1}{2 } \left [ u \ln(u ) + ( 1-u)\ln(1-u)\right ] .\ ] ]
fix a graphon @xmath21 with edge density @xmath2 .
we can always write such a graphon as @xmath79 where @xmath80 is the constant function on @xmath81 ^ 2 $ ] with value @xmath82 .
we then compute @xmath83 ^ 2 } { { \delta g}}(x , y)\,dxdy + 3e\int_{[0,1]^3 } { { \delta g}}(x , y ) { { \delta g}}(y , z)\,dxdydz\cr & & + \int_{[0,1]^3 } { { \delta g}}(x , y ) { { \delta g}}(y , z ) { { \delta g}}(z , x)\ , dxdydz.\end{aligned}\ ] ] the first term on the right hand side is zero , since @xmath84 ^ 2 } \delta g(x , y)\ , dxdy = \delta e = 0 $ ] .
if we think of @xmath85 as the integral kernel of the hermitian trace class operator @xmath86 on @xmath87)$ ] , then using the inner product @xmath88 and trace @xmath89 we can rewrite the remaining terms as @xmath90 where @xmath91 is the constant function on @xmath81 $ ] . note that the first term is non - negative . using again the fact that @xmath84 ^ 2 } { { \delta g}}(x , y)\,dxdy=0 $ ] , @xmath92 ^ 2 } [ i_0(e+{{\delta g}}(x , y))-{{\delta g}}(x , y ) i_0'(e ) -i_0(e)]\,dxdy \cr & = & \int_{[0,1]^2 } \frac{i_0(e+{{\delta g}}(x , y))-{{\delta g}}(x , y ) i_0'(e ) -i_0(e)}{{{\delta g}}(x , y)^2 } { { \delta g}}(x , y)^2\,dxdy \cr & \ge & f_-(e ) \int_{[0,1]^2 } { { \delta g}}(x , y)^2\,dxdy , \end{aligned}\ ] ] where @xmath93 and @xmath94 is a positive number less than or equal to @xmath95 .
[ lem ] @xmath96 , with equality if and only if @xmath86 is a rank 1 operator . since @xmath86 is an hermitian trace class operator it has pure discrete spectrum . if @xmath97 are the eigenvalues of @xmath86 , then @xmath98^{3/2}.\ ] ] if @xmath86 has rank one , then @xmath99^{3/2}$ ] .
if @xmath86 has rank bigger than 1 , then @xmath100 is strictly smaller than @xmath101 .
we next give an estimate for @xmath102 when @xmath37 .
if @xmath103 , then @xmath104^{3/2 } \le \left ( \frac{\delta i}{f_-(e)}\right ) ^{3/2}.\ ] ] this implies that @xmath105 using @xmath106 this also implies a linear estimate @xmath107 for @xmath108 . finally , we estimate @xmath102 when @xmath109 . since @xmath110 , and since @xmath111 , we have @xmath112 so @xmath113
we now generalize theorem [ thm0 ] to graph models where we keep track of the number of graph homomorphisms @xmath114 for some fixed graph @xmath49 , not necessarily triangles .
we can compute the entropy of graphs with @xmath115 within @xmath116 of @xmath2 and @xmath117 within @xmath116 of @xmath3 , and define the entropies @xmath118 and @xmath33 exactly as in equation ( [ entropydef ] ) .
the proof of theorem [ thm1 ] carries over almost word - for - word to show the following .
[ thm3 ] for any possible pair @xmath25 , @xmath119 $ ] , where the maximum is over all graphons @xmath21 with @xmath75 and @xmath120 .
note that if @xmath49 has @xmath121 edges the constant graphon @xmath122 satisfies @xmath123 .
[ thm4 ] for fixed @xmath124 the entropy density @xmath33 achieves its maximum at @xmath125 and is not differentiable with respect to @xmath3 at that point .
following the proof of theorem [ thm0 ] , we write @xmath126 and expand both @xmath102 and @xmath127 in terms of @xmath128 .
the estimate ( [ ibelow ] ) still applies .
the only difference is the expansion of @xmath127 .
since @xmath127 is the integral of a polynomial expression in @xmath21 , we can expand @xmath129 as a polynomial in @xmath128 .
this must take the form @xmath130 ^ 2 } h_1(x , y ) \delta g(x , y)\ , dx dy + \int_{[0,1]^4 } h_2(w , x , y , z ) \delta g(w , x ) \delta g(y , z)\ , dw dx dy dz \cr & & + \int_{[0,1]^3 } h_3(x , y , z ) \delta g(x , y ) \delta g(y , z)\ , dxdydz + \cdots,\end{aligned}\ ] ] where the non - negative functions @xmath131 , @xmath132 , @xmath133 , etc .
, are computed from the graphon from which we are perturbing . however , that graphon is a constant @xmath122 , so each function @xmath134 is also a constant .
thus there are non - negative constants @xmath135 , @xmath136 , @xmath137 , such that @xmath138 ^ 2 } \delta g(x , y ) \ , dxdy\ & + & c_2 \int_{[0,1]^4 } \delta g(w , x)\delta g(y , z)\ , dw dx dy dz \cr & + & c_3 \int_{[0,1]^3 } \delta g(x , y ) \delta g(y , z)\ , dx dy dz + \cdots\end{aligned}\ ] ] the first two terms integrate to zero , while any subsequent terms are bounded by a multiple of @xmath139 .
since there are only a finite number of terms , @xmath140 is bounded above by a constant multiple of @xmath139 while @xmath141 is bounded below by a constant multiple of @xmath139 . combining these observations yields the analog of theorem [ thm0 ] , and
we conclude that @xmath33 can not have a 2-sided derivative with respect to @xmath3 at @xmath125 . a more careful analysis of the terms in the sum ( [ bigsum ] )
shows that each term is either positive - definite , is dominated by a positive - definite term , or scales as @xmath142 or higher , implying that the concavity of @xmath33 just below the curve @xmath125 is the same as for the triangle model .
however , this analysis is not needed for the proof of theorem [ thm4 ] and has been omitted . there
do exist some graphs @xmath49 , such as `` @xmath121-stars '' with @xmath121 edges and one vertex on all of them , such that the lowest value of @xmath3 for fixed @xmath2 is on the ` erds - rnyi curve ' , @xmath143 .
for such graphs the analysis of what happens for @xmath103 is moot and @xmath33 may have a 1-sided derivative at @xmath144 .
we return temporarily to the special case in which @xmath49 is a triangle .
note that it has been fundamental to our analysis to use the optimization characterization of @xmath33 of theorem [ thm1 ] ( theorem 3.1 in @xcite ) . treating this as an optimization with constraints
, one might naturally introduce lagrange multipliers @xmath145 and consider the following optimization system , @xmath146;\ e(g)=e;\ t(g)=t,\ ] ] namely maximize @xmath147 for fixed @xmath148 and then adjust @xmath148 to achieve the desired values of @xmath18 and @xmath19 .
the ` free energy density ' @xmath149 is directly related to the normalization in exponential random graph models and basic information in such models is simply obtainable from it @xcite .
it can be considered the legendre transform of @xmath33 , but since the domain of @xmath33 is not convex , the relationship between @xmath33 and @xmath150 must be more complicated than is common for legendre transforms . in particular , although it has been proven ( @xcite ) that @xmath150 has singularities as a function of @xmath148 ( see figure 2 ) it does not seem straightforward to use this to prove singularities in @xmath33 .
this is what necessitated the different approach we have taken here .
we will try to clarify the relationship between @xmath150 and @xmath33 through differences in the optimization characterizations of these quantities .
.4truein , for @xmath151,title="fig:",width=288 ] as one crosses the curve in figure 2 by increasing @xmath152 at fixed @xmath153 , the unique graphon maximizing @xmath154 jumps from lower to higher value of @xmath18 , but still @xmath155 @xcite .
we emphasize that whenever @xmath151 , one is on the erds - rnyi curve @xmath36 indicated in figure 1 @xcite .
this is significant in interpreting the singularities of @xmath33 and @xmath150 .
the singularities or ` transition ' characterized in theorem [ thm0 ] and associated with _
crossing _ the erds - rnyi curve is presumably between graphs of different character but similar densities ; we expect that graphons maximizing @xmath33 , for @xmath156 , are related to those ( discussed below ) for the upper boundary of its domain @xmath24 , while for @xmath157 they are related to those for the lower boundary of @xmath24 .
( the latter are the subject of @xcite . ) on the other hand , the transition in figure 2 , associated with varying @xmath148 , is between graphs of similar character ( independent edges ) but different densities .
this phenomenon is unrelated to the transition of theorem [ thm0 ] , although still associated with the erds - rnyi curve , and which we understand as follows .
assume one optimizes @xmath154 for fixed @xmath148 , where @xmath148 is adjusted so that maximizing graphons @xmath21
satisfy @xmath75 and @xmath76 to match the desired values of @xmath25 in which we are interested for @xmath33 .
it may happen that for special @xmath148 there are also optimizing @xmath21 with other densities , @xmath158 .
this degeneracy is what is occurring precisely for the @xmath148 on the singularity curve of figure 2 .
all such @xmath21 clearly solve the maximization problem for @xmath159 $ ] ; they are appearing together when we fix @xmath148 because the value of @xmath154 happens to be the same for all these @xmath21 , a phenomenon of no particular relevance to the original optimization problem of @xmath33 .
so in this sense degenerate solutions in the lagrange multiplier method can be misleading ; they point to a ` transition ' which is foreign to the maximization problem for @xmath33 .
we next consider other features of the lagrange multiplier method .
one issue of importance to those who study exponential random graph models is that for no @xmath152 is there a maximizer @xmath21 of the free energy density @xmath160 satisfying @xmath161 , though there clearly are such optimizers of the entropy density @xmath162 as we see for instance from figure 1 .
[ thm5 ] for every @xmath152 and every maximizer @xmath21 of @xmath163 , @xmath164 .
suppose the graphon @xmath165 satisfies @xmath166 ^ 3 $ ] and maximizes the free energy @xmath167 for some @xmath153 and @xmath152 .
it follows from theorem 4.2 in @xcite that @xmath168 .
let @xmath122 be the constant graphon with the same edge density as @xmath169 .
since @xmath109 , @xmath170 .
also , @xmath171 , since for given edge density @xmath172 is maximized at @xmath122 .
but then @xmath173 , and @xmath165 is not a maximizer , which is a contradition .
having established in theorem [ thm0 ] a phase transition on the erds - rnyi curve , we consider the forms of the graphons that maximize @xmath33 on each side of the curve .
we previously @xcite determined the optimizing graphons on the lower boundary of the region @xmath24 , including the scalloped curve .
we now compute the optimizing graphons on the upper boundary and on the curve @xmath174 below the erds - rnyi line .
let @xmath177 be the operator on @xmath178 $ ] with integral kernel @xmath21 .
we already know that @xmath179 , with equality if and only if @xmath177 is rank 1 . however , @xmath180 with equality if and only if @xmath181 almost everywhere , i.e. @xmath182 or 1 almost everywhere . combining the two results
, we have that @xmath183 , with equality if and only if two conditions are met : @xmath184 for some positive function @xmath116 , ( i.e. @xmath177 has rank one ) , and @xmath185 is a 01 function , implying that @xmath186 is a 0 - 1 function . by applying a measure - preserving transformation to @xmath81 $ ] we can assume that @xmath116 is the characteristic function of an interval @xmath187 $ ] .
we then compute @xmath188 and @xmath36 . in short ,
each point on the upper boundary for the allowed region in the @xmath2@xmath3 plane is achieved by a unique reduced graphon , namely the equivalence class of the graphon ( [ top ] ) .
when @xmath174 and @xmath189 , the graphon @xmath190 where @xmath191 , maximizes @xmath33 .
furthermore , every maximizing graphon is of the form @xmath192 for some mearsure - preserving transformation @xmath193 .
we use perturbation theory , writing @xmath194 .
when @xmath174 , the @xmath195 derivative @xmath196 is positive for @xmath0 even and zero for @xmath0 odd .
this means that @xmath197/{x^2}$ ] is a convex function of @xmath198 ( since it is a power series in @xmath198 with positive coefficients ) .
this allows us to find a formula for @xmath128 that simultaneously maximizes @xmath199 for fixed @xmath200 , minimizes the positive - definite quadratic term in @xmath129 ( to be zero ) , and minimizes @xmath141 for fixed @xmath200 .
this must therefore be a minimizer of the rate function and a maximizer of the entropy .
we assume throughout that @xmath201 .
now we try to minimize @xmath84 } i[e+c\alpha(x)\alpha(y)]\ , dxdy$ ] . by convexity ,
this is minimized when @xmath210 ^ 2 $ ] is constant , which means that @xmath211 is constant .
since the integral of @xmath116 is zero , we must have @xmath212 on a set of measure 1/2 and @xmath213 on a set of measure 1/2 .
up to measure - preserving automorphism , we can assume that @xmath214 this means that any graphon that minimizes @xmath102 for fixed @xmath174 and fixed @xmath215 must be @xmath58 , up to a measure - preserving transformation .
we proved in theorem [ thm5 ] that maximizing graphons for @xmath33 for @xmath216 can not be found using lagrange multipliers .
we now show that this also applies to certain values of @xmath37 , starting with @xmath174 and @xmath3 close to @xmath217 .
now consider the optimization using lagrange multipliers .
the euler - lagrange equations are : @xmath221 + \beta_1 + \beta_2 h(x , y)=0,\ ] ] where @xmath222 } g(x , z ) g(y , z)\ , dz\ ] ] is the first variation of @xmath19 with respect to @xmath185 . for our @xmath223 ,
this becomes : @xmath224 \cr \beta_1 + 3 \beta_2\left ( \frac{1}{4}+\epsilon^2\right ) & = & i_0'\left ( \frac{1}{2}-\epsilon\right ) = \frac{1}{2}\ln\left[\frac{\frac{1}{2}-\epsilon}{\frac{1}{2}+\epsilon}\right],\end{aligned}\ ] ] which are satisfied if and only if @xmath225.\ ] ] notice that @xmath153 and @xmath152 diverge as @xmath226 ( equivalently , as @xmath227 ) .
however , solutions to the euler - lagrange equations are not necessarily local maxima of @xmath160 .
it is easy to check by differentiation of ( [ halfentropy ] ) that there are @xmath228 such that @xmath229 is strictly concave on @xmath230 but strictly convex on @xmath231 $ ] .
convexity implies that @xmath58 is not a maximizer for @xmath232 for @xmath233 , but is rather a local _ minimizer with respect to variation of @xmath3 , and so there are no @xmath148 which can lead to the maximizers of @xmath234 for @xmath3 just below @xmath217 .
while the precise calculation was done for @xmath174 using equation ( [ halfentropy ] ) , this phenomenon is simply due to inequality ( [ strong ] ) , and actually occurs for all @xmath2 , not just for @xmath174 .
in fact from the proof of theorem [ thm4 ] this phenomenon can be extended to subgraphs @xmath49 other than triangles .
[ however , as noted above , for some @xmath49 the erds - rnyi curve is actually the lower boundary of the domain of the entropy , in which case there are no ` missing ' points below it . ] .2truein * acknowledgements : * we gratefully acknowledge useful discussions with mei yin . .2truein _ | we prove that the asymptotic entropy of large simple graphs , as a function of fixed edge and triangle densities , is nondifferentiable along a certain curve . |
according to the predictions of unified gauge theories , supersymmetry , supergravity , and string theory , there would exist a number of light and massless particles @xcite .
an exchange of such particles between two bodies gives rise to an additional force .
additional fundamental forces at short distances were intensively studied , in particular during last few years following the hypothesis about `` large '' supplementary spatial dimensions proposed by arkami - hamed , dimopoulos , and dvali @xcite , building on earlier ideas in refs @xcite . for a review of theoretical works and recent experimental results ,
see @xcite .
this hypothesis could be verified using neutrons because the absence of an electric charge allows one to strongly suppress the false electromagnetic effects @xcite .
it was noticed in @xcite that the measurement of the neutron quantum states in the earth s gravitational field @xcite is sensitive to such extra forces in the sub - micrometer range . in the case of @xmath0 extra dimensions ,
the characteristic range is just in the nanometer domain @xcite which is accesible in this experiment .
the first attempt to establish a model - dependent boundary in the range 110 @xmath1 m , was presented in ref .
@xcite .
an effective gravitational interaction in presence of an additional yukawa - type force is parametrized as : @xmath2 here , @xmath3 is the newtonian gravitational constant , @xmath4 and @xmath5 are interacting masses , @xmath6 their relative distance , @xmath7 and @xmath8 are strength and characteristic range of this hypothetical interaction .
the experiment @xcite consists in the measurement of the neutron flux through a slit between a horizontal mirror on bottom and a scatterer / absorber on top as a function of the slit size @xmath9 .
this dependence is sensitive to the presence of quantum states of neutrons in the potential well formed by the earth s gravitational field and the mirror . in particular ,
the neutron flux was measured to be equal to zero within the experimental accuracy if the slit size @xmath9 was smaller than the characteristic spatial size ( a quasiclasssical turning point height ) of the lowest quantum state of @xmath10 m in this potential well . the neutron flux at the slit size @xmath11 m was lower by at least a factor of 200 than that for the lowest quantum state(@xmath12 m ) .
if an additional short - range force of sufficiently high strength would act between neutrons and the mirror then it would modify the quantum states parameters : an attractive force would `` compress '' the wave functions towards the mirror , while a repulsive force would shift them up . in this experiment ,
no deviation from the expected values was observed within the experimental accuracy .
this accuracy is defined by the uncertainty in the slit size which can be conservatively estimated as @xmath13 30 % for the lowest quantum state @xcite .
the motion of neutrons in this system over the vertical axis @xmath14 could be considered , within first and quite good approximation , as an one dimentional problem for which the mirror provides an infinitely high potential .
the interaction between neutrons and the earth is described by the first term in equation ( [ intergen ] ) and can be approximated by the usual linear potential ( @xmath15 ) : @xmath16 with @xmath17 , @xmath18 being the earth s radius , @xmath19 its mass , @xmath20 the neutron mass .
the second term in equation ( [ intergen ] ) introduces an additional interaction . due to the short range of this interaction , its main contribution
is provided by the interaction of neutrons with a thin surface layer of the mirror and the scatterer . in this article , we estimate an upper limit on an additional attractive short - range force , which could be established from this experiment in a model - independent way . we show that it could not be significantly improved in any more sophisticated model .
let us estimate firstly the interaction of neutrons with the mirror .
if the mirror s density is constant and equals to @xmath21 , then an additional potential of the interaction between neutrons and the mirror is given by : @xmath22 the volume integral is calculated over the mirror bulk : @xmath23 , @xmath24 ( in fact , over the neutron s vicinity with the size of the order of a few @xmath8 due to the exponential convergency of these integrals ) . it can be calculated analytically for small @xmath8 : @xmath25 with @xmath26 . the simplest upper limit on the strength of an additional interaction follows from the condition that this additional interaction does not create itself any bound state
. it is known @xcite that for an exponential attractive ( @xmath27 ) potential ( [ newint ] ) this means that : @xmath28 this condition gives a boundary for an additional potential strength : @xmath29 @xmath30 being the earth s averaged density . in this experiment , both densities are close to each other @xmath31 , therefore their ratio @xmath32 is close to 1 .
however an adequate choice of the mirror material ( coating ) would easily allow one to gain a factor of 35 in the sensitivity in future experiments .
one obtains the following numerical boundary : @xmath33 here , 1 @xmath1 m is chosen as a natural scale for this experiment .
this limit is presented in figure [ figlimit ] in comparison with the limits coming from the casimir - like and van der waals force measurement experiments @xcite .
one can note that , in the realistic case , one has to establish a condition of non existence of an additional bound state for the sum of ( [ mgz ] ) and ( [ newint ] ) but not for the interaction ( [ newint ] ) alone .
a presence of the linear potential modifies slightly the critical value in ( [ limsimple ] ) .
for instance , for @xmath34 m it equals approximately to 1.0 and for @xmath35 m it equals to 0.74 . for smaller @xmath8 ,
this value is evident to tend to 0.72 .
it is possible to explain qualitatively why the strength of an additional interaction should be higher in presence of the @xmath36-potential than without it .
when a bound state just appears , then its wave function is extremely spread .
if one adds a supplementary `` external '' confining potential , it does not allow the wave function to be spreaded and thus one needs a stronger potential to create a bound state .
the range of presented @xmath8 is 1 nm
10 @xmath1 m .
a deviation from a straight line in the solid curve at 1 nm is due to the finite range of increase of the mirror effective nuclear potential ( impurities on the surface and its roughness ) .
the same effect at @xmath37 m is due to an `` interference '' of the potentials ( [ mgz ] ) and ( [ newint ] ) .
following from the experiment @xcite ( the solid line ) in comparison with that from the measurement of the casimir and the van der waals forces @xcite ( the dashed line ) .
the dotted line shows a limit which can be easily obtained by an improvement of this experiment .
the solid horizontal line represents the limit established from the atomic experiment @xcite .
[ figlimit ] ] it is interesting to compare this analytical limit ( [ statesnumbound ] ) to an analogous expression obtained in @xcite for the casimir force - like experiments .
the simplest boundary @xmath38 following from these experiments is given by a formula : @xmath39 here @xmath40 is a gap separation and @xmath41 is a constant depending on the geometry of different experiments .
this function increases exponentially when @xmath8 tends to zero .
this behaviour is clearly seen in figure [ figlimit ] . in the experiment
@xcite we obtain @xmath42 and @xmath43 increases only as @xmath44 ! this difference between ( [ cas ] ) and ( [ neut ] ) means that , in principle , for any @xmath41 and @xmath45 , one could find a domain of sufficiently small but experimentally observable @xmath8 in which the limit obtained from our experiment is stronger than that obtained from a casimir force - like experiment . up to now , we have not discussed the scatterer .
as we can see from figure [ figlimit ] , a competitive limit could be established only for @xmath8 much smaller than the scatterer roughness amplitude of the order of 1 @xmath1 m .
therefore an influence of the scatterer is negligible .
unfortunately , this experiment does not allow us to establish a competitive limit for a repulsive interaction .
in this case , there could be no `` additional '' bound state . here , instead of the condition of `` non - existence '' of a bound state , one could consider the critical slit size for which the first bound state appears in this system .
such an approach would be model - dependent due to uncertainties in the description of the interaction of neutrons with the scatterer . nevertheless , it is possible to obtain a simple analytical expression for small @xmath8 and to show explicitly a difference in sensitivity of this experiment to an attractive and to a repulsive additional interactions .
let us remind the reader that the quasiclassical approximation for the potential composed by the linear term @xmath36 and by the infinite mirror potential describes the problem very well . for instance , it gives the ground state energy , for which the quasiclassical approximation is usually expected to be not valid , within 1 % @xcite .
the bohr - sommerfeld quantification condition in this case assumes that the quasiclassical wave function @xmath46 with @xmath47 ( @xmath48 , @xmath49 being the turning point ) equals to zero at @xmath50 ( at the inpenetrable mirror ) .
this condition , for @xmath51-th quantum state , means @xmath52 and the corresponding energies are : @xmath53 with an additional short - range repulsive interaction @xmath54 , the problem could be solved in an analogous way .
one can choose such @xmath55 that @xmath56 .
for @xmath57 , the @xmath36 potential is negligible and the problem can be solved analytically @xcite . for @xmath58
, one can use the quasiclassical approximation . at @xmath59 the quasiclassical wave function should be equal to the exact one : @xmath60 with the phase shift @xmath61.\end{aligned}\ ] ] here @xmath62 , @xmath63 , @xmath64 is the euler constant , @xmath65 and @xmath66 are the modified bessel functions of pure imaginary argument .
this phase shift modifies the bohr - sommerfeld quantification condition and introduces a shift @xmath67 in the level s energy .
if this shift is small with respect to the nonperturbated value , one obtains the following relation between the energy shift @xmath67 and the phase shift @xmath68 : @xmath69^{1/2 } = \delta_0 .\end{aligned}\ ] ] for very strong repulsive potential , @xmath70 , the phase shift is equal to @xmath71 an experimental upper limit on the energy shift @xmath67 would impose an upper limit on @xmath7 for a repulsive interaction : @xmath72 with @xmath73 or @xmath74 direct comparison of relation ( [ repulsalph ] ) to ( [ limsimple ] ) and ( [ statesbound ] ) shows that the limit ( [ repulsalph ] ) at small @xmath8 is sufficiently less restrictive than that for an attractive one ( [ statesbound ] ) due to the exponential factor .
note that the presented approach could be applied as well for an attractive interaction @xmath75 . in this case , the phase shift is equal to @xmath76.\end{aligned}\ ] ] here @xmath77 and @xmath78 are the bessel functions .
this expression is maximal in the vicinity of the first bessel function s zero @xmath79 where it can be developped as @xmath80 one thus obtains the relation between the strength of an attractive potential and the energy shift @xmath81 : @xmath82 with @xmath83 , @xmath84 , @xmath85 .
if one omits the second term in the latter equation , one obtains exactly the relations ( [ limsimple ] ) and ( [ statesbound ] ) representing a condition of non - existance of an additional bound state .
let us mention that this formula allows us to estimate the capability of this experiment to further improve an upper limit on @xmath7 in the case of an attractive interaction : in order to improve significantly ( by an order of magnitude ) the limit ( [ statesbound ] ) in a given range of @xmath8 it is necessary to determine experimentally and theoretically the parameters of quantum states with a precision @xmath86 .
the considerations presented above are valid only if the neutron population in the lowest quantum state in such a system ( with an additional interaction included ) if sufficiently high to provide a measurable signal / noise ratio .
as we have mentioned in the introduction , the experiment @xcite would allow one to identify an additional quantum state if its occupation number would not be suppressed by more than a factor of 200 compared to that for other states . in order to calculate the occupation numbers ,
let us start with a general expression for the probability of a rapid transition from a state @xmath87 with the wave function @xmath88 to a state @xmath51 with the wave function @xmath89 which is given by a formula @xcite : @xmath90 for a few initial quantum states , the probability @xmath91 is a sum ( an integral for continius spectrum ) over them : @xmath92 with the occupation numbers @xmath93 of initial states . to obtain an analytical expression for the occupation numbers ,
let us consider a simplified model of a harmonic oscillator in a final state and a plane wave in an initial one .
an explicite analytical shape of the final state wave function does not play a role ( the only important parameter is its spatial size ) and would not modify considerably the occupation numbers .
the used wave functions are equal to @xmath94{\pi x_0 ^ 2 } } \frac{1}{\sqrt{2^n n ! } } \exp \left(-\frac{x^2}{2x_0 ^ 2}\right ) h_n\left ( \frac{x}{x_0}\right).\end{aligned}\ ] ] and @xmath95 here @xmath96 are the hermite polynomials and @xmath97 represents a characteristic geometric size of the problem
. if initial states are populated according to the gaussian law with a chercteristic momentum @xmath98 then @xmath99 and all integrals ( [ fewtran ] ) can be calculated analytically @xcite .
the occupation numbers of the final states appear to be equal to : @xmath100 here @xmath101 is the legendre polinomial .
for instance , for the lowest states with @xmath102 and @xmath103 : @xmath104 if @xmath105 then the occupation numbers are approximately equal for all states : @xmath106 let us apply these expressions to our problem .
for the gravitational quantum states , @xmath107 m ; the vertical velocity distribution has a characteristic velocity of @xmath108 cm / s .
for these states , @xmath109 and all states should have approximately the same occupation numbers .
if an additional bound state were created by the interaction ( [ newint ] ) then the characteristic size of such a state should be of the order of the range of the interaction @xmath8 ( or bigger ) .
for the interaction range , for which this experiment establishes a competitive limit , one obtains @xmath110 for @xmath111 nm and @xmath112 for @xmath113 nm . as we mentioned previously , if such a state exists it would be detected in this experiment .
moreover , these numbers represent a lower estimation for the occupation numbers because they assume the state to be deeply bound . for a just appearing state ,
the characteristic size @xmath97 would be sufficiently bigger than the range of the interaction and would be close to that of the unperturbed states in the gravitational potential . thus the population would not be almost suppressed .
to complete our analysis in the nanometer range , let us note that a competitive limit follows from a recent experiment with antiprotonic atoms as well .
the idea of such an analysis is the same as in ref .
an additional force between a nucleus of mass @xmath19 and an antiproton would change the spectrum of such an atom .
the effective orbit radius @xmath114 for usually studied antiproton - nucleus atoms is about a few hundreeds fm .
therefore , for the range of a hypothetical yukawa force @xmath115 , an additional potential would have a @xmath116 behaviour and would change the coulomb potential : @xmath117 the energy spectrum @xmath118 of an atom with a reduced mass @xmath1 as well as the transition frequences @xmath119 would be modified as well : @xmath120 thus , the constant @xmath7 is related to the shift @xmath121 in the transition frequencies by a formula @xmath122 an upper experimental limit on the transition frequency shift @xmath123 imposes an upper limit on @xmath124 .
the most precise measurement of the energy spectrum of antiprotonic atoms was done for p@xmath125he@xmath126 and p@xmath127he@xmath126 atoms by the asakusa collaboration at the antiproton decelerator at cern @xcite . in this experiment ,
13 electromagnetic transitions between different levels of these atoms were investigated .
no deviation was found from the values expected within the qed calculations @xcite .
equation ( [ alpvisfreq ] ) allows @xmath7 to be extracted .
the averaged over these 13 transitions value of @xmath7 is equal : @xmath128 and is compatible with zero .
a @xmath129 upper limit on @xmath124 from this experiment is : @xmath130 .
an upper limit to an additional attractive force is established from the measurement of quantum states of neutrons in the earth s gravitational field .
relatively high sensitivity of the experiment @xcite to a hypothetical additional force is due to the following factors : firstly , no `` background '' electromagnetic interactions ; secondly , the characteristic size of the neutron wave function in the quantum states fits well to the range of interest for the short - range forces ; finally , non - negligible probability to find neutrons ( quantum - mechanical object ) at distances much closer to the mirror than the average value of 10 @xmath1 m .
however , it can be easily improved in the same kind of experiment with some evident modifications , for instance , one can choose a mirror material ( coating ) with higher density .
a more significant gain in the sensitivity could be achieved in a dedicated neutron experiment which will be presented in a forthcoming paper .
99 see , for instance , h. murayama , g.g .
raffelt , c. hagmann , k. van bibber , and l.j .
rosenberg , in _ review of particle physics , phys .
* d 66 * , 334 ( 2002 ) and references therein .
n. arkani - hamed , s. dimopoulos , and g. dvali , phys .
* b 429 * , 263 ( 1998 ) ; phys . rev .
* d59 * , 086004 ( 1999 ) .
g. carugno , z. fontana , r. onofrio , c. rizzo , phys .
rev . * d55 * , 6591 ( 1997 ) ; g. bressi , g. carugno , a. galavni , r. onofrio , g. ruoso , class . quantum grav . * 17 * , 2365 ( 2000 ) ; g. bressi , g. carugno , a. galavni , r. onofrio , g. ruoso , f. veronese , class . quantum grav . * 18 * , 3943 ( 2001 ) ; g. bressi , g. carugno , a. galavni , r. onofrio , g. ruoso , phys . rev . lett . * 88 * 041804 ( 2002 ) .
nesvizhevsky , h.g .
brner , a.m. gagarski , a.k .
petukhov , g.a .
petrov , h. abele , s. baessler , g. divkovic , f.j .
ruess , t. stferle , a. westphal , a.v .
strelkov , k.v .
protasov , a.yu .
voronin , phys .
* d 67 * , 102002 ( 2003 ) . | an upper limit to non - newtonian attractive forces is obtained from the measurement of quantum states of neutrons in the earth s gravitational field .
this limit improves the existing contraints in the nanometer range . |
_ quandles _ , an algebraic structure associated to oriented knots and links , were introduced by joyce and matveev independently in 1982 in @xcite .
_ racks _ , the analogous structure associated to framed knots and links , were introduced by fenn and rourke in 1992 in @xcite . soon thereafter , _ biracks _ were introduced in @xcite and later the special cases known as _ biquandles _ were studied in papers such as @xcite . a homology and cohomology theory associated to racks
was introduced in @xcite . in @xcite ,
a subcomplex of the rack chain complex was identified in the case when our rack is a quandle , and cocycles in the quotient complex ( known as _ quandle cocycles _ ) were used to enhance the quandle counting invariant , yielding _ cjkls quandle cocycle invariants_. in @xcite , the degenerate subcomplex was generalized to the case of non - quandle racks with finite rack rank , yielding an analogous enhancement of the rack counting invariant via _
@xmath0-reduced cocycles_. in @xcite , yang - baxter ( co)homology was defined as a natural generalization of the quandle ( co)homology for biquandles , but the boundary map was difficult to define combinatorially for arbitrary dimensions , making it impossible to define the degenerate subcomplex in general . nevertheless , in the special case of biquandles , _
reduced 2cocyles _ were defined which allowed enhancement of the biquandle counting invariant , generalizing the cjkls cocycle invariants . in this paper
, we give a reformulation of the birack structure in terms of actions of a set by an _ augmentation group _ generalizing the augmented quandle and augmented rack structures defined in @xcite and @xcite .
our reformulation also restores the original approach taken in @xcite of using the `` sideways operations '' as the primary operations , as opposed to the more usual approach of using the `` direct operations '' as primary .
this approach enables us to define a ( co)homology theory for biracks with a fully combinatorial formula for the boundary map , which we are able to employ to identify the degenerate subcomplex associated to @xmath0-phone cord moves for arbitrary biracks of finite characteristic , generalizing the previous cases of quandle , @xmath0-reduced rack , and yang - baxter ( co)homology . as an application we use @xmath0-reduced augmented birack 2-cocycles to define cocycle enhancements of the birack counting invariant .
the paper is organized as follows . in section [ ab ]
we introduce augmented biracks . in section [ abh ]
we define augmented birack ( co)homology and discuss relationships with previously studied ( co)homology theories .
section [ abci ] deals with augmented birack cocycles and enhancement of the counting invariant . in section [ e ]
we give some examples of the new cocycle invariants and their computation .
we end in section [ q ] with some questions for future research .
[ def1 ] an augmented birack defines a birack map @xmath1 as defined in previous work by setting @xmath2 the @xmath3-actions @xmath4 are the components of the sideways map in the notation of previous papers .
the geometric motivation for augmented biracks come from labeling semiarcs in an oriented framed link diagram with elements of @xmath5 ; each reidemeister move yields a set of necessary and sufficient conditions for labelings before and after the move to correspond bijectively .
@xmath6 the names are chosen so that if we orient a crossing , positive or negative , with the strands oriented upward , then the unbarred actions go left - to - right and the barred actions go right - to - left , with @xmath7 and @xmath8 standing for `` above '' and `` below '' .
thus , @xmath9 is the result of @xmath10 going above @xmath11 left - to - right and @xmath12 is the result of @xmath11 going below @xmath10 from right - to - left .
the element @xmath13 is the _ kink map _ which encodes the change of semiarc labels when going through a positive kink .
@xmath14 in particular , each @xmath15-labeling of a framed oriented knot or link diagram before a framed type i move corresponds to a unique @xmath15-labeling after the move .
if @xmath16 is the identity element in @xmath3 , our augmented birack is an _ augmented biquandle _ ; labelings of an oriented link by an augmented biquandle are independent of framing . axiom ( ii )
is equivalent to the condition that the map @xmath17 defined by @xmath18 is a bijection with inverse @xmath19 note that the condition that the components of @xmath20 are bijective is not sufficient to make @xmath20 bijective ; for instance , if @xmath5 is any abelian group , the map @xmath21 has bijective component maps but is not bijective as a map of pairs .
the maps @xmath22 are the components of the inverse of the sideways map ; we can interpret them as labeling rules going right to left . at negatively oriented crossings ,
the top and bottom labels are switched .
@xmath23 note that @xmath15-labelings of a framed knot or link correpond bijectively before and after both forms of type ii moves : _ direct type ii _
moves where both strands are oriented in the same direction @xmath24 and _ reverse type ii _
moves in which the strands are oriented in opposite directions .
@xmath25 axiom ( iii ) encodes the conditions arising from the reidemeister iii move : @xmath26 thus by construction we have if @xmath27 and @xmath28 are oriented framed links related by oriented framed reidemeister moves and @xmath15 is an augmented birack , then there is a bijection between the set of labelings of @xmath27 by @xmath15 , denoted @xmath29 , and the set of labelings and the set of labelings of @xmath28 by @xmath15 , denoted @xmath30 .
let us now consider the case when @xmath5 is a finite set . for any framed oriented link @xmath27 with @xmath31 crossings ,
there are at most @xmath32 possible @xmath15-labelings of @xmath27 , so @xmath29 is a positive integer - valued invariant of framed oriented links . more generally , if we choose an ordering on the components of a @xmath33-component link @xmath27 , then framings on @xmath27 correspond to elements @xmath34 and we have a @xmath33-dimensional integral lattice of framings of @xmath27 .
if @xmath5 is a finite set , then @xmath3 is a subgroup of the symmetric group @xmath35 ; in particular there is a unique smallest positive integer @xmath0 such that @xmath36 .
this @xmath0 is called the _
characteristic _ or _ birack rank _ of the augmented birack @xmath15 .
the value of @xmath29 is unchanged by _
@xmath0-phone cord moves _ : @xmath37 in particular , framed oriented links which are equivalent by framed oriented reidemeister moves and @xmath0-phone cord moves have the same @xmath29-values , and links which differ only by framing with framing vectors equivalent mod @xmath0 have the same @xmath29-values .
hence , the @xmath33-dimensional lattice of values of @xmath29 is tiled with a @xmath33-dimensional tile of side length @xmath0 .
we can thus obtain an invariant of the unframed link by summing the @xmath29-values over a single tile .
let @xmath38 be an augmented birack .
let @xmath39 $ ] be the free abelian group generated by ordered @xmath31-tuples of elements of @xmath5 and let @xmath40 . for @xmath41
, define maps @xmath42 by @xmath43 and @xmath44 where the @xmath45 indicates that the entry is deleted , i.e. @xmath46 [ thm : boundary ] the map @xmath47 given by @xmath48 is a boundary map ; the map @xmath49 given by @xmath50 is the corresponding coboundary map .
the quotient group @xmath51 is the _
@xmath31th augmented birack homology _ of @xmath38 , and the quotient group @xmath52 is the _
@xmath31th augmented birack cohomology _ of @xmath38 .
to prove theorem [ thm : boundary ] , we will find it convenient to first prove a few key lemmas .
. then @xmath54 .
we compute @xmath55 obtaining the input vector with the entries in the @xmath56th and @xmath57th positions deleted . on the other hand ,
if we first delete the @xmath56th entry , each entry with subscript greater than @xmath56 is now shifted into one lower position ; in particular , @xmath58 is now in the @xmath59st position and we have @xmath60 as required .
[ partial1 ] the map @xmath61 defined by @xmath62 is a boundary map .
if we apply @xmath63 twice , each term with first summation index less than the second summation index is matched by an equal term with first summation index greater than the second summation index but of opposite sign : @xmath64 if @xmath53 we have @xmath65 .
on the one hand , @xmath66 on the other hand , applying @xmath67 first shifts @xmath68 into the @xmath59 position and we have @xmath69 as required . if @xmath53 we have @xmath70 . on the one hand , @xmath71 as above , applying @xmath72 shifts the positions of the entries with indices greater than @xmath56 , and we have @xmath73 as required .
the final lemma depends on the augmented birack axioms . if @xmath53 we have @xmath74 .
we have @xmath75 while again applying @xmath72 first shifts the positions of the entries with indices greater than @xmath56 , and we have @xmath76 and the two are equal after application of the augmented birack axioms . [ partial2 ]
the map @xmath77 defined by @xmath78 is a boundary map . as with @xmath63 , we observe that every term in @xmath79 with @xmath53 is matched by an equal term with @xmath80 but with opposite sign .
( of theorem [ thm : boundary ] ) we must check that @xmath81 .
our lemmas show that each term in the sum with @xmath53 is matched by an equal term with opposite sign with @xmath80 .
we have @xmath82 the sets of @xmath0-degenerate chains form a subcomplex of @xmath83 .
we must show that @xmath84 implies @xmath85 . using linearity
it is enough to prove that @xmath86 is @xmath0-degenerate .
let @xmath87 , we have : @xmath88 \\ & = & \sum_{k=1}^n \partial ( x_1,\dots , x_{j-1},\pi^{k}(x_j),\pi^{k-1}(x_j),x_{j+2},\dots , x_n ) \\\\ & = & \sum_{k=1}^n \left\ { \sum_{i=1}^{j-1 } ( -1)^{i } [ ( x_1,\dots,\widehat{x_{i}},\dots , \pi^k(x_j ) , \pi^{k-1}(x_j ) , x_{j+2 } , \dots , x_n ) \right . \\ & & \left . - ( \beta_{x_i}(x_1 ) , \dots , \beta_{x_i}(x_{i-1 } ) , \widehat{x_i } , \alpha_{x_i } ( x_{i+1}),\dots , \alpha_{x_i}(\pi^k(x_j ) ) , \alpha_{x_i}(\pi^{k-1}(x_j ) ) , \alpha_{x_i}(x_{j+2 } ) , \dots , \alpha_{x_i}(x_n ) ) ] \right\ } \\ \\ & & + \left\{\sum_{k=1}^n \ { ( -1)^j [ ( x_1,\dots , x_{j-1 } , \pi^ { k-1}(x_j ) , x_{j+2 } , \dots , x_n ) \right . \\ & & - \left . ( \beta_{\pi^k(x_j ) } ( x_1 ) , \dots , \beta_{\pi^k(x_j ) } ( x_{j-1 } ) , \alpha_{\pi^k(x_j ) } ( \pi^{k-1}(x_j ) ) , \alpha_{\pi^k(x_j ) } ( x_{i+2 } ) , \dots , \alpha_{\pi^k(x_j ) } ( x_n))\right ] \\ & & + ( -1)^{j+1}\left [ ( x_1,\dots , x_{j-1 } , \pi^k(x_j ) , x_{j+2 } , \dots , x_n ) \right . \\ & & - \left .
\left.(\beta_{\pi^{k-1}(x_j ) } ( x_1 ) , \dots , \beta_{\pi^{k-1}(x_j ) } ( x_{j-1 } ) , \beta_{\pi^{k-1}(x_j ) } ( x_{j } ) , \alpha_{\pi^{k-1}(x_j ) } ( x_{j+2 } ) , \dots , \alpha_{\pi^{k-1}(x_j ) } ( x_n ) ) \right ] \right\ } \\ \\ & & + \sum_{k=1}^n \left\{\sum_{i = j+2}^{n } ( -1)^{i } \left [ ( x_1,\dots , \pi^{k}(x_j ) , \pi^{k-1}(x_j ) , \dots , \widehat{x_i},\dots , x_n ) \right .
\right.\\ & & \left.\left .
- ( \beta_{x_i}(x_1 ) , \dots , \beta_{x_i}(x_{j-1 } ) , \beta_{x_i}(\pi^k(x_j ) ) , \beta_{x_i}(\pi^{k-1}(x_j ) ) , \dots , \widehat{x_i } , \alpha_{x_i}(x_{i+1 } ) , \dots , \alpha_{x_i}(x_n ) ) \right ] \right\ } \hfill \quad\quad\quad(1)\ ] ] where as usual @xmath89 means @xmath90 .
now the rest of the proof is based on the following two facts : ( 1 ) @xmath91 and ( 2 ) @xmath92 which is obtained by induction from axiom ( i ) in the definition [ def1 ] ( of augmented birack ) .
the following sum vanishes : @xmath93 \right . \\ & & - \left [ ( x_1,\dots , x_{j-1 } , \pi^k(x_j ) , x_{j+2 } , \dots , x_n ) \right . \\ & & - \left .
\left.(\beta_{\pi^{k-1}(x_j ) } ( x_1 ) , \dots , \beta_{\pi^{k-1}(x_j ) } ( x_{j-1 } ) , \beta_{\pi^{k-1}(x_j ) } ( x_{j } ) , \alpha_{\pi^{k-1}(x_j ) } ( x_{j+2 } ) , \dots , \alpha_{\pi^{k-1}(x_j ) } ( x_n ) ) \right ] \right\}\end{aligned}\ ] ] because @xmath92 and @xmath91 .
the rest of the sums can be written as combination of degenerate elements as in the proof of theorem 2 in @xcite by the authors .
in this section we will use augmented birack cocycles to enhance the augmented birack counting invariant analogously to previous work .
let @xmath94 be an oriented framed link diagram with framing vector @xmath95 and a labeling @xmath96 by an augmented birack @xmath15 of characteristic @xmath0 .
for a choice of @xmath97 , we define an integer - valued signature of the labeling called a _
boltzmann weight _ by adding contributions from each crossing as pictured below .
orienting the crossing so that both strands are oriented upward , each crossing contributes @xmath98 evaluated on the pair of labels on the left side of the crossing with the understrand label listed first .
@xmath99 then as we can easily verify , the boltzmann weight @xmath100 is unchanged by framed oriented reidemeister moves and @xmath0-phone cord moves .
starting with move iii , note that @xmath101 implies that @xmath102 and in particular we have @xmath103 then both sides of the reidemeister iii move contribute the same amount to the boltzmann weight : @xmath104 both sides of both type ii moves contribute @xmath105 to the bolztmann weight : @xmath106 @xmath107 similarly , both sides of the framed type i moves contribute zero ; here we use the alternate form of the framed type i move for clarity , with @xmath108 : @xmath109 finally , the @xmath0-phone cord move contributes a degenerate @xmath0-chain : @xmath110 putting it all together , we have our main result : let @xmath27 be an oriented unframed link of @xmath33 components and @xmath15 a finite augmented birack of characteristic @xmath0 . for each @xmath111 ,
the multiset @xmath112 and polynomial @xmath113 defined by @xmath114 and @xmath115 are invariants of @xmath27 known as the _ augmented birack 2-cocycle invariants _ of @xmath27 . as in quandle homology , we have let @xmath15 be an augmented birack . if @xmath116 is a coboundary , then for any @xmath15-labeling @xmath117 of a framed link @xmath94 the boltzmann weight @xmath118 .
if @xmath116 is a coboundary , then there is a map @xmath119 such that @xmath120 .
then for any @xmath121 we have @xmath122 and the boltzmann weight can be pictured at a crossing as below .
@xmath123 in particular , every semiarc labeled @xmath11 contributes a @xmath124 at its tail and a @xmath125 at its head , so each semiarc contributes zero to the boltzmann weight .
cohomologous cocycles define the same @xmath113 and @xmath112 invariants .
in this section we collect a few examples of the augmented birack cocycle invariants and their computation .
in this section we collect a few open questions for future research . in the case of quandle homology , many results are known involving the long exact sequence , the delayed fibonacci sequence in the dimensions of the homology groups for certain quandles , etc . which of these results extend to augmented birack homology ?
in @xcite , yang - baxter ( co)homology was paired with @xmath20-(co)homology to define an enhancement of the virtual biquandle counting invariant .
a future paper will consider the relationship between augmented birack ( co)homology and @xmath20-cohomology .
department of mathematics + louisiana state university + baton rouge , la 70803 - 4918 , + + department of mathematics , + university of south florida , + 4202 e fowler ave . , + tampa , fl 33620 + + and + + department of mathematics , + claremont mckenna college , + 850 columbia ave . , + claremont , ca 91711 | we introduce augmented biracks and define a ( co)homology theory associated to augmented biracks .
the new homology theory extends the previously studied yang - baxter homology with a combinatorial formulation for the boundary map and specializes to @xmath0-reduced rack homology when the birack is a rack .
we introduce augmented birack 2-cocycle invariants of classical and virtual knots and links and provide examples .
keywords : biracks , rack homology , enhancements of counting invariants , cocycle invariants 2010 msc : 57m27 , 57m25 |
after more than 25 years of considerable theoretical and experimental efforts , it appears that qcd is the theory of strong interactions .
ideally , in high - energy qcd one needs one single piece of information from the experiments : the value of @xmath0 . starting from that measured value , every observable can be computed from first principles . in practice
this is not feasible , since we do nt know how to perform calculations in terms of the hadrons that experiments measure in their detectors .
perturbation theory offers a viable way out , since it allows to prove , at least formally , the so - called factorization theorems .
these give explicit prescriptions to write physical observables as the convolution of short- and long - distance parts , up to terms suppressed by the power of some large scale .
we can imagine this factorization to occur at an arbitrary scale @xmath1 ; with a suitable choice of @xmath1 the short - distance pieces , which are entirely expressed in terms of quarks and gluons , are perturbatively calculable .
the long - distance pieces ( such as parton densities ) can not be computed in perturbation theory , but their dependence on @xmath1 can .
furthermore , they are universal , which means that they do nt depend on short - distance physics , but solely on the nature of the hadrons involved , which is a key factor for perturbative qcd to have predictive power .
pending a general solution of qcd , the computing framework based on perturbation theory may be regarded as a hypothesis , which needs to be supported , or disproved , by experimental observations .
countless tests have indeed been successful , convincing us of the correctness of this approach and of the capability of qcd to describe strong interactions ; in many areas precise measurements , rather than tests , are being carried out .
this success may give to the non - expert the impression that current efforts in theoretical qcd are perhaps technically appealing , but not compelling physics - wise .
to counter this view , it is worth reminding that in the past decade the studies of several technically - involved problems , such as computations to next - to - leading order accuracy , resummation techniques , and monte carlo simulations , have been key factors to the outstanding achievements of lep , slc , hera , and tevatron .
however , the solutions devised so far are not sufficient any longer . for an improvement of the accuracy in the extraction of @xmath0 , for a deeper understanding of the interplay between perturbative and non - perturbative physics , and for a realistic modelling of tevatron run ii and lhc physics , new ideas and computations
. it must be clear that such investigations are not only relevant to the study of qcd itself , but also to a variety of other issues , from sm precision tests to searches of beyond - the - sm physics .
besides , a few unsatisfactory results remain in qcd , which deserve further studies . needless to say , this review can not be complete , and i ll have to leave out several interesting results , such as new nlo calculations , progress in small-@xmath2 physics , diffraction , fully numerical computations , and spin physics . also , i ll not present most of the recent experimental results in high - energy qcd , since they can be found in k. long s writeup @xcite .
i d rather use a few phenomenological examples to discuss some theoretical advancements , and open problems .
related topics can be found in z. bern s writeup @xcite .
i ll quote papers submitted to this conference as [ s - nnn ] , s and nnn being the session and paper numbers respectively .
heavy flavour production is one of the most extensively studied topics in qcd .
an impressive amount of data is available , for basically all kinds of colliding particles .
the non - vanishing quark mass allows the definition of open - heavy - quark cross sections ( whereas for light quarks one must convolute the short - distance cross sections with fragmentation functions , in order to cancel final - state collinear divergences ) . on the other hand ,
the presence of the mass makes the calculation of the matrix elements more involved . the breakthrough was the computation of total @xmath3 hadroproduction rates to nlo accuracy @xcite , readily extended to other production processes and more exclusive final states @xcite .
the resummation of various classes of large logarithms affecting these fixed - order computations , such as threshold , large-@xmath4 , and small-@xmath2 logs , has also been accomplished , typically at the next - to - leading log ( nll ) accuracy . in those kinematical regions not affected by large logs , the mass of the heavy quark sets the hard scale .
furthermore , the impact of effects of non - perturbative origin ( such as colour drag or intrinsic @xmath5 ) is known to be larger the smaller the quark mass and cm energy .
thus , top physics is expected to be the ideal testing ground for perturbative computations .
rate at the tevatron.,title="fig : " ] -2.2pc the agreement between nlo results @xcite ( dashed lines the band is the spread of the prediction due to scale variation ) , and tevatron run i data @xcite , shown in fig . [
fig : topattev ] for total @xmath6 rates , appears in fact to be satisfactory .
the inclusion of soft - gluon effects ( solid lines ) , resummed to nll accuracy according to the computation of ref .
@xcite , is seen to increase only marginally the nlo prediction , while sizably reducing the scale uncertainty .
top production appears therefore under perturbative control .
more stringent tests will be performed in run ii : the errors on mass and rate will be smaller , and measurements will be performed of more exclusive @xmath6 observables and of single - top cross section ( for which fully differential nlo results are now available @xcite ) .
bottom quarks are copiously produced at colliders , and precise data for single - inclusive distributions have been available for a long time .
it is well known ( see ref .
@xcite for a review ) that nlo predictions are about a factor of two smaller than data at the sps and at the tevatron ( on the other hand , the shape of the @xmath4 spectrum of the centrally - produced @xmath7 is fairly well described by qcd ) . in a recently - published measurement @xcite of the @xmath8 @xmath4 spectrum ,
cdf find that the average data / theory ratio is @xmath9 .
however , this worrisome result is largely due to an improper computation of the nlo cross section .
let me remind that the spectrum of a @xmath7-flavoured meson @xmath10 is computed as follows : = dzd_td(z ; ) ( p_t - z_t ) , [ btobsec ] where @xmath11 ( @xmath12 ) is the transverse momentum of @xmath7 ( @xmath10 ) , @xmath13 is the cross section for open-@xmath7 production , and @xmath14 is the non - perturbative fragmentation function ( npff ) , which describes the @xmath15 fragmentation .
npff is not calculable from first principles , and the free parameter it contains ( @xmath16 ) is fitted to data after assuming a functional form in @xmath17 ( such as peterson @xcite , kartvelishvili @xcite , etc ) .
this fit is typically performed using eq .
( [ btobsec ] ) , identifying the l.h.s . with @xmath18 data .
it follows that the value of @xmath16 is strictly correlated to the short - distance cross section @xmath13 used in the fitting procedure , and thus is _ non - physical_. when eq .
( [ btobsec ] ) is used to predict @xmath10-meson cross sections ,
it is therefore mandatory to make consistent choices for @xmath16 and @xmath13 . this has not been done in ref .
@xcite : for @xmath13 , the nlo result of ref .
@xcite is used , but the value of @xmath16 adopted ( 0.006 ) has been derived in the context of a lo , rather than nlo , computation . on the other hand , if a more appropriate value of @xmath16 is chosen ( @xmath190.002 @xcite ) , the theoretical prediction increases by a mere 20% @xcite , still rather far from the data .
there are , however , a couple of observations which save the day .
first , one has to remark that in the upper end of the @xmath4 range probed by cdf ( @xmath20 gev ) , large-@xmath21 effects may start to show up .
therefore , fonll computations @xcite should be used rather than nlo ones .
the fonll formalism consistently combines ( i.e. , avoids overcounting ) the nlo result with the cross section in which @xmath22 logs are resummed to nll accuracy ( such resummed cross section is sometimes incorrectly referred to as `` massless '' ) .
thus , fonll can describe both the small-@xmath4 ( @xmath23 , where resummed results do nt make sense ) and the large-@xmath4 ( @xmath24 , where nlo results are not reliable ) regimes . the second observation concerns again the npff : @xmath25 is a rather steeply falling function , and one can approximate it with @xmath26 in the whole @xmath11 range ; then ( from eq .
( [ btobsec ] ) ) @xmath27 , where @xmath28 is the @xmath29 mellin moment of the npff .
this fact has been noticed some time ago @xcite , and @xmath30 is seen to approximate the exact result fairly well @xcite . since @xmath315 ( at the tevatron ) , it follows that , in order to have an accurate prediction for the @xmath4 spectrum in hadroproduction , it is mandatory that the first few mellin moments computed with @xmath32 agreed with those measured . in ref .
@xcite , it is pointed out that this is not the case , in spite of the fact that the prediction for the inclusive @xmath7 cross section in @xmath18 collisions , obtained with _ the same _ @xmath32 , displays an excellent agreement with the data .
there may seem to be a contradiction in this statement : if the shape is reproduced well , why this is not true for mellin moments ?
the reason is that when fitting @xmath32 one excludes the region of large @xmath17 , since it is affected by sudakov logs , and by complex non - perturbative effects which are unlikely to be described by the npff . on the other hand ,
the large-@xmath17 region is important for the computation of @xmath33 ( because of the factor @xmath34 in the integrand ) .
therefore , for the purpose of predicting @xmath10-meson spectra at colliders , ref .
@xcite advocates the procedure of fitting the npff directly in the @xmath35-space .
a fit to the second moment ( denoted as @xmath36 fit henceforth ) is found to fit well all the @xmath33 s for @xmath35 up to 10 ; and although kartvelishvili s form is used , peterson s gives comparable results .
data vs theory @xcite.,title="fig : " ] -2.2pc using the fonll computation , and a @xmath36 fit for the npff , the average data / theory ratio reduces to @xmath37 @xcite . taking the scale uncertainty into account , @xmath8 data appear to be compatible with qcd predictions ( see fig . [
fig : bpluscn ] ) .
if one wants to avoid the pitfalls of npff s , an alternative possibility consists in considering @xmath7-jets rather than @xmath10 mesons , since in this case the npff simply does nt enter the cross section .
the comparison between nlo predictions for @xmath7-jets @xcite and d0 measurements @xcite is indeed satisfactory : data are consistent with theory in the range @xmath38 gev .
overall , one can conclude that @xmath7 data at the tevatron are reasonably described by nlo qcd .
it is worth mentioning that some existing results , presented in terms of @xmath7-quark cross sections , are likely affected by the findings of ref .
@xcite , and need to be reconsidered . among the various mechanisms which can further increase the theoretical predictions , small-@xmath2 @xcite and threshold resummations @xcite will probably play a secondary role wrt nnlo contributions , which are expected to be large given the size of the k - factor at the nlo .
i now turn to the case of charm production .
a thorough discussion on this topic is beyond the scope of this review , and i ll only give the briefest of the summaries ( which will not do any justice to the field ) .
lep data for total rates are in agreement with nlo qcd predictions @xcite ; the shapes of single - inclusive @xmath39 spectra in @xmath40 collisions are as predicted by nlo qcd @xcite , whereas normalization is off by a factor 1.52 , but still consistent with qcd when theoretical uncertainties are taken into account .
the vast majority of fixed target hadro- and photoproduction data are well described by nlo computations , but only if predictions for single - inclusive distributions and correlations are supplemented by some parametrizations of non - perturbative phenomena ( such as intrinsic @xmath5 ) . at hera ,
dis data are in agreement with nlo qcd results @xcite . in photoproduction
, some concerns have arisen in the past because of the discrepancy between zeus and h1 measurements in the comparison with theory : h1 @xcite appears to be in perfect agreement with qcd , whereas zeus @xcite is at places ( for intermediate @xmath4 s and large @xmath41 s ) incompatible with nlo predictions .
zeus have submitted to this conference [ 5 - 786 ] data with unprecedented coverage at large @xmath4 .
the comparison to fonll predictions @xcite , shown in fig .
[ fig : zeusdstar ] , appears to be satisfactory , although data are marginally harder than theory .
the agreement improves if a @xmath36 fit for the npff is adopted ( this result is still preliminary ) .
data in @xmath42 collisions vs fonll predictions , title="fig : " ] -2.2pc let me finally mention the increasing amount of measurements for @xmath7 rates from fixed - target @xcite , hera @xcite , [ 5 - 783,5 - 784,5 - 785,5 - 1013,5 - 1014 ] , and lep @xcite , [ 5 - 366,5 - 475 ] experiments .
a summary of the situation , in the form of ratios data / nlo qcd , is presented in fig .
[ fig : brates ] .
while the fixed - target measurements are in overall agreement with qcd , hera and lep measurements are largely incompatible with theory .
i find this hard to reconcile with the results presented so far , and the size of the discrepancy also makes it problematic to find an explanation in terms of beyond - the - sm physics , let alone higher orders in qcd .
rates.,title="fig : " ] -2.2pc it is necessary to note that in many cases the experimental results are extrapolated to the full phase space from a rather narrow visible region .
it is encouraging to note that in the cases of the recent zeus measurements in dis [ 5 - 783 ] and photoproduction [ 5 - 785 ] ( full boxes in fig . [ fig : brates ] ) , results for single - inclusive distributions are given too , which are seen to be fully compatible with the corresponding nlo predictions .
the necessity of understanding long - distance effects in final - state measurements is not peculiar to @xmath7 hadroproduction .
the hadron - parton duality assumption states that there is a class of observables ( such as jet variables or event shapes in @xmath18 collisions ) whose description in terms of quarks and gluons is expected to reproduce the data , up to terms suppressed by some inverse power of the hard scale @xmath43 of the process ( power corrections ) .
these terms are usually estimated by comparing the parton- and hadron - level predictions of monte carlo ( mc ) generators .
this procedure is not really satisfactory , since mc parameters are tuned to data ( which creates a bias on the `` predictions '' for power - suppressed effects ) , and since the definition of parton- and hadron - level is far from being straightforward .
it is remarkable that we can get insight on non - perturbative physics from perturbative considerations .
the perturbative series , being asymptotic , can be summed to all orders only after defining a summation procedure ( in a rather arbitrary manner ) ; one assumes that this technical manipulation mimics the role played in nature by non - perturbative effects , which are necessary for qcd to be self - consistent .
the summation procedure must eliminate the divergence of the perturbative series , but some finite quantities are left unconstrained .
thus , the idea is to use the ambiguities of the summation procedure to study non - perturbative effects .
although the regularization of the divergence can be technically very complicated , it can always be seen as a prescription to deal with the landau pole of @xmath0 . the idea of ref .
@xcite ( dmw from now on ) is to bypass such a prescription by _ defining _ @xmath0 in the infrared ( ir ) region , assuming its universality ; thus , @xmath0 should effectively measure confinement effects in inclusive quantities ( in order to use such an @xmath0 in actual computations , it is also necessary to assume that the concepts of quarks and gluons still make sense in the ir ) . after giving the gluon
a fake ( `` trigger '' ) mass @xmath1 , the ( fully inclusive ) observable under study is computed in perturbation theory ; the small-@xmath1 behaviour determines the power @xmath44 of the leading power - suppressed term @xmath45 .
the coefficient @xmath46 can not be computed , but can be expressed as an integral of @xmath47 over @xmath48 , with @xmath49 .
since all the ( non computable ) power - correction effects are contained in this coefficient , one effectively gets a factorization formula .
a class of observables of great physical relevance is that of event shapes for which @xmath50 ; their mean values can be computed with the dmw approach .
the property of @xmath0 universality is _ formally _ even more far - reaching ; for example , it implies that the non - perturbative effects it describes exponentiate for observables that do so @xcite .
this offers the possibility of studying not only mean values , but distributions .
the results of dmw for mean values can then be recovered by an expansion of the ( sudakov ) exponent and subsequent average .
more precisely , if @xmath51 denotes the observable , the sudakov is expanded in the region @xmath52 , and only the first non - trivial term in the expansion is kept . in this context
, factorization derives from the hypothesis of @xmath0 ir universality , which is an effective description of long - distance effects , and as such must be insensitive to the details of parton dynamics .
for example , these details are irrelevant when one sums inclusively over the decay products of any parton branchings .
it has been observed @xcite that inclusiveness is lost to a certain extent when recoil effects ( i.e. , higher orders ) are considered , and thus factorization breaks down in this case .
one may still insist that factorization holds , and devise a procedure to systematically account for those effects which would break it in a naive treatment @xcite .
the results for mean values and distributions of event shapes can be written as follows : & & t= t_pert + c_tp , [ evmean ] + & & ( t)= .|_pert(t - c_tp ) , [ evdistr ] + & & p= , [ pshift ] + & & _ i_0(_i)= _ 0^_idk(k ) . here , @xmath53 is a computable coefficient , `` pert '' means perturbatively - computed , and @xmath54 includes the two - loop results of refs .
@xcite ; in principle , @xmath54 can depend upon @xmath51 , but ( accidentally ) it does not .
the @xmath55 and @xmath56 terms in eq .
( [ pshift ] ) are there to avoid double counting with the perturbatively - computed part , which shows again that long- and short - distance effects are correlated . since no small parameter is involved in the two - loop computation of @xmath54 , one may wonder whether factorization could be spoiled beyond two loops .
although it is _ argued _ that this is not the case @xcite , a comparison with the data is mandatory .
also notice that the formulae above need to be modified when more complicated kinematical effects have to be described , as in the case of broadenings @xcite .
( [ evmean])([pshift ] ) are used to fit the data in terms of @xmath55 and @xmath57 .
updated analyses relevant to @xmath18 collisions have been presented to this conference [ 5 - 228,5 - 229,5 - 389 ] , @xcite see also @xcite .
results obtained from mean values are satisfactory , with comparable @xmath58 s obtained from different observables , and @xmath59 values fairly consistent with the world average .
the situation worsens in the case of distributions : @xmath58 universality holds at @xmath1925% level ( 12 @xmath60 ) , and @xmath59 values are systematically lower than the world average , especially for observables such as wide broadening @xmath61 and heavy jet mass @xmath62 .
this fact is disturbing since the same data for distributions , with hadronization effects described by mc s , return @xmath59 values in much better agreement with the world average . from event shapes in @xmath18 collisions.,title="fig : " ] -2.2pc
these findings are summarized in fig .
[ fig : asres ] , where the results obtained with the dmw approach are shown as boxes and crosses for mean values and distributions respectively ; the world average @xmath63 @xcite is also shown .
data have been taken in the petra , lep and lep2 energy range , and analyses have been presented by aleph [ 5 - 296 ] , delphi [ 5 - 228,5 - 229 ] , jade [ 5 - 389 ] @xcite , l3 [ 5 - 495 ] , and opal [ 5 - 368 ] .
one of the two results presented in ref .
@xcite , reported as the uppermost cross in fig .
[ fig : asres ] , is obtained by excluding @xmath61 from the fit .
recently , the nll resummation of many dis event shapes has been achieved ( @xcite , and references therein ) .
power - correction effects can then be studied similarly to what done for @xmath18 collisions , and the results are intriguing .
[ fig : asazdis ] ( taken from ref .
@xcite ) presents the @xmath59 and @xmath58 values obtained from fitting event shape _ distributions _ in dis , and event shape _ means _ in @xmath18 .
the dis and @xmath18 results are largely consistent , which implies , in view of what shown in fig .
[ fig : asres ] , that event shape distributions in @xmath18 and dis prefer different values for @xmath59 . and @xmath58 from dis and @xmath18 event shapes . from @xcite.,title="fig : " ] -2.2pc although not compelling from the statistical point of view , these results for event shape distributions in the dmw approach may hint to the necessity of a more complete description of hadronization effects .
as mentioned before , eq .
( [ evdistr ] ) results from keeping the first non - trivial term in a taylor expansion . if no expansion is made , from rather general factorization arguments in the two - jet limit @xmath64 one
gets the following formula @xcite ( t)= _ 0^tq df_t ( ) .|_pert ( t-/q ) , [ evshape ] where @xmath65 is known as _ shape function_. dmw formulae are recovered with @xmath66 ; in the general case , the first mellin moment of @xmath67 has the same meaning as @xmath58 of dmw . with eq .
( [ evshape ] ) it is not necessary to assume that @xmath68 ( all terms @xmath69 are now expected to be included ) , and therefore the fit ranges can be extended .
similarly to @xmath58 , @xmath67 can not be computed from first principles ; thus , in order not to lose predictive power , a functional form depending on a small number of parameters must be assumed @xcite , keeping in mind that qcd dynamics and lorentz invariance considerations @xcite can be used to severely constrain the form of a more general shape function , from which @xmath67 is derived , independently of phenomenological arguments . in a different approach ( dge @xcite ) , whose final result has the same form as eq .
( [ evshape ] ) , it is suggested to combine sudakov and renormalon resummations in a single formalism .
a renormalon ambiguity appears in the exponent , and the prescription necessary to resolve it can be naturally formulated in terms of a shape function , automatically constraining its functional form .
although the @xmath67 which one gets in dge is consistent with the one obtained in refs .
@xcite , it has to be stressed that the perturbative result @xmath70 in eq .
( [ evshape ] ) is different in the two formalisms , since dge includes a class of subleading logs of renormalon origin . in @xmath18 physics ,
( [ evshape ] ) gives satisfactory results : a good fit to the second moments of 1-thrust , @xmath62 , and @xmath71 parameter is obtained in ref .
@xcite , and the fits for thrust and @xmath62 of ref .
@xcite are in better agreement in the @xmath72 plane wrt those obtained with dmw . on the other hand , according to ref .
@xcite , @xmath73 ( the theory error is estimated to be around 5% ) . therefore , it seems that a more refined treatment of non - perturbative effects , which is helpful in other respects , is not what one needs in order to get larger @xmath0 values .
in a couple of interesting analyses , delphi adopted rather unconventional methods to study event shapes . in ref .
@xcite , event shape distributions were compared to fixed - order @xmath74 ( nlo ) results ( i.e. , resummation has not been included ) , using the renormalization scale as a free parameter in the fit , and correcting for hadronization effects with mc s .
although i m not aware of any theoretical consideration which justifies such a procedure ( called in ref .
@xcite `` experimental optimization of the scale '' ) , the @xmath0 values obtained from different observables display a remarkable consistency , and they are also consistent with those obtained by using nll - resummed predictions . in [ 5 - 228 ] ( see also ref .
@xcite ) event shape means were shown to give mutually consistent @xmath0 values in the context of a renormalization group approach ( rgi @xcite ) , without needing any hadronization corrections .
the final @xmath0 results of refs . @xcite and [ 5 - 228 ] are in excellent agreement with the world average .
i interpret these findings as the indication that ( at least in a given cm energy range ) the uncertainties affecting theoretical predictions at the nlo are larger than or of the same order as the power - suppressed effects that one aims to study .
it thus appears that the computation of event shapes at the nnlo is necessary for a deeper understanding of this matter . in summary , in the past few years a solid progress has been achieved in the understanding of power - suppressed effects in @xmath18 collisions and in dis .
although models such as dmw can successfully describe many features of the data , some aspects deserve further studies . in some cases ,
improvements are obtained within approaches which refine the treatment of the non - perturbative part , using a shape function , but the computation of the next order in perturbation theory will likely be necessary in order to obtain a more consistent overall picture .
it is worth recalling that the study of hadron - mass effects has been found @xcite to induce further power - suppressed terms , some of which can be eliminated by adopting a suitable definition for the observables ( e - scheme ) . furthermore , more stringent tests of the models for power - suppressed effects should be performed using observables with more complicated kinematic structure and/or gluons at the lo ( see refs . @xcite and references therein ) . finally , it appears to be mandatory to extend the studies of such models to the case of hadronic collisions ( for jet observables in particular ) , where little work has been done so far .
bottom production at the tevatron and event shapes in @xmath18 collisions are a couple of examples which provide physical motivations to increase the precision of the perturbative computations .
if tree @xmath75-point functions contribute to a given reaction at the lo , the n@xmath76lo result ( i.e. , of relative order @xmath77 wrt to the lo ) will get contributions from the @xmath78-loop , @xmath79-point functions , with @xmath80 .
there are basically three major steps to make in order to get physical predictions : _ i _ ) explicit computation of all the @xmath78-loop , @xmath79-point functions ; _ ii _ ) cancellation of soft and collinear divergences ( which i ll denote improperly as ir cancellation henceforth ) ; _ iii _ ) numerical integration of the _ finite _ result obtained from _ i _ ) and _ ii _ ) , with mc techniques to allow more flexibility .
one should also mention that uv renormalization has in general to be carried out ; however , this is basically textbook matter by now , and thus i ll not deal with it in the following . for nlo computations ( @xmath81 ) , steps _
i)__iii ) _ appear to be understood .
one - loop integrals have been computed up to five external legs @xcite ; the case of @xmath82 can not probably be handled with feynman - diagram techniques only , and still awaits for a general solution ( see @xcite ) .
subtraction @xcite and slicing @xcite methods , to achieve ir cancellation with semi - analytical techniques , have been available for a long time . in their modern versions
@xcite they are formulated in an universal ( i.e. , process- and @xmath75-independent ) way , which simplifies step _ iii ) _ considerably , and allows the computation of any ir safe observable ( no matter how exclusive ) .
no modification is needed in order to incorporate new one - loop results .
attempts to achieve ir cancellation through full numerical computations @xcite are still in a preliminary stage , reproducing known results for three - jet production in @xmath18 collisions . only a handful of production processes have been computed to nnlo accuracy and beyond : results are available for dis coefficient functions @xcite and for the drell - yan k - factor @xcite at @xmath74 , and for the rate @xmath83 hadrons @xcite at @xmath84 ( i ll later deal with inclusive higgs production in some details ) .
all these computations are inclusive enough to allow a complete analytical integration over the phase - space of final - state partons .
such an integration is not possible in general , and an nnlo process - independent and exclusive formulation of ir cancellation will likely proceed through semi - analytical techniques , similar to those adopted at the nlo .
therefore , one needs to know the ir - divergent pieces of all the quantities contributing to the cross section at the nnlo .
these are ( i still assume that the lo gets contribution from the tree @xmath75-point functions ) : _ a ) _ tree - level @xmath85-parton amplitudes squared @xmath86 ; _ b ) _ the interference between tree - level @xmath87-parton amplitudes and one - loop @xmath87-parton amplitudes @xmath88 ; _ c ) _ the interference between tree - level @xmath75-parton amplitudes and two - loop @xmath75-parton amplitudes @xmath89 ; _ d ) _ one - loop @xmath75-parton amplitudes squared @xmath90 .
the ir divergences of @xmath86 result from having two soft partons , or three collinear partons , or one soft parton plus two other collinear partons , or two pairs of collinear partons .
only the latter configuration is trivial , in the sense that the corresponding singular behaviour of @xmath86 can be obtained from known nlo results ; the other limits have been studied in refs .
@xcite , and can be generally cast in the form of a reduced @xmath75-resolved - parton matrix element , times a suitable kernel .
the problem of combining these singular pieces into local ir counterterms , and of integrating these counterterms over the appropriate region of the phase space , is still unsolved .
one also needs to know the ir divergences of @xmath91 when one parton is soft or two partons are collinear : this is a new feature of nnlo computations , since at the nlo all partons in a virtual contribution are resolved .
these limits are also known @xcite . finally , the general form of the residues of the poles @xmath92 , @xmath93 , and @xmath94 appearing in @xmath95 has been given in ref .
@xcite , without computing any two - loop integrals ( see also @xcite ) .
the pole terms found in ref .
@xcite must precisely match those resulting from the explicit computations of two - loop integrals .
a lot of progress has been made in the past couple of years in such computations , and now all the two - loop @xmath96 and @xmath97 amplitudes are available .
first , all the ( very many ) tensor integrals are reduced to a much smaller number of master scalar integrals , thanks to integration - by - part identities @xcite ( previously used for two - point functions ) , and lorentz - invariance identities @xcite ; integration - by - part identities for @xmath75-leg , @xmath78-loop integrals were also shown @xcite to be equivalent to those for @xmath98-leg , @xmath99-loop integrals .
the problem of actually computing the master integrals is of a different nature .
the breakthrough @xcite was the use of a mellin - barnes representation for the propagators in the computation of planar and non - planar massless double box integrals ( expanded in the dimensional - regularization parameter @xmath100 ) .
negative space - dimension techniques @xcite can also be applied to simpler topologies @xcite .
the computation of master integrals is mapped onto the problem of solving differential equations in the approach of refs .
this approach has been adopted to compute all of the double box master integrals with one off - shell leg @xcite , not all of which had been computed with mellin - barnes techniques @xcite .
the master integrals so far computed , together with the reduction - to - master - integral techniques , _ would _ allow the computation of @xmath101 jets , and of two - jet production in hadronic collisions ( and a few other processes : see for example ref .
@xcite for a discussion unfortunately , these processes do nt include heavy flavour hadroproduction ) if one knew how to achieve ir cancellation for a generic observable at the nnlo . in hadronic physics ,
the computation of nnlo short - distance cross sections is not sufficient to get nnlo - accurate predictions , since nnlo - evolved pdf s are also necessary .
nnlo - evolved pdf s require the computation of altarelli - parisi splitting functions to three loops .
it turns out to be convenient to perform such a computation in mellin space ; the results for the first few mellin moments @xcite ( together with constraints on the small-@xmath2 behaviour ) have been used @xcite to obtain _ approximate _
expressions for the splitting functions in the @xmath2 space .
very recently , the complete three - loop computation of the @xmath102 part of the non - singlet structure function in dis has become available @xcite , from which the corresponding coefficient functions ( relevant to n@xmath103lo computations ) and splitting function can be extracted ; the latter has been cross - checked against the approximate results mentioned above , and full agreement has been found . although the complete expressions for the splitting functions will not appear soon @xcite , this fairly impressive result gives confidence on the accuracy of the approximate solution of ref .
one has to keep in mind that , in order for any pdf set ( such as nnlo - mrst @xcite ) to actually be of nnlo accuracy , not only the three - loop splitting functions are needed , but also _ all _ the short - distance cross sections used in the fits must be computed to nnlo . at present
, this is the case for the dis coefficient functions only ( which implies that the approximation is generally good , the bulk of the data being from dis ) ; for example , in the case of drell - yan the @xmath104 distribution needed for the fits has only nlo accuracy ( the drell - yan nnlo k - factor is used for normalization purposes @xcite ) .
this is another hint of the necessity of solving the problem of ir cancellation at the nnlo in a general way . in view of enormous amount of work done and yet to be done , one may ask whether the final outcome will justify such an effort .
the answer is certainly positive , but one must keep in mind that nnlo computations will not automatically mean _ precision _ physics . as shown in the case of @xmath18 event shapes , short- and large - distance effects are always correlated , and any advancement in perturbation theory should be complemented by a deeper understanding of this correlation .
furthermore , no precision study in hadronic collisions can be made without an accurate assessment of uncertainties due to pdfs .
this matter is now receiving considerable attention : see for example ref .
@xcite for a review .
nnlo computations will certainly play a major role in those cases in which nlo results still give an unsatisfactory description of data : examples are the processes with large k - factors ( such as @xmath7 production ) , or the observables which require a better description in terms of kinematics ( such as jet profiles ) .
a large - k - factor process is the direct production of sm higgs at hadron colliders . at the nlo
, the exact computation @xcite for the @xmath105 channel is found to be in excellent agreement with approximate results @xcite based on keeping the leading term of an expansion in @xmath106 ( the agreement further improves if the full dependence on @xmath106 is kept in the lo term ) .
therefore , one assumes the same to hold at the nnlo , and computes the nnlo contribution in the @xmath107 limit .
this is feasible since the effective lagrangian @xmath108 has been obtained @xcite to @xmath109 ( actually , one order larger than necessary here ) . exploiting the technique of ref .
@xcite , the two - loop virtual correction to @xmath110 has been obtained in ref .
the missing contributions to the physical cross section ( including @xmath111 and @xmath112 channels ) have been presented in ref .
@xcite , where an expansion for @xmath113 ( @xmath114 ) has been used to compute the double - real contribution .
terms up to @xmath115 have been included .
the @xmath113 expansion is expected to work well , since the gluon density is rather peaked towards small @xmath2 s , and thus the cm energy available at the partonic level is never too far from threshold . as argued in ref .
@xcite , collinear radiation also gives a sizable effect .
methods similar to @xcite have been used in refs .
@xcite , with the result of ref .
@xcite , to give early estimates of the complete nnlo rate .
these estimates are seen to agree well with the result of ref .
@xcite , which confirms the soft - collinear dominance in inclusive higgs production at colliders . finally ,
the result of ref .
@xcite has been found to agree to 1% level or better with the computation of ref .
@xcite , where the double - real contribution is also evaluated exactly ( it is interesting to notice that in ref .
@xcite the phase - space integrals relevant to real - emission terms are computed with techniques used so far only for loop diagrams it remains to be seen whether this method can be generalized to more exclusive observables ) .
-2.2pc as shown in fig .
[ fig : higgsatlhc ] , the inclusion of nnlo corrections seems to suggest that effects beyond this order are negligible .
the scale dependence is reduced wrt the one observed at the nlo ( see ref .
the dominance of the region @xmath116 implies the potential relevance of soft - gluon resummation .
preliminary results @xcite indicate that nnll resummation enhances the nnlo rate by 56% ( 1215% ) at the lhc ( tevatron ) , for @xmath117 gev .
the result for the fully - inclusive rate also serves to compute a slightly less inclusive observable , namely the rate for higgs+jets , with the @xmath4 of any jets imposed to be smaller than a fixed quantity @xmath118 ( jet veto ) .
such an observable , which should help in reducing the background due to the decay channel @xmath119 , has been computed in ref .
@xcite by subtracting the anti - vetoed jet cross section @xmath120 ( obtained in ref .
@xcite ) from the inclusive nnlo result discussed so far .
the study of more exclusive observables , which implies the understanding of ir cancellation at nnlo , will certainly prove useful in the future . in this case
, the dominance of the region @xmath116 might not be as strong as in the case of inclusive rates .
although it is unlikely that the region @xmath121 will play any role in phenomenological studies , it is worth recalling that in this region the approximation @xmath107 is not expected to work well : in the full theory the dominant contribution for @xmath122 is single - logarithmic , whereas double logs are also found in the large-@xmath123 theory .
the latter terms have been identified explicitly in ref .
@xcite with @xmath5-factorization arguments ; at the nnlo , they are seen to coincide with those resulting from the explicit computation of ref .
@xcite , thus providing a cross check impossible to achieve in the comparison of ref @xcite with ref . @xcite .
monte carlo ( mc ) programs are essential tools in experimental physics , giving fully - fledged descriptions of hadronic final states which can not be obtained in fixed - order computations .
schematically , an mc works as follows : for a given process , which at the lo receives contribution from @xmath124 reactions , @xmath125-particle configurations are generated , according to exact tree - level matrix element ( me ) computations .
the quarks and gluons ( partons henceforth ) among these primary particles are then allowed to emit more quarks and gluons , which are obtained from a parton shower or dipole cascade _ approximation _ to qcd dynamics .
this implies that mc s can not simulate the emission of final - state hard ( i.e. , with large relative transverse momenta ; thus , hard is synonymous of resolved here ) partons other than the primary ones obtained from me computations .
furthermore , total rates are accurate to lo .
although these problems are always present in mc simulations , they become acute when cm energies grow large , since in this case channels with large numbers of well - separated jets are phenomenologically very important and , correspondingly , total rates need to be computed to an accuracy better than lo .
two strategies can be devised in order to improve mc s .
the first aims at having @xmath126 extra hard partons in the final state ; thus , in the example given above , the number of final - state hard particles would increase from @xmath127 to @xmath128 .
this approach is usually referred to as _ matrix element corrections _ , since the mc must use the @xmath129-particle me s to generate the correct hard kinematics ; more details are given in sect .
[ sec : mecorr ] .
the second strategy also aims at simulating the production of @xmath128 hard particles , but improves the computation of rates as well , to n@xmath130lo accuracy .
a discussion is given in sect .
[ sec : mcatnlo ] .
there are basically two major problems in the implementation of me corrections .
the first problem is that of achieving a fast computation of the me s themselves for the largest possible @xmath128 , and an efficient phase - space generation .
the second problem stems from the fact that multi - parton me s are ir divergent .
clearly , in hard - particle configurations ir divergences do nt appear ; however , the definition of what hard means is , to a large extent , arbitrary . in practice ,
hardness is achieved by imposing some cuts on suitable partonic variables , such as @xmath4 s and @xmath131-distances @xmath132 in hadronic collisions .
i collectively denote these cuts by @xmath133 .
one assumes that @xmath75 hard partons will result ( after the shower ) into @xmath75 jets ; but , with a probability depending on @xmath133 , a given @xmath75-jet event could also result from @xmath134 hard partons .
this means that , when generating events at a fixed @xmath128 number of primary particles , physical observables in general depend upon @xmath133 ; i refer to this as the @xmath133-bias problem .
any solution to the @xmath133-bias problem implies a procedure to combine consistently the treatment of me s with different @xmath128 s . here , the difficulty is that of avoiding double counting , that is , the generation of the same kinematical configuration more often than prescribed by qcd .
the vast majority of recent approaches to me corrections address only the first of the two problems mentioned above .
a considerable amount of work has been devoted to the coding of hadronic processes with vector bosons / higgs plus heavy quarks in the final state , which can not be found ( regardless of the number of extra partons ) in standard mc s .
the complexity of hard - process generation for large @xmath128 suggests to implement it in a package ( which i call me generator ) distinct from the shower mc .
the me generator stores a set of hard configurations in a file ( event file ) ; the event file is eventually read by the mc , which uses the hard configurations as initial conditions for the showers .
the advantage of this procedure is that it is completely modular : one given event file can be read by different mc s , and conversely one mc can read event files produced by different me generators .
it is clearly convenient to reach an agreement on the format of such event records : this is now available @xcite ( les houches accord # 1 ) .
ready - to - use me generators ( with different numbers of hard processes implemented ) are acermc @xcite , alpgen @xcite , and madgraph / madevent @xcite .
related work , at present set up to function only with pythia , has been presented by the comphep @xcite and grace @xcite groups .
all of these me generators use feynman - diagram techniques in the computation of me s , except alpgen , which uses the iterative algorithm alpha @xcite ( see also @xcite ) .
when using an me generator , the cuts @xmath133 must be looser than those used to define the observables .
for example , the @xmath4-cut imposed at the parton level must be smaller than the minimum @xmath4 of any jets
. on the other hand , the cuts should not be too loose : the looser the cuts , the larger the probability of getting a @xmath75-jet event starting from @xmath134 hard partons .
thus , @xmath133 must be chosen in a range which is somewhat dependent upon the observables that one wants to study .
this implies that , strictly speaking , the combination of an me generator with a shower mc is _ not _ an event generator , since the event record depends upon the observables .
this happens precisely because such a combination is affected by the @xmath133-bias problem : it is therefore necessary to assess its impact on physical observables .
an example is given in fig .
[ fig : wjetalpgen ] , obtained using alpgen+herwig ( any other me generator and shower mc would give equivalent results for the same observable ) .
the plot presents jet rates , integrated over @xmath135 , for the hardest jet in w+3-jet events at the tevatron , versus the parton separation @xmath136 imposed at the level of me generation .
jets are reconstructed with the cone algorithm , with @xmath137 .
rates are normalized to the result obtained with @xmath138 .
the conclusion is that , in the `` reasonable '' range @xmath139 , the physical prediction has a @xmath140 dependence on @xmath136 .
the existence of a @xmath133 dependence should always be kept in mind when using an me generator + shower mc combination , because it affects the precision of the prediction . on the other hand , this is a rather modest price to pay : for multi - jet observables , ordinary mc s can underestimate the cross section by orders of magnitude , and the use of me generators is mandatory .
-2.2pc as mentioned before , the @xmath133 bias can be avoided by suitably combining the generation of me s with different @xmath128 . early proposal for me corrections
@xcite achieved this in the case @xmath141 .
the solution for arbitrary @xmath126 appears to be more complicated ; it has been fully implemented for shower mc s in @xmath18 collisions @xcite in the case of jet production ; along similar lines , proposal for colour dipole mc s @xcite and shower mc s in hadronic collisions @xcite have also been made . the idea of ref .
@xcite is the following . _
a ) _ integrate all the @xmath142 me s by imposing @xmath143 for any pairs of partons @xmath144 , with @xmath145 a fixed parameter and @xmath146 the interparton distance defined according to the @xmath5-algorithm @xcite . _
b ) _ choose statistically an @xmath126 , using the rates computed in _
c ) _ generate a @xmath147-parton configuration using the exact @xmath142 me , and reweight it with a suitable combination of sudakov form factors ( corresponding to the probability of no other branchings ) . _
d ) _ use the configuration generated in _
c ) _ as initial condition for a _ vetoed _ shower .
a vetoed shower proceeds as the usual one , except that it forbids all branchings @xmath148 with @xmath149 without stopping the scale evolution . in ref @xcite
, @xmath145 plays the role of @xmath133 .
although the selection of an @xmath126 value has a leading - log dependence on @xmath145 , it can be proved that this dependence is cancelled up to next - to - next - to - leading logs in physical observables @xcite .
parameter on @xmath150.,title="fig : " ] -2.2pc this is illustrated in fig .
[ fig : meand ] , where the mean value of the @xmath151 parameter @xcite ( full circles ) is plotted versus the value of @xmath145 , and compared to the measurement of ref .
@xcite ( band ) .
figure [ fig : meand ] clearly documents that the consistent combination of me s with different @xmath128 largely reduces the dependence of observables on @xmath133 .
this , however , comes at the price of modifying the shower algorithm .
furthermore , the procedure is more computing - intensive , since _ all _
@xmath126 s need to be considered ( this being impossible in practice , in @xmath18 collisions the procedure has an error of @xmath152 if only me s with @xmath153 partons are considered ; in ref .
@xcite , @xmath154 ) .
the problem of fully matching mc s with higher - order computations can be seen as an upgrade of me corrections : not only one wants to describe the kinematics of @xmath128 hard particles correctly , but the information on n@xmath130lo rates must also be included .
first attempts at solving this problem have only recently become available , and only for the case @xmath155 .
following ref .
@xcite , let me denote by @xmath156 the improved mc we aim at constructing .
the naive idea , of defining an @xmath156 by multiplying an mc with me corrections by the nlo k - factor , is simply not acceptable : the inclusiveness of the k - factor does not really fit well into the exclusive framework of an mc . thus
, any @xmath156 must involve the computation of virtual me s .
this is the reason why the construction of an @xmath156 is conceptually more complicated than me corrections : the ir divergences of the virtual me s can only be cancelled by computing real - emission me s with one soft parton or two collinear partons .
these soft and collinear configurations never occur in ordinary mc s ; in the case of me corrections , the cuts @xmath133 are specifically introduced to avoid them .
the presence of virtual me s also requires a less - intuitive definition of double counting @xcite : in the context of @xmath156 s , double counting may correspond to either an excess or a deficit in the prediction .
this generalization is necessary , since the me s used in @xmath156 s are not positive - definite . in order to describe in more details current approaches to @xmath156 ,
i adopt the toy model of ref .
@xcite , in which a system @xmath157 ( say , a quark ) can emit `` photons '' , massless particles with only one degree of freedom ( say , the energy ) .
the initial energy of @xmath157 is 1 , which becomes @xmath158 after the emission of one photon of energy @xmath159 .
the lo , virtual , and real contributions to the nlo cross section are ( @xmath160 is the coupling constant ) : ( ) _ b&=&b(x ) , [ born ] + ( ) _ v&=&a(+v)(x ) , [ virt ] + ( ) _ r&=&a , [ real ] where @xmath161 in eqs .
( [ born ] ) and ( [ virt ] ) reminds that there s no emission of real photons , and the ir divergence @xmath162 in eq .
( [ virt ] ) results from the loop integration over virtual photon momentum in @xmath163 dimensions .
i denote by @xmath164 the configuration of the system plus up to one photon , with @xmath165 in the case of the lo or virtual contributions ( eqs .
( [ born ] ) and ( [ virt ] ) respectively ) , and @xmath166 in the case of the real contribution ( eq .
( [ real ] ) ) .
energy conservation is understood , and therefore in the configuration @xmath164 the system has energy @xmath167 .
the function @xmath168 characterizes real emissions ; its specific form is irrelevant , except that , for ir cancellation to occur , it must fulfil @xmath169 for @xmath170 . with one photon emission at most , any observable @xmath171
can be represented by a function @xmath172 ; the computation of its expectation value @xmath173 can be achieved through standard techniques for ir cancellation : & = & bo(s,0)+a[nloslicing ] in the slicing method @xcite , and & = & _ 0 ^ 1 dx [ nlosubtint ] in the subtraction method @xcite . in an mc approach
, the system can undergo an arbitrary number of photon emissions .
i denote by @xmath174 the distribution in the observable @xmath171 obtained with mc methods ; this notation reminds that in a standard mc the initial condition for the shower is @xmath175 ( the lo kinematics ) , and that the total rate is @xmath10 ( the lo rate , see eq .
( [ born ] ) ) .
the most straightforward implementation of an @xmath156 can then be done by analogy : since two kinematical configurations , @xmath176 and @xmath175 , appear in the nlo cross section , one can use both of them as initial conditions for the showers . in order to recover the correct total rate , each event resulting from a shower with initial condition @xmath164
will be weighted with the coefficient of @xmath172 which appears in eq .
( [ nloslicing ] ) or in eq .
( [ nlosubtint ] ) . using eq .
( [ nlosubtint ] ) , one gets ( ) & = & _ 0 ^ 1 dx . [ imcnlonaive ] unfortunately , this naive approach does not work .
the weights @xmath177 and @xmath178 are ir - divergent at @xmath179 ; since the corresponding showers have different initial conditions @xmath176 and @xmath175 , it would take an infinite amount of time to cancel the divergences ( in other words , unweighting is impossible ) .
this is not a practical problem , is a fundamental one : the cancellation works for inclusive quantities , and the shower is exclusive .
so the main problem in the construction of an @xmath156 can be reformulated as follows : how to achieve ir cancellation , without giving up the exclusive properties of the showers .
besides , eq .
( [ imcnlonaive ] ) also suffers from double counting . in the context of the slicing method
, an approach has been proposed @xcite which exploits an idea of refs .
@xcite ( see also @xcite ) . the slicing parameter @xmath180 in eq . ( [ nloslicing ] )
is fixed to the value @xmath181 , by imposing that no @xmath175 contribution be present in the nlo cross section : b+a(b_0 + v)=0 .
[ getdeltaz ] this effectively restricts the energy of the real photons emitted to the range @xmath182 ( see eq .
( [ nloslicing ] ) ) .
this range is further partitioned by means of an arbitrary parameter @xmath183 .
one starts by generating the emission of a real photon with energy @xmath2 distributed according to @xmath177 .
then , if @xmath184 , the real - emission kinematics @xmath176 is mapped onto the lo kinematics @xmath175 ( in other words , the photon with energy @xmath2 is thrown away ) .
the configuration @xmath175 is used as initial condition for the shower , requiring the shower to forbid photon energies larger than @xmath183 . if @xmath185 , the real emission is kept , and @xmath176 is used as initial condition for the shower .
the corresponding formula is : & & ( ) = a_0 ^ 1 dx . [ phivmc ] the advantage of eq . ( [ phivmc ] ) is that it is manifestly positive - definite , and that its implementation can be carried out with little or no knowledge of the structure of the mc . on the other hand
, it can be shown @xcite that eq .
( [ phivmc ] ) still has double counting , and does not have a formal perturbative expansion in @xmath160 .
these problems arise since the technique adopted to deal with the ir cancellation is not exclusive enough : eq .
( [ getdeltaz ] ) is an integral equation ( the @xmath186 term is due to an integral over soft - photon configurations in the real - emission contribution ) .
although the absence of a perturbative expansion can be seen as a minor drawback from a practical point of view , the impact of the double counting should be assessed for each observable studied , by considering the dependence of physical predictions upon @xmath183 .
other approaches @xcite are based on the subtraction method .
one can observe that ordinary mc s do contain the information on the leading ir singular behaviour of nlo me s .
formally , this implies that the @xmath187 term in the perturbative expansion of the mc result can act as a _
counterterm to the ir divergences at the nlo ( strictly speaking , this is not exactly true in the case of large - angle soft gluon emission , and a few technical complications arise see ref .
@xcite ) . furthermore
, the form of the counterterm does not depend upon the observable studied .
thus , ir cancellation is achieved locally , but without any reference to a specific observable : this allows to implement it at the level of event generation , without giving up the exclusive treatment of the branchings .
the prescription of ref .
@xcite is & & ( ) = _ 0 ^ 1 dx , [ imcfive ] where the @xmath188-dependent quantities are the @xmath189 term of the mc result . the local ir cancellation mentioned before shows in the fact that the coefficients of @xmath190 and @xmath191 in eq .
( [ imcfive ] ) are finite , since the condition @xmath192 for @xmath170 always holds , regardless of the specific mc used .
therefore , these coefficients can be given as weights to the showers with @xmath176 and @xmath175 initial conditions respectively .
( [ imcfive ] ) does not have double counting , and features a smooth matching between the soft- and hard - emission regions of the phase space , without the need to introduce any extra parameter such as @xmath183 . the price to pay for
this is the presence of negative weights ( which however do not spoil the probabilistic interpretation of the results ) .
furthermore , one needs to know details of the mc ( @xmath188 ) in order to implement eq .
( [ imcfive ] ) : this seems unavoidable , since one should expect different mc s to match differently with a given nlo computation . the first qcd implementation of eq .
( [ imcfive ] ) has been presented in ref .
@xcite . the approach of refs .
@xcite uses a technique similar to that of ref .
@xcite , based on the definition of a local ir counterterm . at variance with ref .
@xcite , where the resummation of large logs is performed to ll accuracy , refs .
@xcite advocate an nll ( or beyond ) resummation ; for this to happen , it is argued that the standard formulation of collinear factorization must be extended .
this approach has been fully formulated @xcite only in the unphysical @xmath193 theory so far .
current qcd implementations do not include gluon emission .
it seems appropriate to start this section by mentioning a couple of phenomenological issues which would have deserved more attention .
one is the problem affecting the single - inclusive , isolated - photon measurements at the tevatron .
if d0 data are considered @xcite , a moderate disagreement with nlo qcd is present in the low-@xmath4 , central-@xmath41 region , which disappears when the ratio @xmath194 is considered .
the situation worsens in the case of cdf data @xcite : not only the discrepancy with qcd is statistically more significant for cross sections , but the measured ratio @xmath195 also disagrees with theory .
the consistent inclusion of recoil effects , along the lines of refs .
@xcite , might increase the qcd prediction at small @xmath4 s , and thus reduce the discrepancy .
the second problem affects the single - inclusive jet cross section as measured by d0 @xcite , the jets being reconstructed with a @xmath5-algorithm .
the data display a rather poor agreement with theory for @xmath196 gev ; this is disappointing , given the excellent results obtained with the cone algorithm ( as far as the cone algorithm is concerned , i should also mention here that the previously reported excess of data over theory at large @xmath4 has now completely disappeared : nlo qcd perfectly reproduces the data , if updated pdf sets are used .
an explanation in terms of pdfs has been already given in the past , but the pdf set used at that time resulted from an _ ad hoc _ fit named `` hj '' by the cteq collaboration .
this is not necessary any longer , since in the newest pdf releases the gluon density of the _ best _ fit naturally results to be hj - like .
@xcite for a discussion on this point ) .
it is hard to believe that the discrepancy in the case of the @xmath5-algorithm is the signal of a serious problem in qcd ( since it would probably affect the cone algorithm as well ) ; however , it may indicate a deficiency in current mc simulations , or in the understanding of hadronization and/or detector effects , which would surely worsen at the lhc energies .
it has to be remarked that the @xmath5-algorithm appears to work well at hera . in general ,
the capability of qcd to describe hard production processes is quite remarkable ( with the exception of @xmath7-production at lep and hera ) .
the definition of a formalism for the computation of exclusive observables at the nnlo is one of the most challenging and hot topics in perturbative qcd , which will have an important impact on phenomenological studies and will be crucial in improving the precision of @xmath0 ( and other fundamental parameters ) measurements at colliders , exclusive drell - yan production will necessarily have to be calculated to nnlo , in order to match the experimental precision .
the substantial progress made in the past couple of years in the computation of two - loop integrals and three - loop splitting functions , and the nnlo results for direct inclusive sm - higgs production , are certainly very encouraging .
nnlo results will also help to understand better the interplay between soft and hard physics ; the @xmath18 and dis environments will serve as a laboratory for the more involved case of hadronic collisions , where most of the work remains to be done .
monte carlos are much better equipped to face the challenges of tevatron run ii and lhc .
the implementation of new processes in matrix element generators proceeds steadily and , although there is still work to be done on the general structure of matrix element corrections , these techniques are rather well established by now .
formalisms for @xmath156 s have received considerable attention in the recent past ; the field has still to reach a mature stage , and new ideas will certainly be presented in the near future .
a rather obvious but quite challenging step is that of merging matrix element corrections and @xmath156 approaches .
p. nason , s. dawson and r. k. ellis , nucl .
b * 303 * ( 1988 ) 607 .
w. beenakker , w. l. van neerven , r. meng , g. a. schuler and j. smith , nucl .
b * 351 * ( 1991 ) 507 .
r. k. ellis and p. nason , nucl .
b * 312 * ( 1989 ) 551 .
j. smith and w. l. van neerven , nucl .
b * 374 * ( 1992 ) 36 .
m. drees , m. kramer , j. zunft and p. m. zerwas , phys .
b * 306 * ( 1993 ) 371 .
p. nason , s. dawson and r. k. ellis , nucl . phys .
b * 327 * ( 1989 ) 49 [ erratum - ibid .
b * 335 * ( 1990 ) 260 ] .
m. l. mangano , p. nason and g. ridolfi , nucl .
b * 373 * ( 1992 ) 295 .
s. frixione , m. l. mangano , p. nason and g. ridolfi , nucl .
b * 412 * ( 1994 ) 225 [ arxiv : hep - ph/9306337 ] .
e. laenen , s. riemersma , j. smith and w. l. van neerven , nucl .
b * 392 * ( 1993 ) 162 .
e. laenen , s. riemersma , j. smith and w. l. van neerven , nucl .
b * 392 * ( 1993 ) 229 .
t. affolder _ et al . _
[ cdf collaboration ] , phys .
d * 64 * ( 2001 ) 032002 [ arxiv : hep - ex/0101036 ] . v. m. abazov _ et al . _
[ d0 collaboration ] , arxiv : hep - ex/0205019 .
r. bonciani , s. catani , m. l. mangano and p. nason , nucl .
b * 529 * ( 1998 ) 424 [ arxiv : hep - ph/9801375 ] .
b. w. harris , e. laenen , l. phaf , z. sullivan and s. weinzierl , phys .
d * 66 * ( 2002 ) 054024 [ arxiv : hep - ph/0207055 ] .
s. frixione , m. l. mangano , p. nason and g. ridolfi , adv .
high energy phys .
* 15 * ( 1998 ) 609 [ arxiv : hep - ph/9702287 ] .
d. acosta _ et al .
_ [ cdf collaboration ] , phys . rev .
d * 65 * ( 2002 ) 052005 [ arxiv : hep - ph/0111359 ] . c. peterson , d. schlatter , i. schmitt and p. m. zerwas , phys .
d * 27 * ( 1983 ) 105 .
v. g. kartvelishvili , a. k. likhoded and v. a. petrov , phys .
b * 78 * ( 1978 ) 615 .
p. nason and c. oleari , nucl .
b * 565 * ( 2000 ) 245 [ arxiv : hep - ph/9903541 ] .
m. cacciari and p. nason , phys . rev .
* 89 * ( 2002 ) 122003 [ arxiv : hep - ph/0204025 ] . m. cacciari , m. greco and p. nason , jhep * 9805 * ( 1998 ) 007 [ arxiv : hep - ph/9803400 ] .
p. nason _ et al .
_ , arxiv : hep - ph/0003142 . s. frixione and m. l. mangano , nucl .
b * 483 * ( 1997 ) 321 [ arxiv : hep - ph/9605270 ] .
b. abbott _ et al . _
[ d0 collaboration ] , phys .
* 85 * ( 2000 ) 5068 [ arxiv : hep - ex/0008021 ] .
s. catani , m. ciafaloni and f. hautmann , nucl . phys . b * 366 * ( 1991 ) 135 .
j. c. collins and r. k. ellis , nucl .
b * 360 * ( 1991 ) 3 .
s. frixione , m. kramer and e. laenen , nucl .
b * 571 * ( 2000 ) 169 [ arxiv : hep - ph/9908483 ] . c. adloff
[ h1 collaboration ] , nucl .
b * 545 * ( 1999 ) 21 [ arxiv : hep - ex/9812023 ] .
j. breitweg _ et al . _
[ zeus collaboration ] , eur .
j. c * 6 * ( 1999 ) 67 [ arxiv : hep - ex/9807008 ] .
m. cacciari , s. frixione and p. nason , jhep * 0103 * ( 2001 ) 006 [ arxiv : hep - ph/0102134 ] .
s. frixione and p. nason , jhep * 0203 * ( 2002 ) 053 [ arxiv : hep - ph/0201281 ] .
i. abt _ et al .
_ [ hera - b collaboration ] , arxiv : hep - ex/0205106 . c. adloff _ et al .
_ [ h1 collaboration ] , phys .
b * 467 * ( 1999 ) 156 [ erratum - ibid .
b * 518 * ( 2001 ) 331 ] [ arxiv : hep - ex/9909029 ] .
j. breitweg _ et al . _
[ zeus collaboration ] , eur .
j. c * 18 * ( 2001 ) 625 [ arxiv : hep - ex/0011081 ] .
m. acciarri _ et al . _
[ l3 collaboration ] , phys .
b * 503 * ( 2001 ) 10 [ arxiv : hep - ex/0011070 ] .
y. l. dokshitzer and b. r. webber , phys .
b * 352 * ( 1995 ) 451 [ arxiv : hep - ph/9504219 ] .
y. l. dokshitzer , g. marchesini and b. r. webber , nucl .
b * 469 * ( 1996 ) 93 [ arxiv : hep - ph/9512336 ] .
g. p. korchemsky and g. sterman , nucl .
b * 437 * ( 1995 ) 415 [ arxiv : hep - ph/9411211 ] .
g. p. korchemsky and g. sterman , arxiv : hep - ph/9505391 .
y. l. dokshitzer and b. r. webber , phys .
b * 404 * ( 1997 ) 321 [ arxiv : hep - ph/9704298 ] .
p. nason and m. h. seymour , nucl .
b * 454 * ( 1995 ) 291 [ arxiv : hep - ph/9506317 ] .
y. l. dokshitzer , a. lucenti , g. marchesini and g. p. salam , nucl .
b * 511 * ( 1998 ) 396 [ erratum - ibid .
b * 593 * ( 2001 ) 729 ] [ arxiv : hep - ph/9707532 ] .
y. l. dokshitzer , a. lucenti , g. marchesini and g. p. salam , jhep * 9805 * ( 1998 ) 003 [ arxiv : hep - ph/9802381 ] .
m. dasgupta and b. r. webber , jhep * 9810 * ( 1998 ) 001 [ arxiv : hep - ph/9809247 ] . y. l. dokshitzer , g. marchesini and g. p. salam , eur .
j. directc * 3 * ( 1999 ) 1 [ arxiv : hep - ph/9812487 ] .
a. movilla fernandez , arxiv : hep - ex/0209022 .
p. a. movilla fernandez , s. bethke , o. biebel and s. kluth , eur .
j. c * 22 * ( 2001 ) 1 [ arxiv : hep - ex/0105059 ] .
s. bethke , eur .
j. directc * 1 * ( 2002 ) 1 [ arxiv : hep - ex/0112032 ] . m. dasgupta and g. p. salam , jhep * 0208 * ( 2002 ) 032 [ arxiv : hep - ph/0208073 ] .
g. p. salam , arxiv : hep - ph/0207147 .
korchemsky and g. sterman , nucl .
b * 555 * ( 1999 ) 335 [ arxiv : hep - ph/9902341 ] .
g. p. korchemsky and s. tafat , jhep * 0010 * ( 2000 ) 010 [ arxiv : hep - ph/0007005 ] .
e. gardi and j. rathsman , nucl .
b * 609 * ( 2001 ) 123 [ arxiv : hep - ph/0103217 ] .
e. gardi and j. rathsman , nucl .
b * 638 * ( 2002 ) 243 [ arxiv : hep - ph/0201019 ] .
a. v. belitsky , g. p. korchemsky and g. sterman , phys .
b * 515 * ( 2001 ) 297 [ arxiv : hep - ph/0106308 ] .
e. gardi , nucl .
b * 622 * ( 2002 ) 365 [ arxiv : hep - ph/0108222 ] .
p. abreu _ et al .
_ [ delphi collaboration ] , eur .
j. c * 14 * ( 2000 ) 557 [ arxiv : hep - ex/0002026 ] .
d. wicke , arxiv : hep - ph/0205299 .
a. dhar and v. gupta , phys .
d * 29 * ( 1984 ) 2822 .
g. p. salam and d. wicke , jhep * 0105 * ( 2001 ) 061 [ arxiv : hep - ph/0102343 ] .
a. banfi , y. l. dokshitzer , g. marchesini and g. zanderighi , jhep * 0105 * ( 2001 ) 040 [ arxiv : hep - ph/0104162 ] .
a. banfi , g. marchesini , g. smye and g. zanderighi , jhep * 0108 * ( 2001 ) 047 [ arxiv : hep - ph/0106278 ] .
z. bern , l. j. dixon and d. a. kosower , nucl .
b * 412 * ( 1994 ) 751 [ arxiv : hep - ph/9306240 ] .
r. k. ellis , d. a. ross and a. e. terrano , nucl .
b * 178 * ( 1981 ) 421 .
k. fabricius , i. schmitt , g. kramer and g. schierholz , z. phys .
c * 11 * ( 1981 ) 315 .
w. t. giele and e. w. glover , phys .
d * 46 * ( 1992 ) 1980 .
w. t. giele , e. w. glover and d. a. kosower , nucl .
b * 403 * ( 1993 ) 633 [ arxiv : hep - ph/9302225 ] .
s. frixione , z. kunszt and a. signer , nucl .
b * 467 * ( 1996 ) 399 [ arxiv : hep - ph/9512328 ] . s. catani and m. h. seymour , phys .
b * 378 * ( 1996 ) 287 [ arxiv : hep - ph/9602277 ] .
s. catani and m. h. seymour , nucl .
b * 485 * ( 1997 ) 291 [ erratum - ibid .
b * 510 * ( 1997 ) 503 ] [ arxiv : hep - ph/9605323 ] .
s. frixione , nucl .
b * 507 * ( 1997 ) 295 [ arxiv : hep - ph/9706545 ] .
z. nagy and z. trocsanyi , nucl .
b * 486 * ( 1997 ) 189 [ arxiv : hep - ph/9610498 ] .
d. e. soper , phys .
* 81 * ( 1998 ) 2638 [ arxiv : hep - ph/9804454 ] .
d. e. soper , phys .
d * 62 * ( 2000 ) 014009 [ arxiv : hep - ph/9910292 ] .
m. kramer and d. e. soper , phys .
d * 66 * ( 2002 ) 054017 [ arxiv : hep - ph/0204113 ] .
e. b. zijlstra and w. l. van neerven , nucl .
b * 383 * ( 1992 ) 525 .
e. b. zijlstra and w. l. van neerven , phys .
b * 297 * ( 1992 ) 377 .
r. hamberg , w. l. van neerven and t. matsuura , nucl .
b * 359 * ( 1991 ) 343 .
w. l. van neerven and e. b. zijlstra , nucl .
b * 382 * ( 1992 ) 11 .
s. g. gorishnii , a. l. kataev and s. a. larin , phys .
b * 259 * ( 1991 ) 144 .
f. a. berends and w. t. giele , nucl .
b * 313 * ( 1989 ) 595 .
j. m. campbell and e. w. glover , nucl .
b * 527 * ( 1998 ) 264 [ arxiv : hep - ph/9710255 ] .
s. catani and m. grazzini , phys .
b * 446 * ( 1999 ) 143 [ arxiv : hep - ph/9810389 ] .
v. del duca , a. frizzo and f. maltoni , nucl .
b * 568 * ( 2000 ) 211 [ arxiv : hep - ph/9909464 ] .
s. catani and m. grazzini , nucl .
b * 570 * ( 2000 ) 287 [ arxiv : hep - ph/9908523 ] .
z. bern , l. j. dixon , d. c. dunbar and d. a. kosower , nucl .
b * 425 * ( 1994 ) 217 [ arxiv : hep - ph/9403226 ] .
z. bern and g. chalmers , nucl .
b * 447 * ( 1995 ) 465 [ arxiv : hep - ph/9503236 ] .
z. bern , v. del duca and c. r. schmidt , phys .
b * 445 * ( 1998 ) 168 [ arxiv : hep - ph/9810409 ] .
d. a. kosower , nucl .
b * 552 * ( 1999 ) 319 [ arxiv : hep - ph/9901201 ] .
d. a. kosower and p. uwer , nucl .
b * 563 * ( 1999 ) 477 [ arxiv : hep - ph/9903515 ] .
z. bern , v. del duca , w. b. kilgore and c. r. schmidt , phys .
d * 60 * ( 1999 ) 116001 [ arxiv : hep - ph/9903516 ] .
s. catani and m. grazzini , nucl .
b * 591 * ( 2000 ) 435 [ arxiv : hep - ph/0007142 ] .
s. catani , phys .
b * 427 * ( 1998 ) 161 [ arxiv : hep - ph/9802439 ] .
g. sterman and m. e. tejeda - yeomans , arxiv : hep - ph/0210130 .
f. v. tkachov , phys .
b * 100 * ( 1981 ) 65 .
k. g. chetyrkin and f. v. tkachov , nucl .
b * 192 * ( 1981 ) 159 .
t. gehrmann and e. remiddi , nucl .
b * 580 * ( 2000 ) 485 [ arxiv : hep - ph/9912329 ] .
p. a. baikov and v. a. smirnov , phys .
b * 477 * ( 2000 ) 367 [ arxiv : hep - ph/0001192 ] . v. a. smirnov , phys . lett .
b * 460 * ( 1999 ) 397 [ arxiv : hep - ph/9905323 ] .
j. b. tausk , phys .
b * 469 * ( 1999 ) 225 [ arxiv : hep - ph/9909506 ] .
i. g. halliday and r. m. ricotta , phys .
b * 193 * ( 1987 ) 241 . c. anastasiou , e. w. glover and c. oleari , nucl .
b * 565 * ( 2000 ) 445 [ arxiv : hep - ph/9907523 ] .
a. t. suzuki and a. g. schmidt , j. phys .
a * 35 * ( 2002 ) 151 [ arxiv : hep - th/0110047 ] .
a. v. kotikov , phys .
b * 267 * ( 1991 ) 123 .
e. remiddi , nuovo cim . a * 110 * ( 1997 ) 1435 [ arxiv : hep - th/9711188 ] .
t. gehrmann and e. remiddi , nucl .
b * 601 * ( 2001 ) 248 [ arxiv : hep - ph/0008287 ] .
t. gehrmann and e. remiddi , nucl .
b * 601 * ( 2001 ) 287 [ arxiv : hep - ph/0101124 ] .
v. a. smirnov , phys .
b * 491 * ( 2000 ) 130 [ arxiv : hep - ph/0007032 ] .
v. a. smirnov , phys .
b * 500 * ( 2001 ) 330 [ arxiv : hep - ph/0011056 ] .
w. giele _ et al .
_ , arxiv : hep - ph/0204316 . s. a. larin , t. van ritbergen and j. a. vermaseren , nucl .
b * 427 * ( 1994 ) 41 .
s. a. larin , p. nogueira , t. van ritbergen and j. a. vermaseren , nucl .
b * 492 * ( 1997 ) 338 [ arxiv : hep - ph/9605317 ] .
a. retey and j. a. vermaseren , nucl .
b * 604 * ( 2001 ) 281 [ arxiv : hep - ph/0007294 ] .
w. l. van neerven and a. vogt , nucl .
b * 568 * ( 2000 ) 263 [ arxiv : hep - ph/9907472 ] .
w. l. van neerven and a. vogt , nucl .
b * 588 * ( 2000 ) 345 [ arxiv : hep - ph/0006154 ] .
w. l. van neerven and a. vogt , phys .
b * 490 * ( 2000 ) 111 [ arxiv : hep - ph/0007362 ] . s. moch , j. a. vermaseren and a. vogt , arxiv : hep - ph/0209100 .
a. d. martin , r. g. roberts , w. j. stirling and r. s. thorne , phys .
b * 531 * ( 2002 ) 216 [ arxiv : hep - ph/0201127 ] .
m. spira , a. djouadi , d. graudenz and p. m. zerwas , nucl .
b * 453 * ( 1995 ) 17 [ arxiv : hep - ph/9504378 ] .
s. dawson , nucl .
b * 359 * ( 1991 ) 283 .
a. djouadi , m. spira and p. m. zerwas , phys .
b * 264 * ( 1991 ) 440 .
k. g. chetyrkin , b. a. kniehl and m. steinhauser , phys .
* 79 * ( 1997 ) 353 [ arxiv : hep - ph/9705240 ] .
k. g. chetyrkin , b. a. kniehl and m. steinhauser , nucl .
b * 510 * ( 1998 ) 61 [ arxiv : hep - ph/9708255 ] .
r. v. harlander , phys .
b * 492 * ( 2000 ) 74 [ arxiv : hep - ph/0007289 ] .
r. v. harlander and w. b. kilgore , phys .
* 88 * ( 2002 ) 201801 [ arxiv : hep - ph/0201206 ] . m. kramer , e. laenen and m. spira , nucl . phys .
b * 511 * ( 1998 ) 523 [ arxiv : hep - ph/9611272 ] .
s. catani , d. de florian and m. grazzini , jhep * 0105 * ( 2001 ) 025 [ arxiv : hep - ph/0102227 ] .
r. v. harlander and w. b. kilgore , phys .
d * 64 * ( 2001 ) 013015 [ arxiv : hep - ph/0102241 ] . c. anastasiou and k. melnikov , arxiv : hep - ph/0207004 .
m. grazzini , arxiv : hep - ph/0209302 .
s. catani , d. de florian and m. grazzini , jhep * 0201 * ( 2002 ) 015 [ arxiv : hep - ph/0111164 ] .
d. de florian , m. grazzini and z. kunszt , phys .
* 82 * ( 1999 ) 5209 [ arxiv : hep - ph/9902483 ] .
f. hautmann , phys .
b * 535 * ( 2002 ) 159 [ arxiv : hep - ph/0203140 ] .
e. boos _ et al .
_ , arxiv : hep - ph/0109068 . b. p. kersevan and e. richter - was , arxiv : hep - ph/0201302 .
f. caravaglios , m. l. mangano , m. moretti and r. pittau , nucl .
b * 539 * ( 1999 ) 215 [ arxiv : hep - ph/9807570 ] .
m. l. mangano , m. moretti and r. pittau , nucl .
b * 632 * ( 2002 ) 343 [ arxiv : hep - ph/0108069 ] .
m. l. mangano , m. moretti , f. piccinini , r. pittau and a. d. polosa , arxiv : hep - ph/0206293 .
a. pukhov _ et al .
_ , arxiv : hep - ph/9908288 .
a. s. belyaev _ et al .
_ , arxiv : hep - ph/0101232 . k. sato , s. tsuno , j. fujimoto , t. ishikawa , y. kurihara and s. odaka , arxiv : hep - ph/0104237 .
s. tsuno , k. sato , j. fujimoto , t. ishikawa , y. kurihara , s. odaka and t. abe , arxiv : hep - ph/0204222 .
t. stelzer and w. f. long , comput .
commun .
* 81 * ( 1994 ) 357 [ arxiv : hep - ph/9401258 ] .
f. maltoni and t. stelzer , arxiv : hep - ph/0208156 .
f. caravaglios and m. moretti , phys .
b * 358 * ( 1995 ) 332 [ arxiv : hep - ph/9507237 ] .
d. draggiotis , r. h. kleiss and c. g. papadopoulos , eur .
j. c * 24 * ( 2002 ) 447 [ arxiv : hep - ph/0202201 ] .
g. gustafson , phys .
b * 175 * ( 1986 ) 453 .
m. bengtsson and t. sjostrand , phys .
b * 185 * ( 1987 ) 435 .
m. bengtsson and t. sjostrand , nucl .
b * 289 * ( 1987 ) 810 .
g. gustafson and u. pettersson , nucl .
b * 306 * ( 1988 ) 746 .
m. h. seymour , comput .
commun .
* 90 * ( 1995 ) 95 [ arxiv : hep - ph/9410414 ] . s. catani , f. krauss , r. kuhn and b. r. webber , jhep * 0111 * ( 2001 ) 063 [ arxiv : hep - ph/0109231 ] .
r. kuhn , f. krauss , b. ivanyi and g. soff , comput .
commun .
* 134 * ( 2001 ) 223 [ arxiv : hep - ph/0004270 ] .
l. lonnblad , jhep * 0205 * ( 2002 ) 046 [ arxiv : hep - ph/0112284 ] . f. krauss , jhep * 0208 * ( 2002 ) 015 [ arxiv : hep - ph/0205283 ] .
s. catani , y. l. dokshitzer , m. olsson , g. turnock and b. r. webber , phys .
b * 269 * ( 1991 ) 432 .
g. parisi , phys .
b * 74 * ( 1978 ) 65 .
b. adeva _ et al . _
[ l3 collaboration ] , z. phys .
c * 55 * ( 1992 ) 39 .
s. frixione and b. r. webber , jhep * 0206 * ( 2002 ) 029 [ arxiv : hep - ph/0204244 ] .
b. potter , phys .
d * 63 * ( 2001 ) 114017 [ hep - ph/0007172 ] ; b. potter and t. schorner , phys .
b * 517 * ( 2001 ) 86 [ hep - ph/0104261 ] .
m. dobbs , phys .
d * 64 * ( 2001 ) 034016 [ hep - ph/0103174 ] ; m. dobbs , phys .
d * 65 * ( 2002 ) 094011 [ hep - ph/0111234 ] .
h. baer and m. h. reno , phys .
d * 44 * ( 1991 ) 3375 .
h. baer and m. h. reno , phys .
d * 45 * ( 1992 ) 1503 ; h. baer and m. h. reno , phys .
d * 54 * ( 1996 ) 2017 [ hep - ph/9603209 ] .
j. collins , jhep * 0005 * ( 2000 ) 004 [ hep - ph/0001040 ]
. y. chen , j. c. collins and n. tkachuk , jhep * 0106 * ( 2001 ) 015 [ hep - ph/0105291 ] .
j. collins , phys .
d * 65 * ( 2002 ) 094016 [ hep - ph/0110113 ] .
y. chen , j. collins and x. zu , hep - ph/0110257 ; y. kurihara , arxiv : hep - ph/0207214 . s. frixione and b. r. webber , arxiv : hep - ph/0207182
. b. abbott _ et al . _
[ d0 collaboration ] , phys .
* 84 * ( 2000 )
2786 [ arxiv : hep - ex/9912017 ] . v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
* 87 * ( 2001 ) 251805 [ arxiv : hep - ex/0106026 ] .
d. acosta _ et al .
_ [ cdf collaboration ] , phys . rev .
d * 65 * ( 2002 ) 112003 [ arxiv : hep - ex/0201004 ] .
e. laenen , g. sterman and w. vogelsang , phys .
d * 63 * ( 2001 ) 114018 [ arxiv : hep - ph/0010080 ] .
a. kulesza , g. sterman and w. vogelsang , phys .
d * 66 * ( 2002 ) 014011 [ arxiv : hep - ph/0202251 ] . v. m. abazov _ et al . _
[ d0 collaboration ] , phys .
b * 525 * ( 2002 ) 211 [ arxiv : hep - ex/0109041 ] .
j. pumplin , d. r. stump , j. huston , h. l. lai , p. nadolsky and w. k. tung , jhep * 0207 * ( 2002 ) 012 [ arxiv : hep - ph/0201195 ] . | i review recent results in qcd at high energy , emphasizing the role of higher - order computations , power corrections , and monte carlo simulations in the study of a few discrepancies between data and perturbative predictions , and discussing future prospects . |
cerium intermetallics are rich in exciting phenomena due to various manifestations of the hybridization between a conduction band and 4f electrons ( @xmath3 ) based on the kondo interaction @xcite . among the most peculiar effects
one may list the formation of heavy fermions , unconventional superconductivity , quantum critical phenomena and non - fermi liquid behavior .
microscopic models for heavy fermion ( hf ) systems assume very often that the strength of @xmath3 hybridization is momentum independent @xcite . on the other hand ,
there are theoretical considerations @xcite indicating a strong momentum dependence of @xmath3 with possible maxima and nodes .
this complex variation of the @xmath3 amplitude is expected to result from the symmetry of f - electrons , which participate in the formation of the kondo singlet .
therefore , the issue of @xmath3 anisotropy deserves more attention in the experimental investigation of f - electron systems .
a typical spectroscopic manifestation of the kondo effect is the abrikosov - suhl resonance also called the kondo resonance ( kr ) which is a narrow high intensity peak in the spectral function located close to the fermi energy ( @xmath4 ) @xcite . in case of ce
based systems angle resolved photoemission spectroscopy ( arpes ) was able to determine a k - vector dependent intensity variation of the kr @xcite .
it was reported that the increased kr intensity is found at normal emission @xcite or at k - vectors related to bands crossing @xmath4 ( fermi vectors ) @xcite .
recently , it was shown for cecoin@xmath5 @xcite that the f - electron peak intensity depends considerably on a band crossing @xmath4 , which was interpreted in terms of an anisotropic @xmath3 .
optical spectroscopy @xcite also reveals anisotropic hybridisation for ce-115 compound .
another effect of the 4f - conduction band hybridization is kr dispersion or more precisely heavy quasiparticle dispersion , which has also been found in ce systems @xcite .
nevertheless , a complete image of momentum dependent hybridization strength @xmath3 with the occurrence of eventual nodes predicted by theory @xcite remains a challenge for an experiment so far .
the object of our investigations , ce@xmath0co@xmath1si@xmath2 , crystallizes with a derivative of the hexagonal alb@xmath0-type structure with the @xmath6 space group @xcite .
as established from detailed magnetic susceptibility , electrical resistivity and heat capacity measurements of single - crystalline specimens , ce@xmath0co@xmath1si@xmath2 does not order magnetically down to 0.4 k. its physical behavior at low temperatures is governed by strong _
f_-ligand hybridization , leading to enhanced electronic contribution to the specific heat [ @xmath7 200 mj@xmath8(mol@xmath9k@xmath10 ) at 0.4 k ] and kondo - like temperature variations of the electrical resistivity with the characteristic temperature @xmath11 of about 50 k @xcite .
the @xmath12 dependencies are dominated by broad maxima near 80 k , which manifest a crossover from incoherent to coherent kondo regime .
most interestingly , below 10 k , all the bulk characteristics of ce@xmath0co@xmath1si@xmath2 show non - fermi - liquid ( nfl ) features that are compatible with theoretical predictions for griffiths phases .
these should be related to a disorder in ce - si sublattice found recently @xcite .
so far , pes studies performed on ce@xmath0cosi@xmath13 polycrystals delivered only angle - integrated spectra showing a kr @xcite . in this paper
, we present arpes studies of the kondo lattice ce@xmath0co@xmath1si@xmath2 system .
the spectra consist of dispersions originating from surface and bulk states and a kondo peak related mainly to bulk states .
a specific k - vector dependence of kr indicates that certain bulk bands cross @xmath4 with much higher quasiparticle weight than others .
these results point to a momentum and/or band dependent anisotropic @xmath3 hybridization in ce@xmath0co@xmath1si@xmath2 .
single crystals of ce@xmath0co@xmath1si@xmath2 have been grown by the czochralski method and characterized as described elsewhere @xcite .
the arpes experiment was conducted at the ape beamline of elettra synchrotron @xcite with a ses2002 electron spectrometer .
prior to the photoemission studies the crystals were oriented with a laue method .
subsequently , they were cleaved at a pressure of @xmath14 mbar exposing flat surfaces along the ( 10@xmath150 ) crystallographic plane .
the measurements were carried out with linearly polarized radiation , typically at a temperature of 25 k. the energy and wave vector resolution was fixed to 18 mev and 0.01 @xmath16 respectively .
the fermi energy was determined regularly on evaporated gold .
band structure calculations were performed for stoichiometric ce@xmath0cosi@xmath13 using the scalar relativistic version of the full - potential local - orbital ( fplo ) code @xcite with the perdew - wang @xcite exchange - correlation potential .
additional correlations within the lsda+u approach were accomplished employing the around - mean - field scheme @xcite .
an irreducible wedge of the brillouin zone comprised 133 points .
the valence band of ce@xmath0co@xmath1si@xmath2 was investigated by means of photoemission spectroscopy with the photon energies ( h@xmath17 ) corresponding to lower ( h@xmath17=25 ev ) and higher ( h@xmath17=40 ev ) photoionization cross section for ce 4f electrons @xcite .
the spectra obtained by integrating arpes data for h@xmath17=40 ev and h@xmath17=25 ev are shown in fig .
1 . usually , one calculates the difference of the spectra corresponding to these energies to approximate a contribution from 4f electrons .
however , in the case of ce@xmath0co@xmath1si@xmath2 the photoionization cross section for co 3d electrons almost doubles for the excitation energy increase from 25 to 40 ev according to the estimates for free atoms @xcite .
hence , the contribution from both orbitals , ce 4f and co 3d , is estimated by such a subtraction .
the peaks of the well screened f@xmath18 and f@xmath19 final states ( fig .
1 ) indicate a high hybridization between 4f and conduction band electrons .
the characteristic tail of the kondo peak f@xmath18 is located near @xmath4 and its spin - orbit partner f@xmath19 at 275 mev .
the broad peak at @xmath20 2 ev can not solely be attributed to the weakly screened f@xmath21 final state .
this is because the arpes data ( fig .
2 ) reveal high intensity dispersing peaks near this energy , which can not originate from highly localized and weakly hybridized f - electrons .
thus , a large part of the spectral weight at @xmath20 2 ev should be attributed to co 3d electrons .
density of states ( dos ) obtained theoretically ( not shown ) by means of the fplo method with lsda approach indicate that si 3p , co 3d and ce 4f dominate the valence band .
the contribution from si 3p to the photoemission spectra is less significant due to a very low photoionization cross section at h@xmath17=25 and 40 ev .
= 40 ev and 25 ev resulting from angle integration of the arpes data collected along the @xmath22 - @xmath23 direction of ce@xmath0co@xmath1si@xmath2 and their difference.,width=249 ] further insight into the electronic structure of ce@xmath0co@xmath1si@xmath2 is given by band mapping with arpes .
the path scanned in the reciprocal space is shown as a green dashed curve inside the three - dimensional ( 3d ) brillouin zone ( bz ) [ fig .
2 ( a ) ] . this curve is part of a large circle in the @xmath24-a - l - m plane , for which the perpendicular to the surface component of the wave vector ( k@xmath25 ) is unknown . for surface states the problem
is reduced to two dimensions and the band structure is scanned along @xmath22-@xmath23 , which is drawn with a green dashed straight line in the surface brillouin zone ( sbz ) .
as the considered arpes data do not deliver any information about the real value of k@xmath25 both 2d and 3d bands will be described in the sbz for simplicity .
the measurements were performed with h@xmath17=40 ev at a temperature of 25 k , which is below the coherence temperature estimated to be t@xmath26 80 k. kr and spin - orbit partner are not dispersing but vary in intensity .
the other peaks reveal dispersions .
in particular , @xmath27 and @xmath28 bands ( fig .
2 ) clearly cross @xmath4 .
the electron pocket ( @xmath27 ) observed around the @xmath22 points is more shallow near k@xmath29=0 than at k@xmath30 1.5 @xmath16 .
the locations in the reciprocal space scanned by arpes for k@xmath29=0 and k@xmath29=1.5@xmath16 do not share the same out of plane component of the wave vector - k@xmath25 according to the free electron final state model @xcite .
hence , 3d states exhibit different dispersions for the equivalent k@xmath29 but different k@xmath25 in neighboring bzs .
therefore , the @xmath27 pocket is considered to be a 3d band of bulk origin .
the @xmath28 band crossing @xmath4 at @xmath22 is well visible at k@xmath30 1.5 @xmath16 in contrast to k@xmath30 0 @xmath16 .
this parabolic band exhibits the same dispersion in both bzs and is therefore regarded as a surface state .
it has the highest spectral intensity of all features in the spectrum , which is expected for surface states at the photon energy of h@xmath17=40 ev .
this is due to the very small mean free path of electrons for this energy @xcite .
the difference in the intensity at k@xmath30 1.5 @xmath16 and k@xmath30 0 @xmath16 may be attributed to matrix element effects , which favor certain bands with a higher photoemission cross section , depending on the geometry of the experiment , k - vector and probed bz .
this may result in suppression or disappearance of particular bands .
in fact , the @xmath28 band looses intensity when approaching k@xmath29=0 @xmath16 . at this wave vector
the experiment is realized in a pure @xmath31-polarization , which excludes bands with odd parity with respect to the experimental mirror plane . due to this
we can not conclude whether this band crosses @xmath4 near k@xmath29=0 @xmath16 or bends back to higher energies .
cosi@xmath13 by means of fplo with lsda+u approximation for u@xmath32 = 6 ev along ( a ) a standard path connecting high symmetry points and ( b ) for the directions parallel to @xmath24-a .
the high symmetry points are defined in fig 2 ( a ) .
the vertical dashed line denotes k@xmath29=0.6 @xmath16 with the highest kr .
solid ( black ) and dashed ( red ) lines show dispersions corresponding to spin up and down directions respectively . ]
band structure calculations using the fplo method ( fig .
3 ) are helpful for a further interpretation of the arpes results .
theoretical dispersions for bulk ce@xmath0cosi@xmath13 are drawn along a conventional path in the reciprocal space [ fig 3 ( a ) ] and for a set of directions parallel to @xmath24-a [ fig 3 ( b ) ] .
the real path for the arpes studies should be located close to one of the latter directions [ see also fig . 2 ( a ) ] . to consider different strength of correlation the calculations were performed with the ce 4f correlation parameters @xmath33 = 0 ( not shown ) and @xmath33 = 6 ev .
the calculations yield more bands , than observed in the experiment .
a certain number of them can not be seen because of unfavorable matrix elements .
the electron pocket @xmath27 found experimentally near @xmath24 has its counterpart in the calculations for @xmath33 = 6 ev along @xmath24-a and @xmath34-r@xmath35 .
it reaches higher binding energy along @xmath24-a and is a 3d state . on the other hand ,
the exact @xmath28 dispersion is not found in the calculations .
this confirms the surface state origin of this band .
the other bands recorded with a weaker intensity ( fig .
2 ) have corresponding dispersions in the calculations with both @xmath33 = 0 and 6 ev .
the theoretical band structure consists of two components related to opposite spin direction .
this is due to a magnetic ground state predicted by calculations , which is however not found in the experiment .
this discrepancy can be explained by the fact that in the calculations the total energy difference between magnetic and non - magnetic ground state is so low that even at low temperatures the thermal excitations would destroy long range magnetic order .
moreover , kondo scattering present in ce@xmath0co@xmath1si@xmath2 should screen magnetic interactions .
arpes data provide a direct evidence of kr intensity variation with k@xmath29 , while its dispersion is not found at t=25 k , a temperature well below the coherence temperature of the kondo lattice , t@xmath37 80k .
energy distribution curves ( edcs ) for exemplary k - vectors from fig .
2 ( b),(c ) are shown in fig .
the height of the kr ( blue arrow ) was estimated with respect to the background located between kr and the spin - orbit splitting peak after proper normalization of the edcs . the
first surprising fact is that the kondo peak intensity is not directly correlated with the fermi wave vectors found in the same experiment . @xmath4
crossings at k@xmath29=0.3 , 1.1 and 1.9 @xmath16 , which are related to the most prominent bulk state , namely the @xmath27 electron pocket , exhibit a relatively low intensity kondo peak . on the other hand ,
the highest intensity kr is found around k@xmath29=0.6 @xmath16 , where bands crossing the fermi energy are not directly observed .
the arpes measurements were performed a few times to exclude the effects of surface quality and kr intensity dependence on k is reproducible .
it is known that the kr may have a much higher intensity than its parent band , which crosses @xmath4 @xcite .
therefore , it is possible that the band responsible for the high kr intensity is not visible .
bulk 4f electrons are known for a stronger hybridization with the valence band electrons as compared to surface states @xcite and they may form a high intensity kr . hence , the large kr should be assigned to a band , which is not seen in the experiment and should be a bulk state .
this experimental material points to the main conclusion of this letter , namely a large variation of the kondo peak intensity along fermi surface ( fs ) .
kr is much higher for a weak intensity band at k@xmath29=0.6 @xmath16 than for the well visible @xmath27 pocket .
thus , taking into account band intensities one may conclude that the real kr height variation is even much larger .
it is also noteworthy that the intensity of the kr is lower for k@xmath29=0.9 @xmath16 than for k@xmath29=0.6 @xmath16 , wave vectors that are equivalent in the sbz . due to the circular path of the arpes experiment mentioned before these k@xmath29 vectors correspond to different values of k@xmath25 and thus different locations in the bz .
inequality of kr height in these places is consistent with the 3d nature of the related states . in order to assign the highest intensity kr to a particular band
, the results of band structure calculations are shown along @xmath24-a , m - l and the paths in between for lsda+u approximation with u@xmath32=6 ev [ fig .
3 ( b ) ] .
the wave vector 0.6 @xmath16 corresponding to the highest kr is highlighted with a dashed line .
actually , an @xmath4 crossing at 0.6 @xmath16 appears for almost all considered paths parallel to @xmath24-a .
it should be stressed that in the case of each considered path there are more e@xmath38 crossings but the highest intensity kr is observed only at the unique wave vector in arpes .
the kr absence may also be explained by unfavorable photoionization cross section for certain bands .
nevertheless , the main thesis is based on the experimental spectra , namely on the comparison of kr height for the @xmath27 band and at k@xmath29=0.6 @xmath16 .
ce@xmath0co@xmath1si@xmath2 was investigated in the coherent state .
although , the f - state dispersion was not found directly in the experiment , we assume that f - electron related quasiparticles are present .
the high intensity kr peak should refer to a high quasiparticle spectral weight , z(k ) for a given momentum .
its considerable variation along fs indicates also the momentum dependence of @xmath3 .
such a situation was proposed in the theoretical considerations for a ce based cubic system @xcite .
these predicted strongly anisotropic @xmath3 resulting in a variation of z(k ) along the brillouin zone . in the cited report a high quasiparticle weight ( z(k ) @xmath20 1 )
was found only in very small areas of fs along [ 001 ] and equivalent symmetry directions . according to the predictions
, these regions can be easily missed by arpes due to their very limited size in k - space .
although similar calculations are not realized for ce@xmath0co@xmath1si@xmath2 our arpes results support the theoretical predictions qualitatively giving the first evidence of strongly anisotropic @xmath3 in ce@xmath0co@xmath1si@xmath2 system .
in conclusion , we investigated the band structure of ce@xmath0co@xmath1si@xmath2 by means of arpes and fplo calculations with lsda+u approximation . out of a larger number of bands theoretically predicted for bulk ce@xmath0cosi@xmath13 only some can be seen in the experiment .
arpes data reveal a bulk electron pocket near the @xmath24 point ( @xmath27 ) , which is found in the calculations with u@xmath32=6 ev .
the band @xmath28 exhibiting the highest intensity is interpreted as a surface state .
the main contributions from ce 4f electrons to the photoemission spectra are the kondo resonance ( kr ) at @xmath4 associated with the f@xmath18 final state and the peak related to f@xmath19 final state observed at 275 mev .
both peaks are non - dispersing but their intensity varies as a function of the wave vector . in particular , large maximum of kr
is attributed to a specific fermi vector of a bulk band , whereas the other observed bands from the bulk or the surface are characterized with a medium or low quasiparticle weight at fermi vectors .
this represents the arpes evidence of a momentum dependent hybridization between ce 4f and conduction band electrons in ce@xmath0co@xmath1si@xmath2 .
this work has been supported by the ministry of science and higher education in poland within the grant no .
n n202 201 039 .
a part of the measurements was carried out with the equipment purchased thanks to the european regional development fund in the framework of the polish innovation economy operational program ( contract no .
poig.02.01.00 - 12 - 023/08 ) .
hs , ff and fr acknowledge the support by the dfg through for1162 . j.g .
acknowledges the financial support from the national science centre ( ncn ) , on the basis of decision no .
dec-2012/07/b / st3/03027 .
we acknowledge technical support by f. salvador ( cnr - iom ) . | we studied the electronic structure of the kondo lattice system ce@xmath0co@xmath1si@xmath2 by angle - resolved photoemission spectroscopy ( arpes ) .
the spectra obtained below the coherence temperature consist of a kondo resonance , its spin - orbit partner and a number of dispersing bands .
the quasiparticle weight related to the kondo peak depends strongly on fermi vectors associated with bulk bands .
this indicates a highly anisotropic hybridization between conduction band and 4f electrons - @xmath3 in ce@xmath0co@xmath1si@xmath2 . |
recently @xcite studied the mid- and far - infrared emission from an evolving stellar population formed as a burst with a salpeter imf .
infrared emission at @xmath7 m arises from warm dust in the stellar winds of evolved giant stars ; emission at shorter wavelengths @xmath8 m can be largely photospheric @xcite .
most of the time variation of mid - infrared emission from an evolving stellar population results from the declining rate of stellar mass loss from red giants , but @xcite include many additional details .
the mid - infrared emission is far less sensitive to stellar metallicity than age .
this raises the interesting possibility that the mean stellar ages in elliptical galaxies can be determined directly from mid - infrared observations .
we explore this possibility here using archival iso observations at 15@xmath0 m .
the mean ages of stars in elliptical galaxies can be estimated optically from the strength of the stellar balmer lines ( e.g. gonzalez 1993 , worthey 1994 , tantalo et al .
1998 , trager et al .
2000 , kuntschner et al .
these mean age determinations , accurate to 20 - 30 percent , are often interpreted by comparing optical spectra with computed time - dependent spectral energy distributions a single burst of star formation .
most of the stars in elliptical galaxies are thought to be very old , possibly @xmath9 gyrs , but unexpectedly young ages have been found for many ellipticals ( e.g. @xmath2 gyrs ) . evidently the mean stellar age has been skewed downward by the addition of a relatively small component of young stars that may result from a recent merger or internal star formation .
using archival iso data at 15@xmath0 m for 17 early - type galaxies , mostly ellipticals , we explore the possibility that warm circumstellar dust can be used to determine stellar age or the likelihood of a recent merger .
balmer line stellar ages are known for all galaxies in our sample .
we begin by verifying that the surface brightness profiles at 15@xmath0 m are similar to those in the i - band which we assume to be stellar , unaffected by dust emission .
since optically obscuring dust is often observed in the cores of elliptical galaxies , including some in our sample , we used both the integrated ( within @xmath5 ) flux ratio @xmath4 and the surface brightness ratio @xmath1 at a larger radius ( @xmath10 ) as a measure of the mean stellar ages .
the two methods give similar results .
in fact we find that the flux ratio @xmath4 within @xmath5 is similar to the ratios @xmath1 at @xmath10 for all the galaxies in our sample .
apparently , optically obscuring dust in the galactic cores which could in principle contribute to the emission at 15@xmath0 m can not be easily detected at the sensitivity and resolution of iso .
as we show below , a plot of @xmath4 against optically determined ages shows that the mid - infrared emission is indeed sensitive to stellar age provided the galaxies are no older than @xmath11 gyrs .
abnormally large @xmath4 or @xmath1 ratios , suggesting the presence of younger stars , are found in elliptical galaxies that ( 1 ) are thought to have suffered a recent merger ( 2 ) have strong optically obscuring dust , or ( 3 ) have hi or non - thermal radio emission .
the infrared camera isocam on board the iso satellite ( kessler et al .
1996 ) observed about 60 elliptical galaxies in the 4 - 15@xmath0 m region with narrow and wide band filters .
> from these observations we selected 22 early - type galaxies , 17 of which have optically determined stellar ages ( terlevich & forbes 2002 ) . the relevant data for the galaxies in our sample are listed in table 1 . for these sample galaxies
we consider only those observations taken with the smallest pixel scale to maximize the accuracy of the radial and azimuthally averaged mid - infrared emission .
mid - infrared fluxes within small radii are required to detect emission from dust in the galactic cores .
the selected observations were performed in raster mode in the lw3 ( @xmath12 m ) filter .
individual galactic images have been taken with the @xmath13 si : ga array camera with 3@xmath14 pixel scale .
the total on - source integration times for the si : ga array are typically 300 - 600 seconds .
each observation consists of a collection of exposures , each taken with the target galaxy centered at several positions on the array .
two galaxies , ngc 4552 and ic 1409 , have been observed at a plate scale of @xmath15pix and with larger pointing offsets among the individual exposures , allowing raster maps of approximately @xmath16 .
data reduction has been performed using the camera interactive analysis ( cia ) package , version 5.0 @xcite .
the procedure to reduce isocam data starts with the dark current subtraction , for which we used the _ model _ option to take into account the long term drift in the dark current which occurred during the iso mission .
then we removed cosmic ray hits using a sigma clipping filter or multi - resolution method .
this removes most glitches of short duration .
we also visually inspected the frames to manually remove glitches with very long time constants that may affect more than one frame .
si : ga devices are also affected by transient effects in which the response time to any change in illumination depends on the incident photon flux .
this transient behavior needs to be characterized for our specific data set and properly allowed for in recovering the `` stabilized '' electronic signal .
the transient correction was performed using the fouks - schubert method @xcite in which a model of the detector response to incident radiation is used to reconstruct the stabilized fluxes .
we then flat - fielded the data using standard library flats and applied the conversion from electronic units to jy / pix using the standard _ cal
_ calibration data that accompanies the data products .
after each single exposure has been properly reduced , the final image was produced by registering and coadding each frame .
all 22 galaxies were detected at 15@xmath0 m with a s / n in excess of @xmath17 . in all cases
the field of view was larger than the size of the galaxy , allowing an evaluation of the background flux directly from the galaxy maps .
the sky background was calculated by taking the median pixel value from several blank regions in the reduced frames .
we used the reduction routines in the iraf package to perform both the aperture photometry and the azimuthally averaged radial profiles .
one of the goals in this study is to test the sensitivity of mid - infrared surface brightness profiles to the age of the old stellar population in elliptical galaxies using the known age measurements from optical studies . in principle
it might be possible to determine the stellar age at any galactic radius by comparing the surface brightness profiles at 15@xmath0 m and optical wavelengths .
our measurements of the 15@xmath0 m surface brightness profiles may allow us to detect emission from dust whose spatial distribution does not follow the galactic starlight .
complementary observations of the 15@xmath0 m flux through a small aperture centered on the galactic core are designed to detect emission from optically obscuring dusty clouds observed near the centers of many elliptical galaxies in our sample .
we used the task _ ellipse _ in the iraf _ stsdas _ package to perform elliptical isophote analysis .
the isophote fitting algorithm , described by @xcite , computes best - fitting elliptical isophotes .
it applies iterative corrections to the geometrical parameters of a trial ellipse by projecting the fitted harmonic amplitude onto the image plane .
the output provides all geometrical parameters for the ellipse including the semi - major axis @xmath18 and mean isophotal intensity .
the left panels in figure 1 show the 15@xmath0 m surface brightness profiles for each galaxy as a function of the semi - major axis to the 1/4 power .
most galaxies have surface brightness profiles @xmath19 that are generally in very good agreement with a de vaucouleurs profile , which is a straight line in these plots .
the downward curvature of @xmath19 at small @xmath18 arises because the pixel size becomes too large to properly track the steep de vaucouleurs profiles there .
it has been recognized for some time that mid - infrared emission from relatively warm dust in elliptical galaxies closely follows the de vaucouleurs @xmath20 profile of optical starlight @xcite , indicating that the warm dust is photospheric or circumstellar .
the arrows in the profile plots show the location of @xmath5 for each galaxy where @xmath21 is the effective radius .
the right panels in figure 1 show the ratio of the surface brightness at 15@xmath0 m to that in the i band where dust emission is not expected .
i band profiles ( not shown here ) are taken from the studies of @xcite and @xcite . to avoid complications
when comparing the 15@xmath0 m and i - band profiles due to sampling with very different instruments and pixel scales , we discarded the innermost part of the profile of @xmath1 in figure 1 .
the radial profiles for @xmath1 are also truncated at large radii where the low s / n precludes a proper isophotal fit . in the range plotted where high s / n , low geometrical distortion , and reasonably large number of pixels ensure the most reliable data the ratio @xmath1 is approximatively constant with radius for most of the galaxies .
although the stellar metallicity ( and dust production ? ) is expected to decrease with galactic radius , this trend is not reflected in the @xmath1 profiles , which are either flat or increase slowly with projected radius .
this insensitivity to stellar metallicity is consistent with the models of @xcite .
we note that @xcite observe decreasing 15@xmath0m/7@xmath0 m with galactic radius , but dust may contribute at both these wavelengths .
two galaxies ngc 5866 and ngc 5044 have discordant 15@xmath0 m surface brightness profiles .
ngc 5866 is an approximately edge - on s0 galaxy with an extensive dusty disk .
the increase in @xmath1 for @xmath22 visible in figure 1 is due to a slope change in @xmath23 that presumably arises from the bulge - disk transition .
azimuthally averaged surface brightness profiles for this galaxy are of limited value .
it is debatable if this dusty s0 galaxy should be included in our sample of elliptical galaxies , but we have retained it to see if the dust is visible at 15@xmath0 m .
the @xmath24 profile for ngc 5044 also has a feature visible in figure 1 at @xmath25 caused by asymmetric emission toward the east and southeast that is discussed in more detail below . at visual wavelengths stellar ages and metallicities in elliptical galaxies
are evaluated from an analysis of spectral indices , assuming single stellar populations ( ssp ) .
since line index strengths in early - type galaxies have radial gradients @xcite , the representative mean ages and metallicities of elliptical galaxies refer to a fixed aperture . following @xcite and other authors , we compare 15@xmath0 m and i - band fluxes within an aperture of radius @xmath5 .
photometric data reduction of the 15@xmath0 m maps has been performed using the _ apphot _ package in iraf .
these routines are well suited to compute accurate centers and integrated fluxes within the specified circular aperture .
some galaxies in our sample have relatively small effective radii , requiring an accurate centering algorithm .
it is also essential to account for pixels that span across the @xmath5 aperture to retrieve reliable photometry just within @xmath21/8 . whenever possible , i - band fluxes within @xmath5 for our sample galaxies are found by integrating published @xmath26 profiles to this radius . for galaxies without published @xmath26 profiles we used
the total apparent corrected i - magnitude from the lyon / meudon extragalactic database ( leda ) .
then , assuming a de vaucouleurs i - band distribution , we integrated over the @xmath20 profile to determine the flux within the @xmath5 aperture using the appropriate effective radius @xmath21 for each galaxy .
the aperture and surface photometry results are presented in table 2 .
figure 2 shows the @xmath4 ratios based on the photometry inside @xmath21/8 plotted against the average stellar population age of the galaxies as tabulated by @xcite for this same aperture .
most optical ages have errors less than 20% ; age errors are not plotted for clarity .
the lines in figure 2 show the expected evolution of @xmath4 produced by ssp models from @xcite for metallicities z = 0.02 ( solar ) , z = 0.04 and z = 0.008 .
the predicted flux ratios @xmath4 from @xcite are very insensitive to metallicity variation and vary only weakly for stellar ages @xmath3 gyr .
the observed @xmath4 for galaxies with ages @xmath3 gyr suggest that the variation with stellar age is even less than that predicted by @xcite and that the normalization of the theoretical curves may be a bit too low at ages @xmath9 gyrs .
recently @xcite have provided independent optical ages for a large number of the galaxies in our sample , using models for stellar absorption indices computed with variable stellar @xmath27/fe abundance ratios . while their ages tend to be slightly larger than those quoted by @xcite , our conclusions are unaffected if galactic ages from @xcite were used instead .
some of the galaxies with ages @xmath3 gyr are known to contain central dust clouds and disks ( see table 1 ) , but there is only a limited correspondence with the incidence of optically visible dust and the position of galaxies in figure 2 . to illustrate the lack of significant 15@xmath0 m emission from central dust , we plot in figure 3 the relationship between @xmath5 aperture flux ratios @xmath4 and the corresponding surface brightness ratios @xmath1 evaluated at @xmath10 , regarded as the ratio for the stellar emission alone .
if central dust clouds contribute to the 15@xmath0 m emission , this would be revealed in @xmath4 but not in @xmath1 .
the data points in figure 3 lie along the 1:1 line with very little scatter , suggesting that the vertical spread of the data showed in figure 2 can not be explained by central dust clouds .
one of the elliptical galaxies with the largest positive deviation in figure 3 is ngc 5831 which contains no dust visible with hst .
if small , centrally located dust is present in these galaxies , its effect on the 15@xmath0 m emission is not prominent . for stellar ages
@xmath28 gyr the @xmath4 curves in figure 2 from @xcite are very sensitive to stellar age and in general terms this is supported by our observations .
three galaxies in figure 2 ngc 5866 , ngc 6776 and ngc
1316 lie far above the predicted lines for their optically determined ages .
three additional galaxies ngc 1453 , ngc 4278 and ngc 4261 are marginally higher .
we have already mentioned the unusual s0 galaxy ngc 5866 , a galaxy with a massive dusty disk of cold gas that has been detected in hi and co emission ( roberts et al .
1991 ; welch & sage 2003 ) , so it is not surprising that it has an additional component of non - circumstellar mid - infrared dust emission . of the remaining galaxies , four ngc 1316 , ngc 1453 , ngc 4261 and ngc 4278 show evidence of recent mergers , have significant hi ( which is unusual for elliptical galaxies ) or have strong radio emission and dusty disks
. however , the remaining galaxy , ngc 6776 , seems to be rather normal but does have a faint shell that may be merger - related . on balance , therefore , it appears that 15@xmath0 m observations can be used to identify post - merger ellipticals but the theoretical ssp flux ratios plotted in figure 2 may unfortunately be too simplistic to fully describe the complexities of post - merger dust formation .
the agreement of ngc 5018 with the @xcite predictions may be fortuitous since there is some evidence of hi and radio emission from this galaxy .
however , it appears to be significant that all four of the ellipticals with the lowest values of @xmath4 in figure 2 ngc 5831 , ngc 720 , ngc 2300 and ngc 4649 have been observed with hst and found to contain no visible central dust .
in figure 4 we show 15@xmath0 m images of six galaxies from our sample that have low luminosity extensions in various directions from the emission peak which is coincident with the optical galaxy . with a pixel scale of 3@xmath14
the isocam beam size in the lw3 filter is approximately 5@xmath14 fwhm .
thus , some of the features in the central 30@xmath14 of ngc1316 and ngc1453 are clearly resolved . in the remaining galaxies ,
other non - axisymmetric emission features are apparent but are not resolved by these observations . apart from ngc 5044
, the azimuthally averaged 15@xmath0 m surface brightness profiles tend to wash out the faint isophotal extensions visible in figure 4 . in general ,
we have been unable to find emission at other frequencies that correlates with these mid - infrared features .
the exception to this is ngc 1316 ( fornax a ) a galaxy thought to have had recent merger ( schweizer 1980 ; geldzahler & fomalont 1984 ) . dust and gas appear to be falling into ngc 1316 , forming a disk - like pattern @xmath30 in radius in the se - nw direction , roughly parallel to the jet radio axis @xcite . the extended emission in our 15@xmath0 m image is entirely consistent with emission from optically visible dust in these same regions , confirming the interpretation of @xcite .
recently the spitzer space telescope @xcite has observed ngc1316 with the irac @xcite instrument in four mid - infrared bands at 3.6 , 4.5 , 5.8 and 8.0@xmath0 m . the higher sensitivity and spatial resolution of the camera allow us to better resolve the emission in the central @xmath31 of the galaxy .
the 8@xmath0 m image in figure 5 ( left panel ) shows in great detail and at higher s / n the same asymmetrical features present in the iso 15@xmath0 m image ( see also @xcite ) .
many bright knots and an arc - like structure are prominent and well resolved by irac . the right panel in fig .
5 shows a contour plot of the emission at 8.0@xmath0 m once the stellar continuum is subtracted using a scaled 3.6 @xmath0 m image . the non stellar emission
is then superimposed to a b
i hst image ( grayscale ) ( press release stsci - pr99 - 06 , grillmair et al .
1999 ) where the dust distribution is shown in dark extended features .
the correspondence of the non
stellar 8@xmath0 m emission features with the dust seen in absorption by hst is remarkable .
almost every feature visible in the optical absorption study , included the central one , has 8@xmath0 m emission associated . the 8.0@xmath0 m broadband filter sample the emission band centered at `` 7.7@xmath0 m '' from polycyclic aromatic hydrocarbons ( pah ) molecules .
much of the excess emission in this band must then be attributed to pahs , but the coincidence of the bright spots with the 15@xmath0 m image in the region where the spitzer and iso data overlap is indicative that warm small dust grains also contribute to the mid - infrared emission .
perhaps young , luminous stars are forming in the infalling dusty gas , raising the dust temperature sufficiently to emit at 15@xmath0 m .
it is interesting that dust located at large radii in ngc1316 is associated with 15@xmath0 m emission , while optically visible dust in the galactic cores is not .
this may occur because only stars of rather low ( jeans ) mass can form near the galactic cores where the gas pressures are very high .
but the mass of molecular hydrogen in the dusty patches along the central se - nw direction in ngc 1316 , @xmath32 @xmath33 @xcite , exceeds by @xmath34 the mass of dusty cold gas typically observed in elliptical galaxy cores , and this may be the best explanation for the associated star formation .
finally , it seems odd that the low luminosity extensions in the 15@xmath0 m image of ngc 5044 , which we believe to be real , are strong enough to alter the surface brightness profile but still have no easily identifiable correlated emission at other frequencies .
we have found that the infrared emission at 15@xmath0 m from elliptical galaxies , normalized by the i - band emission , can provide an approximate measure of the mean age of galactic stars .
similar to the predictions of @xcite , the central flux ratios @xmath4 generally increase for ( 1 ) ellipticals known to have stellar ages @xmath2 gyrs and particularly for ( 2 ) ellipticals that exhibit unusual attributes associated with recent mergers : dusty disks , hi emission , non - thermal radio emission , shells , etc .
however , if only normal , non - merging galaxies are considered , the central flux ratios @xmath4 for galaxies in our sample shows little or no convincing variation with optically determined stellar ages . since reliable estimates of the dust mass in elliptical galaxies
are not available , it is not possible to quantify the excess 15@xmath0 m emission expected from a given amount of dust .
however , the four ellipticals having the lowest observed values of @xmath4 all appear to be dust - free in hst observations .
some individual galaxies seem inconsistent with these general trends .
ngc 6776 has an unusually large @xmath4 , but only modest optical evidence for a recent merger .
ngc 7626 and ngc 4374 have optically obscuring central dust and double radio lobes but rather normal @xmath4 .
the strong correlation between the flux ratio @xmath4 measured within @xmath5 and @xmath1 evaluated at @xmath10 indicates that 15@xmath0 m emission from optically visible dust in the cores of some of our sample galaxies is not detectable with iso .
this contrasts with the more distant dust in the merging galaxy ngc 1316 which is a local source of 15@xmath0 m emission .
evidently , massive dust - heating stars are forming in ngc 1316 , but not in the dusty cores of many other ellipticals .
as in ngc 1316 , five additional elliptical galaxies show non - stellar extended features in their 15@xmath0 m images . while we believe these features are real , they do not appear associated with emission at other wavelengths
. .4 in we thank l. piovan for providing electronic form of the ssp model outputs .
studies of hot gas and dust in elliptical galaxies at uc santa cruz are supported by nasa grants nag 5 - 8409 & atp02 - 0122 - 0079 and nsf grants ast-9802994 & ast-0098351 for which we are very grateful .
pt is supported in part by nasa / adp grant 21 - 399 - 20 - 01 .
fb is supported in part by a grant miur / prin 0180903 .
this research is based on observations with iso , an esa project with instruments funded by esa member states and with the participation of isas and nasa .
part of this work is based on observations made with the _
spitzer space telescope _ , which is oparated by jet propulsion laboratory , california institute of technology , under nasa contract 1407 .
ngc0720 ........ & 20.80 & -4.8 & e5 & 4.03 & 10.38 & 3.4 & & no & a prototype e5 galaxy central galaxy in fornax ; + & & & & & & & & & cluster ; strong color gradient in stars ; central + & & & & & & & & & color is among blueset of e galaxies + ngc1316 ........ & 18.11 & -1.7 & sabs(0 ) & 7.08 & 10.93 & 3.4 & & yes & evidence of recent merger in h@xmath27 , dust ; @xmath35 detected + & & & & & & & & & in dust patches ; fornax a ( strong radio source ) ; + & & & & & & & & & evidence of merger arp 154 ; 3 gyr old merger + & & & & & & & & & remnant ; 0.5 gyr old merger remnant .
+ ngc1399 ........ & 18.11 & -4.5 & e1 & 3.78 & 10.52 & & & no & iras100@xmath0 m ; low luminosity radio jet .
+ ngc1453 ........ & 43.10 & -5.0 & e23 & 5.85 & 10.37 & 7.6 & & & gas disk ; radio ; hi .
+ ngc2300 ........ & 27.67 & -3.5 & sa0 & 4.83 & 10.41 & 5.0 & & no & shells ; nuclear radio ; a normal e3 .
+ ngc4261 ........ & 31.48 & -4.8 & e3 & 5.65 & 10.70 & 14.4 & & yes & powerful double lobe radio ( 3c270 ) ; radio jets + & & & & & & & & & perpendicular to disk ; strong central optical + & & & & & & & & & emission lines ; uv emission from jet ; + ngc4278 ........ & 16.22 & -4.8 & e1 & 2.60 & 10.24 & 10.7 & yes&yes & large rotating hi+dust disk ; strong + & & & & & & & & & nuclear liner emission lines ; core - halo + & & & & & & & & & radio source with compact core ; + & & & & & & & & & optically visible dust patches ; + ngc4374 ........ & 15.92 & -4.0 & e1 & 4.25 & 10.57 & 11.8 & yes&yes & classical double radio ; no large scale dust at + & & & & & & & & & submm wavelengths ; nuclear dust lane ; + & & & & & & & & & core - jet radio source perp .
to lane ; + & & & & & & & & & one - sided radio `` jet '' ; m84 ; 3c272.1 + ngc4472 ........ & 15.92 & -4.7 & e1/s01(1 ) & 8.03 & 10.90 & 8.5 & yes&yes & radio ; hi ; 2-sided weak radio jet ; + ngc4473 ........ & 16.14 & -4.8 & e5 & 1.96 & 10.15 & & & no & very small edge - on dust and gas disk but not see + & & & & & & & & & with hst ; prototypical e5 galaxy ; + ngc4552 ........ & 15.92 & -4.6 & s01(0 ) & 2.32 & 10.29 & 9.6 & & yes & @xmath36 undetected at @xmath37 ; + & & & & & & & & & modest radio source ; m89 ; + ngc4636 ........ & 15.92 & -4.8 & e0/s01(6 ) & 7.71 & 10.51 & & & yes & nicmos image is smooth , showing no dust + & & & & & & & & & deviates from fund .
plane for e cores ; + & & & & & & & & & radio core plus jets ; + & & & & & & & & & moderately strong broad @xmath36 ; + & & & & & & & & & @xmath36 morphology is asymmetric with hole in center ; + ngc4649 ........ & 15.92 & -4.6 & s01(2 ) & 6.50 & 10.73 & 11.0 & & no & radio ; @xmath35 ; hi ; no fir dust ; + & & & & & & & & & @xmath36 undetected at @xmath38 ; m60 ; + ngc5018 ........ & 30.20 & -4.5 & e3 & 3.81 & 10.57 & 1.5 & & yes & radio ; hi ; dust patches ; + + ngc5044 ........ & 30.20 & -4.8 & e0 & 10.40 & 10.70 & & yes&yes & nuclear radio source ; very bright , asymmetric + & & & & & & & & & diffuse optical line emission ; dusty region is + & & & & & & & & & smaller ; bright dust cloud to east of center ; + ngc5831 ........ & 22.91 & -4.8 & e3 & 3.11 & 10.03 & 2.6 & & no & + ngc5866 ........ & 13.18 & -1.3 & s03 & 4.98 & 10.32 & 1.8 & & big disk & core radio source ; this s0 galaxy has a + & & & & & & & & & strong dust lane along major axis ; + & & & & & & & & & contains hi ; @xmath36 emission ; + ngc6776 ........ & 70.41 & -4.1 & e+ & 9.21 & 10.66 & 3.2 & & & a normal e1 with a very faint shell ; + & & & & & & & & & 843mhz radio ( very faint ) ; + ngc7562 ........ & 39.99 & -4.8 & e23 & 4.65 & 10.46 & 11.0 & & & hi ; a pair with ngc7562a and ngc5619 + & & & & & & & & & with possible tidal connection ; + & & & & & & & & & faint hi detected in old literature ; + ngc7619 ........ & 39.99 & -4.8 & e & 6.20 & 10.58 & 9.0 & & & radio ; a normal e3 ; + ngc7626 ........ & 39.99 & -4.8 & e & 7.37 & 10.61 & 11.7 & yes&yes & radio core - jet with double lobes ; + & & & & & & & & & nuclear dust lane ; + ic 1459 .......... & 18.88 & -4.7 & e1 & 3.57 & 10.37 & & yes&yes & chaotic dust absorption visible ; galaxy + & & & & & & & & & crossed by a disk of ionized gas ; + & & & & & & & & & hi not detected , but co ( @xmath35 ) is ; heavy dust + & & & & & & & & & lane in @xmath39 with more chaotic and patchy dust + & & & & & & & & & further out ; central starlight core is bluer ; + & & & & & & & & & ionized gas disk ; stellar shells ; bright point source + & & & & & & & & & in center ; powerful compact radio core source ; + ngc0720 ........ & 6.8 @xmath40 0.8 & 46.3 @xmath40 11.3 & 0.15 @xmath40 0.04 & 0.17 + ngc1316 ........ & 85.5 @xmath40 7.4 & 157.8@xmath40 22.9 & 0.54 @xmath40 0.09 & + ngc1399 ........ & 12.0 @xmath40 1.3 & 78.0 @xmath40 8.7 & 0.15 @xmath40 0.03 & 0.16 + ngc1453 ........ & 4.2 @xmath40 0.5 & 13.6 @xmath40 3.6 & 0.31 @xmath40 0.05 & + ngc2300 ........ & 5.6 @xmath40 0.8 & 36.5 @xmath40 5.4 & 0.16 @xmath40 0.04 & 0.16 + ngc4261 ........ & 15.3 @xmath40 1.4 & 55.0 @xmath40 7.1 & 0.28 @xmath40 0.05 & 0.24 + ngc4278 ........ & 14.2 @xmath40 1.2 & 55.1 @xmath40 9.2 & 0.26 @xmath40 0.05 & 0.25 + ngc4374 ........ & 27.6 @xmath40 2.6 & 136.0@xmath40 20.2 & 0.20 @xmath40 0.04 & 0.19 + ngc4472 ........ & 59.6 @xmath40 6.5 & 310.0@xmath40 49.1 & 0.19 @xmath40 0.03 & 0.16 + ngc4473 ........ & 7.0 @xmath40 0.9 & 49.0 @xmath40 7.1 & 0.14 @xmath40 0.03 & 0.14 + ngc4552 ........ & 16.9 @xmath40 2.1 & 74.0 @xmath40 14.4 & 0.23 @xmath40 0.05 & 0.17 + ngc4636 ........ & 37.0 @xmath40 1.1 & 121.0@xmath40 15.6 & 0.31 @xmath40 0.04 & 0.29 + ngc4649 ........ & 35.8 @xmath40 3.9 & 214.0@xmath40 29.8 & 0.17 @xmath40 0.03 & 0.17 + ngc5018 ........ & 11.8 @xmath40 1.1 & 29.0 @xmath40 6.0 & 0.41 @xmath40 0.04 & 0.43 + ngc5044 ........ & 8.1 @xmath40 0.9 & 42.0 @xmath40 5.8 & 0.19 @xmath40 0.04 & 0.20 + ngc5831 ........ & 3.5 @xmath40 0.7 & 19.0
@xmath40 3.3 & 0.18 @xmath40 0.05 & 0.14 + ngc5866 ........ & 62.0 @xmath40 3.2 & 63.0 @xmath40 8.5 & 0.98 @xmath40 0.14 & 1.10 + ngc6776 ........ & 5.7 @xmath40 0.5 & 9.9 @xmath40 2.2 & 0.58 @xmath40 0.13 & + ngc7562 ........ & 4.3 @xmath40 0.5 & 19.8 @xmath40 3.9 & 0.22 @xmath40 0.05 & + ngc7619 ........ & 5.7 @xmath40 0.6 & 25.0 @xmath40 4.1 & 0.23 @xmath40 0.04 & + ngc7626 ........ & 4.0 @xmath40 0.7 & 20.4 @xmath40 4.2 & 0.20 @xmath40 0.05 & + ic 1459 ........... & 12.0 @xmath40 1.9 & 57.0 @xmath40 6.2 & 0.21 @xmath40 0.04 & 0.19 + | the mid - infrared ( 10 - 20@xmath0 m ) luminosity of elliptical galaxies is dominated by the integrated emission from circumstellar dust in red giant stars .
as a single stellar population evolves , the rate of dusty mass loss from red giant stars decreases with time , so the mid - infrared luminosity should also decline with stellar age .
to seek such a correlation , we have used archival iso observations to determine surface brightness profiles and central fluxes at 15@xmath0 m in 17 early - type galaxies for which stellar ages have been determined from optical spectral indices .
the radial surface brightness distributions at 15@xmath0 m generally follow the stellar de vaucouleurs profile as expected .
we find that the surface brightness ratio @xmath1 is systematically higher in elliptical galaxies with ages @xmath2 gyrs and in galaxies that exhibit evidence of recent mergers . within the accuracy of our observations
, @xmath1 shows no age dependence for ages @xmath3 gyrs .
the corresponding flux ratios @xmath4 within apertures scaled to the effective radius ( @xmath5 ) are proportional to the @xmath1 ratios at larger galactic radii , indicating that no 15@xmath0 m emission is detected from central dust clouds visible in optical images in some of our sample galaxies .
emission at 15@xmath0 m is observed in non - central massive clouds of dust and cold gas in ngc 1316 , an elliptical galaxy that is thought to have had a recent merger .
recent _
spitzer space telescope _ data also indicate the presence of pah emission at @xmath6 m .
several ellipticals have extended regions of 15@xmath0 m emission that have no obvious counterparts at other frequencies . |
after several decades of research , planets in stellar binary systems now constitute a well - established observational result .
previous examples include p - type orbits , when the planet is found to orbit both binary components , as well as s - type orbits with the planet orbiting only one of the binary components with the second component acting as a perturbator ; see , e.g. , @xcite and @xcite for selected observational results and data .
another topic of significant importance , especially concerning the astrobiological community , are studies of circumstellar and circumbinary habitability .
previous work focused on the traditional concept of @xcite and subsequent studies , where habitability is defined based on the principal possibility that liquid water is able to exist on the surface of an earth - type planet possessing a co@xmath0/h@xmath0o / n@xmath0 atmosphere ; for more sophisticated concepts see , e.g. , @xcite , and references therein .
other relevant investigations concern studies of orbital stability , especially for ( hypothetical ) earth - type planets in stellar habitable zones .
this type of work has been pursued for single as well as multi - planetary and multi - stellar systems ; see , e.g. , @xcite , @xcite , @xcite , for early contributions as well as , e.g. , @xcite and @xcite for more recent work .
some of these efforts resulted in stability catalogs of the habitable zones of the planetary systems known at the time .
@xcite [ paper i and ii ] forwarded a concise approach for the investigation of habitable regions in stellar binary systems , which forms the basis for binhab ( see sect . 2 ) . in sect . 3
, we will give applications to s / st - type habitability for binaries of low mass stars .
our conclusions and outlook are presented in sect .
the method as conveyed has previously been given in paper i and ii ; thus in the following we will focus on the most decisive concepts , which include : ( 1 ) the consideration of a joint constraint comprising orbital stability and a habitable region for a putative system planet through the stellar radiative energy fluxes ( radiative habitable zone " ; rhz ) needs to be met .
( 2 ) the treatment of conservative , general and extended zones of habitability for the various systems , referred to as chz , ghz , and ehz , respectively , following the approach given by @xcite and subsequent work .
( 3 ) the providing of a combined formalism , based on solutions of a fourth - order polynomial , for the assessment of both s - type and p - type habitability .
in particular , mathematical criteria are presented for which kind of system s - type and p - type habitability is realized .
following paper i , five different cases of habitability are identified , which are : s - type and p - type habitability provided by the full extent of the rhzs ; habitability , where the rhzs are truncated by the additional constraint of planetary orbital stability ( referred to as st and pt - type , respectively ) ; and cases of no habitability at all . regarding the treatment of planetary orbital stability ,
the formulae of @xcite are utilized .
binhab is suitable for both circular and elliptical stellar binary components , the topic of paper ii .
figure 1 conveys the flow diagram .
note that for s - type orbits , the orbital stability criterion operates as an upper limit of orbital stability ( see sect .
3 ) , whereas for p - type orbits , it operates as a lower limit of planetary orbital stability . binhab allows to consider general binary systems , including systems containing evolved stars .
surely , for main - sequence stars , there is an intimate coupling between the input parameter t@xmath1 and r@xmath1 ( @xmath2 ) , the stellar effective temperature and radius , on the one hand , and m@xmath1 , the stellar mass , on the other hand , which is however not required for the usage of the code .
the simulations of paper i are largely based on data given in @xcite . to make binhab publicly accessible
, we created a website through which stellar system conditions could be entered , and the equations pertinent to binhab would then be used to find whether any habitable zone or zones exist , and if so , what type . a virtual server was set up and hosted by the ut arlington it department at http://physbinhab.uta.edu , as a dedicated server for the binhab website .
it allows the user to enter binary system parameters ( i.e. , semi - major axis and eccentricity ) , separate stellar parameters for the two stars ( i.e. , temperature , luminosity , and mass ) , and the type of habitable zone the user would like to look for ( i.e. , chz , ghz , or ehz ) .
background information on the different types of habitable zones is also given .
the output states whether there were any habitable zones found , and if so , what type ( s , st , p , or pt ) and the inner and outer limit of any zones .
the website itself uses html ( hypertext markup language ) to interface with the user , by displaying information , accepting user input parameters , and displaying the results of the binhab calculations .
the user s input is passed to php ( php : hypertext preprocessor ) code , which also checks the validity of the input .
if any of the parameters are out of range , the php code creates a warning message , which is passed to the html code and displayed to the user , explaining what needs to be adjusted . if all parameters are acceptable , the php code saves them to a new input file , runs the binhab fortran binary on the input file , and collects the output from the binary . the output is then checked , formatted , and passed to the html code , for display to the user .
the use of php allows for much more processing of the input than is possible with html alone and is necessary for running the fortran binary .
fortran is used for implementing the binhab algorithms , rather than php , for a variety of reasons , including that it is much more widely used in the scientific community and that as a compiled language , it executes much faster than the interpreted php code .
additionally , the fortran code and input files are stored in a segment of the server directory structure that is inaccessible to the website , but accessible to the php code .
hence , the fortran code is not viewable or downloadable from the web , as it can only be accessed by the php code , which runs on the server , and therefore is not directly viewable from the internet . in order to make the website easier to find
, we successfully submitted the url to the two largest search engines : google and bing ( which also drives the yahoo ! search engine ) .
in order to illustrate the capacity of the method , we convey studies of s / st - type habitability for binaries of low - mass main - sequence stars . here
we focus on stars of masses 0.75 m@xmath3 , 0.65 m@xmath3 , and 0.50 m@xmath3 , corresponding to spectral types of k2v , k6v , and m0v .
this type of work is motivated by the observational finding that stars with masses below about 0.8 m@xmath3 constitute nearly 90% of all stars in the milky way ( e.g. , * ? ? ?
* ; * ? ? ?
as an example we focus on s - type and st - type habitable zones in binaries consisting of these types of stars .
in particular , we investigate the role of the eccentricity of the binary system @xmath4 on the width of the s / st - type habitable zones ( if existing ) for systems with @xmath5 au as examples while focusing on results for the ghz ( see fig . 2 ) .
the following aspects are identified : for all stellar mass combinations , stellar habitable zones are found for eccentricities below 0.20 .
for the majority of models pursued , s - type habitability is truncated due to the orbital stability requirement of the putative planet resulting in st habitability classification ( see table 1 )
. stellar pairs with masses of 0.75 m@xmath3 exhibit the broadest rhzs due to their relatively high luminosities ; however , in this case the relatively large stellar masses lead to a significant truncation of circumstellar habitability , resulting in relatively narrow habitable zones . on the other hand ,
stellar pairs with masses of 0.50 m@xmath3 show s / st - type habitable regions up to binary eccentricities of 0.65 ( see table 2 ) , albeit the widths of the habitable zone are relatively small compared to other primary secondary mass combinations , especially for cases of small binary eccentricities
. starting at a well - defined eccentricity , the width of the habitable zones decreases linearly as a fraction of the binary eccentricity for all mass combinations , encompassing both systems of equal and nonequal stellar masses , owing to the truncation criterion @xcite .
consistently smaller widths for the habitable regions are identified for models based on chzs relative to ghzs , as expected .
lccccccc model & 0.0 & 0.1 & 0.2 & 0.3 & 0.4 & 0.5 & 0.6 + m@xmath6 m@xmath3 , m@xmath7 m@xmath3 & st & st & st & ... & ... & ... & ... + m@xmath6 m@xmath3 , m@xmath8 m@xmath3 & st & st & st & ... & ... & ... & ... + m@xmath6 m@xmath3 , m@xmath9 m@xmath3 & st & st & st & ... & ... & ... & ... + m@xmath10 m@xmath3 , m@xmath11 m@xmath3 & s & st & st & st & st & ... & ... + m@xmath10 m@xmath3 , m@xmath9 m@xmath3 & s & st & st & st & st & st & ... + m@xmath12 m@xmath3 , m@xmath9 m@xmath3 & s & s & s & s & st & st & st + lcc model & chz & ghz + m@xmath6 m@xmath3 , m@xmath7 m@xmath3 & 0.12 & 0.20 + m@xmath6 m@xmath3 , m@xmath8 m@xmath3 & 0.15 & 0.23 + m@xmath6 m@xmath3 , m@xmath9 m@xmath3 & 0.19 & 0.27 + m@xmath10 m@xmath3 , m@xmath11 m@xmath3 & 0.43 & 0.48 + m@xmath10 m@xmath3 , m@xmath9 m@xmath3 & 0.46 & 0.51 + m@xmath12 m@xmath3 , m@xmath9 m@xmath3 & 0.62 & 0.65 +
we provided a short description of the features and capacities of the numerical tool binhab hosted at the university of texas at arlington .
it considers a joint constraint including orbital stability and a habitable region for a putative system planet through the stellar radiative energy fluxes , among various other desirable features .
although the code has previously mostly been used to investigate binary systems consisting of main - sequence stars @xcite , it is highly flexible ; it can also be utilized for the calculation of habitable zones for systems containing a subgiant , giant , or supergiant . concerning the latter ,
binhab has already been used to study the habitability of earth - mass planets and moons in the kepler-16 system @xcite , known to host a circumbinary saturn - mass planet .
ultimately , it is our goal to expand the developed methods to multiple stellar systems , which are of notable interest to the scientific community . | the aim of this contribution is to introduce the numerical tool binhab , a publicly accessible code , available at the university of texas at arlington , that allows the calculation of s - type and p - type habitable zones of general binary systems . |
this work is supported by temasek laboratories at national university of singapore through the dsta project pod0613356 . | we propose a method to investigate modular structure in networks based on fitted probabilistic model , where the connection probability between nodes is related to a set of introduced local attributes .
the attributes , as parameters of the empirical model , can be estimated by maximizing the likelihood function of the observed network .
we demonstrate that the distribution of attributes provides an informative visulization of modular networks on low - dimensional space , and suggest the attribute space can be served as a better platform for further network analysis .
networks are widely used to model complex systems with many interactive units@xcite .
usually , each node in such a network represents a distinct individual , and a link is established based on certain measurement of interaction between particular pair of nodes . on the microscopic level ,
the underlying system is fully described by state of each unit defined by several local properties , which also determines the interactions among them through complicated coupling .
thus , the network would be completely determined if local properties of nodes and interactive functions were known . in many applications however , the representative network is the only available data .
it is the purpose of the researchers to infer these crutial information on the underlying system through network analysis .
for instance , if the interaction between two units mainly depends on their local properties , then it is reasonable to expect that the units which have similar connection patterns in network representation , will share some common features in their local properties , and therefore may have similar functioning in the underlying system .
although very friutful , the applicability of this type of analysis is limited by the gap existing between the network - level description and the underlying system .
this is due to the fact that in most situations , the representitive network only models relationships or interactions among units , but not the associated properties or states of units which determine interactions .
so for example , a minor change on the observed network may not be caused necessarily by small perturbations on local properties of units in the underlying system , since interactions may depend on local states in a highly nonlinear way .
similarly , the evolution of the unerlying system due to continuous change of local states may result in abrupt changes on the network structure , such as group merge or split .
to better deal with these problems , a necessary step is to bridge these two deffierent description levels .
however it is generally impossible to completely reconstruct the intrinsic properties of nodes based solely on the strucure of a given network , since the mechanism determining a link may be very complicated . in this paper , we develope a method to describe the system in a middle level between the representitive network and the microscopic description through some simplifications .
we regard the representitive network as one particular realization of an emprical probabilistic model . in this model
, each pair of nodes has certain probability to be connected , and the connection probabilities depends on the distance between local attributes of the corresponding nodes through some function .
the attributes , act as model parameters , can be estimated according to some statistical criterion to best interpret the observed connections . since each data point in the space of attributes
is associated with distinct node , their configuration provides an alternative representation of the observed network .
the advantages of introducing this new representation are twofolds . on the one hand
, the configuration of the attributes provides an informative projection of the network on a low - dimensional space .
the established attribute space can be taken as new platform for further network analysis , where the attribute vectors which have no difference with conventional data source , allow many well developed clustering techniques to be applied directly .
on the other hand , the introduced local attributes are closely related to the unknown properties of the unerlying system by the common observed network .
it may be easier to model evolution of the underlying system as changes of attributes , and study the final influences on network structure through the empiral probability model . in this paper
, we mainly focus on studying modular strucutres in networks . after describing technique details of modeling approach
, we apply the proposed method to some artificial and real - world networks to demonstrate its usefulness on network visualzing and structure analysis .
we also discuss a possible way to extend the proposed technique to deal with more complicated multi - layered modular network .
assume there is an imaginary probabilistic system with @xmath0 nodes , which is characterized by a connection probability matrix @xmath1 , where @xmath2 represents the probability of connection between node @xmath3 and @xmath4 .
the given network described by adjancency matrix @xmath5 is regarded as a realization of this imaginary system . in particular , @xmath6 is treated as a random variable and its observed value is determined by a bernoulli trial according to the probability @xmath2 . for each node in the imaginary system
, there associates a set of local attributes denoted by @xmath7 , where the dimension @xmath8 .
the connection probability @xmath2 is then assumed to depend only on the local attributes @xmath9 and @xmath10 , and can be written as @xmath11 .
obviously , this is a great simplication , and should be regarded as first order approximation .
the function @xmath12 tying connection probability @xmath2 and local attributes @xmath9 and @xmath10 should be chosen depending on the problem at hand . in our study
, we are mainly interested in undirectional unweighted modular networks . considering the symmetric constraint ,
we choose a function @xmath12 which depends only on the distance @xmath13 .
@xmath14 will be closed to @xmath15 when @xmath16 approaching to zeros , and decays rapidly with increasing @xmath16 .
a natrual choise of @xmath12 is in guassian form : @xmath17 , where @xmath18 , defined as @xmath19 is used as a rescaling factor to compensate the distortion when projecting the high dimensional network into a low - dimensional euclidian space due to the inhomogeneity of the degree distribution of the network .
the connection probability thus becomes @xmath20 .
therefore , we have @xmath21 under this framework , the logarithmic likelihood function of observing particular network @xmath22 is @xmath23 and @xmath24 naturally , attributes @xmath25 which maximize above likelihood function are desired .
these maximum likelihood estimates can be obtained by minimizing @xmath26 following steps below : 1 .
starting from arbitary initials @xmath27 , 2 .
caculate the derivatives @xmath28 , 3 .
search in this direction to get @xmath29 , so that @xmath30 gives smallest @xmath26 , 4 .
set @xmath31 , and go back to step 2 .
5 .
quit if stopping conditions are met .
it should be noded that the explicitly casted functional relationship of attributes and connection probabilities may not be the true undelrying mechanism .
this approach nevertheless , can be applied as a useful technique to explore the desired structure in the network , providing the fitted model is good . [ cols="^,^ " , ] c + ( c ) c + ( a ) + + ( b ) + [ fig : fig7 ] american college football teams network has many clusters involved .
the fitted model nevertheless successfully capture this complicated structure .
the amazing representation of the network in atribute space ( shown in figure 5(a ) ) not only manifests clustering structures , but also suggests that several wandering points may not be clearly classfied .
it turns out that these nodes belongs to specific group ( ia independents ) and can not be classifed as one cluster consistently . furthermore , the particular node in the group of conference usa which is classifed as member of the group of western athletic is not due to the flaw of the method but caused by the network construction@xcite .
the attributes representation also suggests some groups may be further divided into smaller subgroups such as the mid american group .
this observation is well supported by the connection patterns of nodes in this group .
these useful information can not be easily obtained by conventional network clustering algorithms , and show the advantages of the proposed modeling approach .
the dolphins social network is another widely studied example .
it can be divided into two big clusters .
one of the cluster may be further divided into three small group as studied in @xcite .
the analysis results for dolphins social network are shown in figure 6 .
the distribution of the attributes correctly reflects @xmath32 large clusters ( as shown in figure 6(b ) ) , but also indicates that there may exists @xmath32 or @xmath33 smaller clusters in the left cluster .
the interesting observation is that the attributes of the nodes corresponding to three small clusters suggested in @xcite are indeed well grouped . also , the surrogate network generated by the fitted model ( as shown in figure 6(c ) ) shows striking similarity with the observed network .
decaying behaviors of fitting errors when the optimization procedure iterated unveil some common feature , which can be seen from the subplots ( upper right ones ) in all examples studied as shown in figure 4(c ) , figure 5(c ) and figure 6(c ) .
it can be divided into two segments , a rapid decrease phase and a gradual change phase . after closely monitoring the movement of the points in attribute space step by step ,
we find that the points are always firstly arranged according to globe structure of the network , and then fine adjusted within each cluster to generate better configuration .
this may partially explained why the suboptimal solutions due to gradient based algorithm also give out good configuration in general .
although modular structures are most commonly observed and studied , there are many different other structures . in our study
, we choose gaussian function to relate local attributes to connection probability , which is particular useful to capture clustering structure .
the proposed modeling approach nevertheless is flexible enough to deal with other structures providing they can be well defined . for example , for a bipartite network , the function @xmath34 , where @xmath12 is similar to what used in the paper , may be a better choise .
the proposed modeling approach can be easily extended to deal with more complicated grouping structure than simple modular structure .
let us consider a simple exention of overlapping two clustering structure .
for example , consider a group of students who make friends based on different factors such as personality or avocation . now
even if each friendship network based on any single factor were well clustered , the observed overall network may have structure quite different from the simple modular network .
we call this kind of network multi - layered modular network .
one such example of @xmath32-layered network is shown in figure 7(a ) .
this particular network is generated by the following way .
suppose the nodes are numbered from @xmath15 to @xmath35 .
we first generate two modular networks separately .
the first modular network has two clusters , one containing node @xmath15 to node @xmath36 and the other containing node @xmath37 to node @xmath35 .
the connection probabilities are @xmath38 for nodes in same cluster and @xmath39 in different clusters .
the second modular network also has two clusters , one containing node @xmath40 to node @xmath41 , and the other containing nodes @xmath15 to @xmath42 and nodes @xmath43 to @xmath35 .
the connection probabilities are slightly different , which are @xmath44 for nodes in same cluster and @xmath45 for nodes in different clusters .
the final network is generated by stacking two modular networks together , and removing repeated links .
to explore such multi - layered structure by the proposed modeling approach , the most straight way is to extend the dimension of the attributes vector .
let @xmath46^t$ ] the new extended attributes .
for each layer , the connection probability can be written as @xmath47 and @xmath48 .
the connection probability of the observed network will be @xmath49 .
following the same procedure described above , we get the attributes representation in extended attribute space .
the distribution of the attributes are shown in figure 7(b ) .
interestingly , the hidden clustering structure are unveiled in different subspace of the extended attribute space . in summary , we proposed a probabilistic modeling approach to analyze networks . under this framework
, the observed network can be regarded as a measurement of certain probabilistic system , where the connection probability of any pair of nodes depends on the properly rescaled distance between the introduced local attributes of the corresponding nodes .
it is remarkable that the configuration of the optimally estimated attributes well represents the intrinsic structure of the observed network , thus provides an very informative way to visualize networks in low - dimensional space
. it can be more effective to make further network structure analysis based on the attribute vectors instead of observed network directly .
the modeling approach can be easily extended to deal with more complicated structures such as multi - layered clustered network . |
among all the fundamental particles the neutrino occupies a unique place in many ways and as such it has shed light on many important aspects of our present understanding of nature and is believed to hold the key to the physics beyond the standard model .
presently , the question of whether the neutrino has mass is one of the outstanding issues in particle physics , astrophysics , cosmology and theoretical physics in general .
there are several theoretical , observational and experimental motivations which justify the searching for possible non - zero neutrino masses ( see i.e. @xcite for excellent older reviews on this matter ) .
fermion masses in general are one of the major problems of the standard model . observation or non - observation of neutrino masses
could introduce an useful new perspective on the subject . if massless they would be the only fermions with this property . a property which is not dictated by a fundamental underlying principle ( at least one which is known presently ) , such as gauge invariance in the case of the photon .
if massive the question is why are their masses so much smaller than those of their charged partners .
although theory alone can not predict neutrino masses , it is certainly true that they are strongly suggested by present theoretical models of elementary particles and most extensions of the standard model definitively require neutrinos to be massive .
they therefore constitute a powerful probe of new physics at a scale larger than the electroweak scale . some hints at accelerator experiments as the the observed indications of spectral distortion and deficit of solar neutrinos and the ratio of atmospheric @xmath1 neutrinos and their zenith distribution are naturally accounted by the oscillations of a massive neutrino .
recent claims of the high - statistics high precision superkamiokande experiment are unambiguous and left little room for the scepticism .
neutrinos are basic ingredients of astrophysics and cosmology . there may be a hot dark matter component ( hdm ) to the universe : simulations of structure formation fit the observations only when one has some 20 % of hdm , the best fit being two neutrinos with a total mass of 4.7 ev . if so , neutrinos would be , at least in quantity , one of the most important ingredients in the universe .
there seems to be however some kind of conflict within cosmology itself : observations of distant objects favor a large cosmological constant instead of hdm
. regardless of mass and oscillations , astrophysical interest in the neutrino and their properties arises from the fact that it is copiously produced in high temperature and/or high density environment and it often dominates the physics of those astrophysical objects .
the interactions of the neutrino with matter is so weak that it passes freely through any ordinary matter existing in the universe .
this makes neutrinos to be a very efficient carrier of energy drain from optically thick objects .
at the same time they give a good probe for the interior of such dense objects .
for example , the solar neutrino flux is , together with heliosysmology , one of the two known probes of the solar core .
a similar statement applies to the type - ii supernovas : the most interesting questions around supernovas , the explosion dynamics itself with the shock revival , and , the synthesis of the heaviest elements by r - process , could be positively affected by changes in the neutrino flux , e.g. by msw active or sterile conversions .
finally , ultra high energy neutrinos are called to be useful probes of diverse distant astrophysical objects .
active galactic nuclei ( agn ) should be copious emitters of @xmath2 s , providing both detectable point sources and an observable diffuse background which is larger in fact than the atmospheric neutrino background in the very high energy range .
the standard model ( sm ) contains three left - handed neutrinos .
the three neutrinos are represented by two - component weyl spinors @xmath3 each describing a left - handed fermion .
they are upper components of weak isodoublets @xmath4 , they have @xmath5 and an unit of the global ith lepton number . these standard model neutrinos are strictly massless .
the only lorenz scalar made out of them is the majorana mass , of the form @xmath6 ; it has the quantum number of a weak isotriplet , with @xmath7 as well as two units of total lepton number . thus to generate a renormalizable majorana mass term at the tree level one needs a higgs isotriplet with two units of lepton number . since in the strict standard model higgs
is a weak isodoublet higgs , there are no tree - level neutrino masses . if we want to consider quantum corrections we should consider effective terms where a weak isotriplet is made out of two isodoublets and which are not invariant under lepton number symmetry .
the conclusion is that the standard model neutrinos are kept massless by global chiral lepton number symmetry ( and more general properties as renormalizability of the theory ) .
however this is a rather formal conclusion , there is no any other independent , compelling theoretical argument in favor of such symmetry . or , with other words
, there is not any reason why we would like to keep it intact .
independent from mass oddities in any other respect neutrinos are very well behaved particles within the sm framework .
some properties are unambiguously known about neutrinos .
the lep line - shape measurement imply that are only three ordinary light neutrinos .
big bang nucleosynthesis ( bbn ) constrains the parameters of possible sterile neutrinos ( which interact and are produced only by mixing ) . _
all the existing _ data on the the weak interaction processes in which neutrinos take part are perfectly described by the standard model charged - current ( cc ) and neutral - current ( nc ) lagrangians : @xmath8 the cc and nc interaction lagrangians conserve the total three lepton numbers @xmath9 while cc interactions determine the notion of flavor neutrinos @xmath10 .
there are no indications in favor of violation of lepton numbers in weak processes . from the existing experiments very strong bounds on branching ratios of rare , lepton number violating ,
processes are obtained , for example : @xmath11 and @xmath12 ( @xcite , 90% cl ) .
any satisfactory model that generates neutrino masses must contain a natural mechanism that explains their small value , relative to that of their charged partners . given the latest experimental indications it would also be desirable that includes justification for light sterile neutrinos and near maximal mixing . to generate neutrino masses without new fermions
, we must break lepton number by adding to the sm higgs fields carrying lepton numbers , one can arrange then to break lepton number explicitly or spontaneously through their interactions .
possibly , the most familiar approach to give neutrino masses is , however , to introduce for each one an electroweak neutral singlet .
this happens naturally in left - right symmetric models where the origin of sm parity violation is ascribed to the spontaneous breaking of a b - l symmetry . in the so(10 ) gut the majorana neutral particle n enters in a natural way in order to complete the matter multiplet , the neutral n is a @xmath13 singlet . according to the scale where they have relevant effects , unification ( i.e. the aforementioned so(10 ) gut ) and weak - scale approaches ( i.e. radiative models ) can be distinguished .
phenomenologically , mass terms can be viewed as describing transitions between right ( r ) and left ( l)-handed states . for a set of four fields : @xmath14 , the most general mass part of the free - field lagrangian can be written as : @xmath15 in terms of the new majorana fields : @xmath16 , @xmath17 , the lagrangian becomes : @xmath18 where the neutrino mass matrix is evident . diagonalizing
this matrix one finds that the physical particle content is given by two majorana mass eigenstates : the inclusion of the majorana mass splits the four degenerate states of the dirac field into two non - degenerate majorana pairs .
if we assume that the states @xmath19 are respectively active and sterile , the majorana masses @xmath20 and @xmath21 transform as weak triplets and singlets respectively .
@xmath22 is an standard dirac mass term .
the neutrino mass matrix can easily be generalized to three or more families , in which case the masses become matrices themselves .
the complete flavor mixing comes from two different parts , the diagonalization of the charged lepton yukawa couplings and that of the neutrino masses . in most simple extensions of the standard model ,
this ckm - like leptonic mixing is totally arbitrary with parameters only to be determined by experiment .
their prediction , as for for the quark hierarchies and mixing , needs further theoretical assumptions ( i.e. ref.@xcite predicting @xmath23 maximal mixing ) .
the case @xmath24 in eq.([e2003 ] ) corresponds to a purely dirac mass term . in this case @xmath19
are degenerate with mass @xmath22 and a four component dirac field can be recovered as @xmath25 .
the dirac mass term allows a conserved lepton number @xmath26 . for an ordinary dirac neutrino
the @xmath27 is active and @xmath28 is weak sterile , an su(2 ) singlet , the mass term describes then a @xmath29 transition and is generated from su(2 ) breaking with a yukawa coupling : @xmath30 one has @xmath31 where v is the vacuum expectation value of the higgs doublet and @xmath32 is the corresponding yukawa coupling .
a neutrino dirac mass is qualitatively just like any other fermion masses , but that leads to the question of why it is so small in comparison with the rest of fermion masses : one would require @xmath33 in order to have @xmath34 ev . or in other words : @xmath35 while for the hadronic sector we have @xmath36 . a pure majorana mass transition term , @xmath20 or @xmath21 terms in lagrangian ( [ e2003 ] ) , describes in fact a particle - antiparticle transition violating lepton number by two units ( @xmath37 ) .
it can be viewed as the creation or annihilation of two neutrinos leading therefore to neutrinoless double beta decay . for n
a gauge singlet , a renormalizable mass term of the type @xmath38 is allowed by sm su(3)xsu(2)xu(1 ) symmetry .
however , it would not be consistent in general with unified symmetries , i.e. with a full so(10 ) symmetry .
in such theory the most straightforward possibility for generating a term like it in the full theory would be to include a * 126 * higgs and a yukawa coupling .
alternatively , one can imagine more complicated interactions containing products of several simpler higgs fields . whatever the concrete model , the main point to have into account is that , it is strongly suggested that @xmath21 is associated with breaking of some unified symmetry , the expected scale for it should be in a large range covering from @xmath39 tev ( left - right models ) to gut scales @xmath40 gev ( from couplings unification ) .
if @xmath27 is active then @xmath41=1 and @xmath20 must be generated by either an elementary higgs triplet or by an effective operator involving two higgs doublets arranged to transform as a triplet .
for an elementary triplet @xmath42 , where @xmath43 is a yukawa coupling and @xmath44 is the triplet vev .
the simplest implementation ( the old gelmini - roncadelli model ) is excluded by the lep data on the z width : the corresponding majoron couples to the z boson increasing significantly its width .
variant models involving explicit lepton number violation or in which the majoron is mainly a weak singlet ( invisible majoron models ) could still be possible .
for an effective operator originated mass , one expects @xmath45 where m is the scale of the new physics which generates the operator .
one can see this easily in the see - saw scheme : the n s communicate with the familiar fermions through the yukawa interactions ( eq.([e2005 ] ) .
if n were massless the effect of eq.([e2005 ] ) would be to generate masses of the same order as ordinary quark and lepton masses .
if n are massive enough yukawa and mass terms could be integrated out to generate an effective term of the type ( l , lepton doublet ) : @xmath46 in the seesaw limit in eq.([e2003 ] ) , taking @xmath47 , the two majorana neutrinos acquire respectively masses @xmath48 , @xmath49 .
there is one heavy neutrino and one neutrino much lighter than the typical dirac fermion mass .
this is a natural way of generating two well separated mass scales .
now we come again to the majorana mass @xmath21 for the sterile neutral n : @xmath21 can vary anywhere from the tev scale to the planck scale among the large quantity of proposed concrete seesaw and related models .
the tev scale models are motivated , i.e. , by left - right symmetric models . with typical @xmath22 s ,
one expects masses of order @xmath50 ev , 10 kev , and 1 mev for the @xmath10 respectively violating cosmological bounds unless heavy neutrinos decay rapidly and invisibly .
gut motivated intermediate scales ( @xmath51 gev ) yield typical masses in the range relevant to hot dark matter , and solar and atmospheric neutrino oscillations . for @xmath52 gev ( some superstring models , gut with multiple breaking stages ) one can obtain light neutrino masses of the order @xmath53 ev , @xmath54 ev , 10 ev ) . such range of masses would allow the interpretation of the solar and atmospheric deficits as , respectively , @xmath55 , @xmath56 oscillations .
@xmath57 could be considered as a dark matter candidate . for @xmath58 ( grand unified seesaw with large higgs representations )
one typically finds smaller masses around @xmath59 , @xmath60 , @xmath61 ) ev somehow more difficult to fit into the present known experimental facts .
models have been proposed where small tree level neutrino masses are obtained without making use of large scales . the model proposed by @xcite ( inspired by previous superstring models ) offers an example of the possibility of having neutrino mass matrices more general than that given by eq.([e2003 ] ) .
the incorporation of additional iso - singlet neutral fermions @xmath62 leads to a matrix of the type : @xmath63 the smallness of neutrino masses is explained directly from the , otherwise left unexplained , smallness of the parameter @xmath64 in such a model .
moreover , there would be neutrino mixing even if the light neutrino remains strictly massless ( @xmath65 ) . the anomalies observed in the solar neutrino flux , atmospheric flux and low energy accelerator experiments
can not all be explained consistently without introducing a light , then neccesarily sterile , neutrino . if all the majorana masses are small , active neutrinos can oscillate into the sterile right handed fields .
light sterile neutrinos can appear in particular see - saw mechanisms if additional assumptions are considered ( singular see - saw models ) with some unavoidable fine tuning .
the alternative to such fine tuning would be seesaw - like suppression for sterile neutrinos involving new unknown interactions , i.e. family symmetries , resulting in substantial additions to the sm , ( i.e. some sophisticated superstring - inspired models , ref.@xcite ) .
finally , weak scale , radiative generated mass models where the neutrino masses are zero at tree level constitute a very different class of models : they explain in principle the smallness of @xmath66 for both active and sterile neutrinos .
different mass scales are generated naturally by different number of loops involved in generating each of them .
the actual implementation generally requires however the ad - hoc introduction of new higgs particles with nonstandard electroweak quantum numbers and lepton number - violating couplings @xcite .
the magnetic dipole moment is another probe of possible new interactions .
majorana neutrinos have identically zero magnetic and electric dipole moments .
flavor transition magnetic moments are allowed however in general for both dirac and majorana neutrinos .
limits obtained from laboratory experiments are of the order of a few @xmath67 and those from stellar physics or cosmology are @xmath68 . in the sm electroweak theory , extended to allow for dirac neutrino masses , the neutrino magnetic dipole moment is nonzero and given , as ( @xcite and references therein ) : @xmath69 where @xmath70 is the bohr magneton .
the proportionality of @xmath71 to the neutrino mass is due to the absence of any interaction of @xmath28 other than its yukawa coupling which generates its mass . in left - right symmetric theories
@xmath71 is proportional to the charged lepton mass : a value of @xmath72 can be reached still too small to have practical astrophysical consequences .
magnetic moment interactions arise in any renormalizable gauge theory only as finite radiative corrections .
the diagrams which generate a magnetic moment will also contribute to the neutrino mass once the external photon line is removed . in the absence of additional symmetries
a large magnetic moment is incompatible with a small neutrino mass .
the way out suggested by voloshin consists in defining a su(2)@xmath73 symmetry acting on the space @xmath74 , magnetic moment terms are singlets under this symmetry . in the limit of exact su(2)@xmath73 the neutrino mass is forbidden but @xmath71 is allowed .
diverse concrete models have been proposed where such symmetry is embedded into an extension of the sm ( left - right symmetries , susy with horizontal gauge symmetries @xcite ) .
no indications in favor of non - zero neutrino masses were found in direct kinematical searches : 1 . from the measurement of the high energy part of the @xmath75 spectrum in the tritium decay : the troitsk and mainz experiments obtain respectively @xmath76 ev @xcite and mainz @xmath77 ev @xcite . both measurements are plagued by interpretation ambiguities : apparition of negative mass squared and bumps at the end of the spectrum .
limits for the muon neutrino mass have been derived using the decay channel @xmath78 at intermediate energy accelerators ( psi , lanl ) .
the present limits are @xmath79 kev @xcite .
a tau neutrino mass of less than 30 mev is well established and confirmed by several experiments : limits of 28 , 30 and 31 mev have also been obtained by the opal , cleo and argus experiments respectively @xcite .
the best upper limit for the @xmath80 neutrino mass has been derived using the decay mode @xmath81 by the aleph collaboration @xcite : @xmath82 18 mev ( 95% cl ) .
many experiments on the search for neutrinoless double - beta decay [ @xmath83 , @xmath84 have been done .
this process is possible only if neutrinos are massive and majorana particles .
the matrix element of the process is proportional to the effective majorana mass @xmath85 .
uncertainties in the precise value of upper limits are important since they depend on theoretical calculations of nuclear matrix elements . from the non - observation of @xmath86 the heidelberg - moscow ge experiment
@xcite draws the limit @xmath87 ev ( 90% cl ) . in the next years
it is expected an increase in sensitivity allowing limits down the @xmath88 ev level .
many short - baseline ( sbl ) neutrino oscillation experiments with reactor and accelerator neutrinos did not find any evidence of neutrino oscillations .
for example experiments looking for @xmath89 or @xmath90 dissaperance @xcite or oscillations @xmath91 @xcite .
the first reactor long - baseline neutrino oscillation experiment chooz found no evidence for neutrino oscillations in the @xmath92 disappearance mode @xcite .
their results imply an exclusion region in the plane of the two - generation mixing parameters ( with normal or sterile neutrinos ) given approximately by @xmath93 for maximum mixing and @xmath94 for large @xmath95 , as shown in fig.([fchooz ] ) .
chooz results are important for the atmospheric deficit problem : as it is seen in fig.([fchooz ] ) they are incompatible with an @xmath96 oscillation hypothesis for the solution of the atmospheric problem .
los alamos lsnd experiment has reported indications of possible @xmath91 oscillations @xcite .
they search for @xmath97 s in excess of the number expected from conventional sources at a liquid scintillator detector located 30 m from a proton beam dump at lampf .
a @xmath98 signal has been detected via the reaction @xmath99 with @xmath100 energy between 36 and 60 mev , followed by a @xmath101 from @xmath102 ( 2.2 mev ) .
a total @xmath97 excess of @xmath103 events has been obtained .
if this excess is attributed to neutrino oscillations of the type @xmath91 , it corresponds to an oscillation probability of @xmath104 .
the results of a similar search for @xmath105 oscillations where the @xmath106 are detected via the cc reaction @xmath107 provide a value for the corresponding oscillation probability of @xmath108 .
the lsnd result has not been confirmed by the karmen experiment . the karmen experiment ( rutherford laboratories ) , following a similar experimental setup as lsnd , searches for @xmath109 produced by @xmath110 oscillations at a mean distance of 17.6 m. the time structure of the neutrino beam is important for the identification of the neutrino induced reactions and for the suppression of the cosmic ray background .
systematic time anomalies not completely understood has been reported .
the 1990 - 1995 and the latest 1997 - 1998 karmen data showed inconclusive results .
they found no events , with an expected background of @xmath111 events , for @xmath91 oscillations , however the positive lsnd result in this channel could not be excluded in total either @xcite . at the end of 1999
, the karmen sensitivity is expected to be able to exclude the whole parameter region of evidence suggested by lsnd if no oscillation signal were found ( fig.[flsnd ] ) .
the first phase of a third pion beam dump experiment designed to set the lsnd - karmen controversy has been approved to run at fermilab .
phase i of boone ( miniboone ) expects a 10 @xmath112 signal ( @xmath113 events ) and thus will make a decisive statement either proving or ruling it out .
plans are to run early 2001 .
additionally , there is a letter of intent of a similar experiment to be carried out at the cern ps @xcite .
indications in the favor of neutrino oscillations were found in all solar neutrino experiments : the homestake cl radiochemical experiment with sensitivity down to the lower energy parts of the @xmath114b neutrino spectrum and to the higher @xmath115 be line .
the two radiochemical @xmath116ga experiments , sage and gallex , which are sensitive to the low energy pp neutrinos and above and the water cerenkov experiments kamiokande and super - kamiokande ( sk ) which can observe only the highest energy @xmath114 b neutrinos .
water cerenkov experiments in addition demonstrate directly that the neutrinos come from the sun showing that recoil electrons are scattered in the direction along the sun - earth axis .
two important points to remark are : a ) the prediction of the existence of a global neutrino deficit is hard to modify due to the constraint of the solar luminosity on pp neutrinos detected at sage - gallex .
b ) the different experiments are sensitive to neutrinos with different energy ranges and combined yield spectroscopic information on the neutrino flux .
intermediate energy neutrinos arise from intermediate steps of the thermonuclear solar cycle
. it may not be impossible to reduce the flux from the last step ( @xmath114b ) , for example by reducing temperature of the center of the sun , but it seems extremely hard to reduce neutrinos from @xmath117be to a large extent , while keeping a reduction of @xmath118b neutrinos production to a modest amount . if minimal standard electroweak theory is correct , the shape of the @xmath118 b neutrino energy spectrum is independent of all solar influences to very high accuracy .
unless the experiments are seriously in error , there must be some problems with either our understanding of the sun or neutrinos . clearly
, the ssm can not account for the data ( see fig.[ssm1 ] ) and possible highly nonstandard solar models are strongly constrained by heliosysmology studies ( see fig.([ssm2 ] ) ) .
there are at least two reasonable particle physics explanations that could account for the suppression of intermediate energy neutrinos .
the first one , neutrino oscillations in vacuum , requires a large mixing angle and a seemingly unnatural fine tuning of neutrino oscillation length with the sun - earth distance for intermediate energy neutrinos .
the second possibility , level - crossing effect oscillations in presence of solar matter and/or magnetic fields of regular and/or chaotic nature ( msw , rsfp ) , requires no fine tuning either for mixing parameter or neutrino mass difference to cause a selective large reduction of the neutrino flux .
this mechanism explains naturally the suppression of intermediate energy neutrinos , leaving the low energy pp neutrino flux intact and high energy @xmath118b neutrinos only loosely suppressed .
concrete range of parameters obtained including the latest sk data will be showed in the next section .
the high precision and high statistics super - kamiokande ( sk ) experiment initiated operation in april 1996 .
a few words about the detector itself .
sk is a 50-kiloton water cerenkov detector located near the old kamiokande detector under a mean overburden of 2700 meter - water - equivalent .
the effective fiducial volume is @xmath119 kt . it is a well understood , well calibrated detector .
the accuracy of the absolute energy scale is estimated to be @xmath120 based on several independent calibration sources : cosmic ray through - going and stopping muons , muon decay electrons , the invariant mass of @xmath121 s produced by neutrino interactions , radioactive source calibration , and , as a novelty in neutrino experiments , a 5 - 16 mev electron linac .
in addition to the ability of recording higher statistics in less time , due to the much larger dimensions of the detector , sk can contain multi - gev muon events making possible for the first time a measurement of the spectrum of @xmath64-like events up to @xmath122 gev / c .
the results from the first 504 days of data from sk ( results presented recently at the neutrino98 conference @xcite ) , combined with data from earlier experiments provide important constraints on the msw and vacuum oscillation solutions for the solar neutrino problem ( snp ) . in the next paragraphs
we will present a summary of the results presented in that conference together with initial analysis in the framework of neutrino oscillations .
the most robust results of the solar neutrino experiments so far are the total observed rates .
the most recent data on rates are summarized in table ( [ t1 ] ) .
total rates alone indicate that the @xmath106 energy spectrum from the sun is distorted .
the ssm flux predictions are inconsistent with the observed rates in solar neutrino experiments at approximately the 20@xmath112 level .
furtherly , there is no linear combination of neutrino fluxes that can fit the available data at the 3@xmath112 level ( fig.([ssm1 ] ) ) .
.neutrino event rates measured by solar neutrino experiments , and corresponding predictions from the ssm ( see ref.@xcite and references therein , we take the int normalization for the ssm data ( @xmath123 errors ) . [ cols="^,^,^,^ " , ] from a two - flavor analysis of the total event rates in the clar , sage , gallex and sk experiments the best @xmath124 fit considering active neutrino oscillations is obtained for @xmath125 ev@xmath126 ( the so called small mixing angle solution , sma ) .
other local @xmath124 minima exist .
the large mixing angle solution ( lma ) occurs at @xmath127 ev@xmath128 , the low solution ( lower probability , low mass ) , at @xmath129 ev@xmath130 .
the vacuum oscillation solution occurs at @xmath131 ev@xmath132 . at this extremely low value for the mass difference
the msw effect is inopperant . for oscillations involving sterile neutrinos (
the matter effective potential is modified in this case ) the lma and low solutions are not allowed and only the ( only slightly modified ) sma solution together with the vacuum solution are still possible . more sophisticated analysis including more than two neutrino species are not available but they would not change so much the previous picture while introducing a much larger technical difficulty .
analysis which consider neutrino oscillations in presence of magnetic fields , the rsfp effect , have also been presented . typically , they yield solutions with @xmath133 ev@xmath134 for both small and large mixing angles .
rsfp solutions are much more ambiguous than pure msw solutions because of neccesity to introduce additional free parameters in order to model the largely unknown intensity and profile of solar magnetic fields .
the recognition of the random nature of solar convective fields and recent theoretical developments in the treatment of schroedinger random equations have partially aliviated this situation , allowing the obtention of snp solutions without the neccesity of a detailed model description ( see recent analysis in @xcite ) .
in addition , random rsfp models predict the production of a sizeable quantity of electron antineutrinos .
presently , antineutrino searches with negative results in kamiokande and sk are welcome because restrict significantly the , uncomfortably large , parameter space of rsfp models . in the future such antineutrinos could be identified both in sk or in sno setting the majorana nature of the neutrino ( @xcite ) .
if msw oscillations are effective , for a certain range of neutrino parameters the observed event rate will depend upon the zenith angle of the sun ( earth matter regeneration effect ) .
after 504 days of data still , due to lack of statistics , the most robust estimator of zenith angle dependence by now is the day - night ( or up - down ) asymmetry , a. the experimental estimation is : @xmath135 the difference is small and not statistically significant but it is in the direction that would be expected from regeneration at earth ( the sun is apparently neutrino brighter at night ) .
taken alone the small value observed for a excludes a large part of the parameter region that is allowed if only the total rates would be considered ( see fig.([solar1 ] ) ) . from the independence from astrophysical causes ,
the shape of the neutrino spectrum determines the shape of the recoil electron energy spectrum produced by neutrino - electron scattering in the detector .
all the neutrino oscillation solutions ( sma , lma , low and vacuum ) provide acceptable , although indeed not excellent fits to the recoil energy spectrum .
the simplest test is to investigate whether the ratio , r , of the observed to the standard energy spectrum is a constant .
the fit of the ratio r to a linear function yields slope values which are incompatible at 99% cl with the hypothesis of no distortion ( see figs.([solar1]-[solar2 ] ) ) . in the case
where all data , the total rates , the zenith - angle dependence and the recoil energy spectrum , is combined the best - fit solution is almost identical to what is obtained for the rates only case . for other solutions , only the sma and vacuum solution survives ( at the 99% cl ) .
the lma and the low solutions are , albeit marginally , ruled out @xcite .
a small but significant discrepancy appears when comparing the predictions from the global best fits for the energy spectrum at high energies ( @xmath136 mev ) with the sk results presented in neutrino98 @xcite . from this discrepancy
it has been speculated that hep neutrinos may affect the solar neutrino energy spectrum .
presently low energy nuclear physics calculations of the rate of the hep reaction are highly uncertain ( a factor of six is allowed ) .
coincidence between expected and measured ratios is improved when the hep flux is allowed to vary as a free parameter ( see fig.([solar6 ] ) ) .
atmospheric neutrinos are the decay products of hadronic showers produced by cosmic ray interactions in the atmosphere .
the experimental ratio r @xmath137 where @xmath138 denotes the ratio of the numbers of @xmath64-like to @xmath139-like neutrino interactions observed in the data or predicted by the simulation has been measured as an estimator of the atmospheric neutrino flavor ratio @xmath140 the individual absolute neutrino flux calculation has a @xmath141 uncertainty .
the flux ratio has been calculated however to an accuracy of better then @xmath142 in the range @xmath143 gev .
the calculated flux ratio has a value of about 2 for energies @xmath144 1 gev and increases with increasing neutrino energy . for neutrino energies higher than a few gev
, the fluxes of upward and downward going neutrinos are expected to be nearly equal ; geomagnetic field effects at these energies are expected to be small because of the relative large geomagnetic rigidity of the primary cosmic rays that produce these neutrinos .
anomalous , statistically significant , low values of the ratio r has been obtained previously in the water cerenkov detectors kamiokande and imb-3 and the soudan-2 for `` sub - gev '' events ( e@xmath145 gev ) .
the nusex and frejus experiments have reported results consistent with no deviation from unity with smaller data samples .
kamiokande experiment observed a value of r smaller than unity in the multi - gev ( e@xmath1461 gev ) energy region as well as a dependence of this ratio on the zenith angle .
imb-3 , with a smaller data sample , reported inconclusive results in a similar energy range but they are not in contradiction with kamiokande observations .
superkamiokande ( sk ) recent results are completely consistent with previous results but they are more accurate than before .
specially significant improvements in accuracy have been obtained in measuring the zenith angular dependence of the neutrino events .
they see that the flux of muon neutrinos going up is smaller than that of downgoing neutrinos : in the sub - gev range ( @xmath147 gev ) , from an exposure of @xmath119 kiloton - years of the sk detector the measured ratio r is : @xmath148 it is not possible to determine from data , whether the observed deviation of r is due to an electron excess of a muon deficit .
the distribution of r with momentum in the sub - gev range is consistent with a flat distribution within the statistical error as it happens with zenith angle distributions ( see right plots in fig.([atm8 ] ) ) . in the multi - gev range , it has been obtained ( for a similar exposure ) a ratio r which is slightly higher than at lower energies @xmath149 for e - like events , the data is apparently consistent with mc . for @xmath64-like events
there is a clear discrepancy between measurement and simulation .
a strong distortion in the shape of the @xmath64-like event zenith angle distribution was observed ( plots ( [ atm4]-[atm8 ] ) ) .
the angular correlation between the neutrino direction and the produced charged lepton direction is much better at higher energies ( @xmath150 ) : the zenith angle distribution of leptons reflects rather accurately that of the neutrinos in this case .
the ratio of the number of upward to downward @xmath64-like events was found to be @xmath151 while the expected value is practically one : @xmath152 the validity of the results has been tested by measuring the azimuth angle distribution of the incoming neutrinos , which is insensitive to a possible influence from neutrino oscillations .
this shape agreed with mc predictions which were nearly flat .
the most obvious solution to the observed discrepancy is @xmath153 flavor neutrino oscillations .
this fits well to the angular distribution , since there is a large difference in the neutrino path - length between upward - going ( @xmath154 km ) and downward - going ( @xmath155 km ) , a zenith angle dependence of r can be interpreted as evidence for neutrino oscillations .
oscillation into sterile neutrinos , @xmath156 , is also a good explanation consistent with data .
@xmath157 oscillations does not fit however so well , they would also conflict laboratory measurements ( chooz , figs.([fchooz]-[atm6 ] ) ) . apart from neutrino
oscillation , no other consistent explanation has been proposed .
evidence for oscillations equals evidence for non - zero neutrino mass within the standard neutrino theory .
the allowed neutrino oscillation parameter regions obtained by kamiokande and sk from different analysis are shown in fig.([atm6 ] ) .
the best fit is obtained for squared mass differences in the range @xmath158 ev@xmath134 , and very large mixing .
unless there is no fine tuning , this suggests a neutrino mass of the order of 0.1 ev .
such a mass implies the neutrino energy density in the universe to be 0.001 of the critical density which is too small to have cosmological consequences .
this is of course a very rough argument : specific models , however , may allow larger neutrino masses quite naturally . from the individual analysis of the data available from neutrino experiments
, it follows that there exist three different scales of neutrino mass squared differences and two different ranges of small and maximal mixing angles , namely : @xmath159 fortunely for the sake of simplicity the neutrino mass scale relevant for hdm is roughly similar to the lsnd one .
the introduction of the former would not change any further conclusion .
but for the same reason , the definitive refutation of lsnd results by karmen or future experiments does not help completely in simplifying the task of finding a consistent framework for all the neutrino phenomenology .
any combination of experimental data which involves only of the two mass scales can be fitted within a three family scenario , but solving simultaneously the solar and atmospheric problems requires generally some unwelcome fine tuning of parameters at the @xmath61 level .
the detailed analysis of ref.@xcite obtains for example that solutions with 3 neutrino families which are compatible with the results from sbl inclusive experiments , lsnd and solar neutrino experiments are possible .
the problem arises when one add the results from chooz , which rule out large atmospheric @xmath160 transitions and zenith dependence from sk atmospheric data one comes to the neccesity of consideration of schemes with four massive neutrinos including a light sterile neutrino . among the numerous possibilities ,
complete mass hierarchy of four neutrinos is not favored by existing data @xcite nor four - neutrino mass spectra with one neutrino mass separated from the group of the three close masses by the lsnd gap ( @xmath39 1 ) ev .
one is left with two possible options where two double - folded groups of close masses are separated by a @xmath161 ev gap : @xmath162 & ( b)&\ \underbrace{\overbrace{\nu_e\to \nu_\tau : \ m_1 < m_2}^{sun } < < \overbrace{\nu_\mu\to\nu_s:\ m_3 < m_4}^{atm}}_{lsnd\sim 1 ev}. \end{aligned}\ ] ] the two models would be distinguishable from the detailed analysis of future solar and atmospheric experiments . for example they may be tested combining future precise
recoil electron spectrum in @xmath163 measured in sk and sno @xcite with the sno spectrum measured in cc absorption .
the sno experiment ( a 1000 t heavy water under - mine detector ) will measure the rates of the charged ( cc ) and neutral ( nc ) current reactions induced by solar neutrinos in deuterium : @xmath164 including the determination of the electron recoil energy in the cc reaction .
only the more energetic @xmath118b solar neutrinos are expected to be detected since the expected sno threshold for cc events is an electron kinetic energy of about 5 mev and the physical threshold for nc dissociation is the binding energy of the deuteron , @xmath165 mev .
if the ( b ) model it is true one expects @xmath166 while in the ( a ) model the ratio would be @xmath167 . the schemes ( a ) and ( b ) give different predictions for the neutrino mass measured in tritium @xmath75-decay and for the effective majorana mass observed in neutrinoless double @xmath75 decay .
respectively we have @xmath168 ( a ) or @xmath169 ( b ) .
thus , if scheme ( a ) is realized in nature this kind of experiments can see the effect of the lsnd neutrino mass . from the classical lep requirement @xmath170 @xcite , it is clear that the fourth neutrino should be a @xmath171 singlet in order to ensure that does not affect the invisible z decay width .
the presence of additional weakly interacting light particles , such as a light sterile @xmath172 , is constrained by bbn since it would enter into equilibrium with the active neutrinos via neutrino oscillations .
the limit @xmath173 ev@xmath134 should be fulfilled in principle .
however systematical uncertainties in the derivation of the bbn bound make any bound too unreliable to be taken at face value and can eventually be avoided @xcite .
taking the most restrictive options ( giving @xmath174 ) only the ( a ) scheme is allowed , one where the sterile neutrino is mainly mixed with the electron neutrino . in the lest restrictive case ( @xmath175 ) both type of models would be allowed .
the theoretical challenges that the present phenomenological situation offers are two at least : to understand origin and , very particularly , the lightness of the sterile neutrino ( apparently requiring a radiatively generated mass ) and to account for the maximal neutrino mixing indicated by the atmospheric data which is at odd from which one could expect from considerations of the mixing in the quark sector .
actually , the existence of light sterile neutrinos could even be beneficial in diverse astrophysical and cosmological scenarios ( supernova nucleosynthesis , hot dark matter , lepton and baryon asymmetries for example ) . in the last years different indications in favor of nonzero neutrino masses and mixing angles have been found .
these evidences include four solar experiments clearly demonstrating an anomaly compared to the predictions of the standard solar model ( ssm ) and a number of other atmospheric experiments , including a high statistics , well calibrated one , demonstrating a quite different anomaly at the earth scale .
one could argue that if we are already beyond the stage of having only circumstantial evidence for new physics , we are still however a long way from having conclusive proof of new physics .
evidence for new physics does not mean the same as evidence for neutrino oscillations but there exists a significant case for neutrino oscillations and hence neutrino masses and mixing as one , indeed the most serious candidate , explanation of the data .
one of the possible alternatives is that one or more of the experiments will turn out to be wrong .
this is possible and even probable , but it is little probable that with all the evidence accumulated by now all the experiments turn out to be simultaneously wrong .
many neutrino experiments are taking data , are going to start or are under preparation : solar neutrino experiments ( sno and borexino are of major interest , also heron , hellaz , icarus , gno and others ) ; lbl reactor ( chooz , palo verde , kamland ) and accelerator experiments ( k2k , minos , icarus and others ) ; sbl experiments ( lsnd , karmen , boone and many others ) . the important problem for any next generation experiment is to find specific and unambiguous experimental probe that the anomalies which has been found are indeed signals of neutrino oscillations and to distinguish among the different neutrino oscillation possibilities ( this is specially important in the solar case ) . among these probes
, we could include : * perhaps the most direct test of sm deviation : to measure the ratio of the flux of @xmath106 s ( via a cc interaction ) to the flux of neutrinos of all types ( @xmath176 , determined by nc interactions ) .
this measurement will be done hopefully by the sno experiment in the near future .
see fig.([fut3 ] ) .
* statistically significant demonstration of an energy - dependent modification of the shape of the electron neutrino spectrum arriving at earth .
besides observing distortion in the shape of @xmath118b neutrinos , it will be very important to make direct measurements of the @xmath117be ( borexino experiment ) and pp ( heron , hellaz ) neutrinos .
* improved observation of a zenith angle effect in atmospheric experiments or their equivalent , a day - night effect in solar experiments . * and least , but by no means the least , independent confirmation by one or more accelerator experiments .
there is a high probability that in the near future we should know much more than now about the fundamental properties of neutrinos and their masses , mixing and their own nature whether dirac or majorana .
e. torrente - lujan .
hep - ph/9807361 .
( to be published in phys . rev .d ) .
e. torrente - lujan .
hep - ph/9807371 , ( to be published in phys .
d. ) e. torrente - lujan .
hep - ph/9807426 .
( to be published in phys . lett . ) sno collaboration , g.t .
ewan et al . , sudbury neutrino observatory proposal .
report n. sno-87 - 12 , 1987 ( unpublished ) ; scientific and technical description of the mark ii sno detector , edited by e.w .
beier and d. sinclair , report n. sno-89 - 15 , 1989 ( unpublished ) .
mc.donald , proc . of the 9th lake louise winter institute , edited by a. astbury et al .
( world scientific , singapore , 1994)p.1 . | we present a short review of the present status of the problem of neutrino masses and mixing including a survey of theoretical motivations and models , experimental searches and implications of recently appeared solar and atmospheric neutrino data , which strongly indicate nonzero neutrino masses .
such data apparently requires the existence of light sterile neutrino in addition to the three active ones and two - generation nearly maximal mixing .
-10 mm -10 mm -10 mm # 1eq .
( [ # 1 ] ) # 1eq .
( [ # 1 ] ) # 1#2#3 _ ann . phys .
( ny ) _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ ann
. rev .
nucl . part .
sci . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ comm .
nucl . part .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astrophys .
j. _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astropart .
phys .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astrophys .
space rev .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astrophys .
space sci .
_ * # 1 * ( 19#2 ) # 3 #1#2#3 _ astron . &
astrophys .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astrophys .
j. lett .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astropart .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ astrophys .
space sci .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ journal europhys .
lett .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ ibid . _
* # 1 * ( 19#2 ) # 3 # 1#2#3 _ nature _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ nucl .
phys .
b ( proc . suppl . ) _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ nucl .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ phys .
lett . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ phys .
rev . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ phys .
rep . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ phys .
rev .
lett . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ particle world _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ prog .
theor .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ j. prog .
part .
nucl .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ comp .
phys .
commun . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ rep .
on prog . in phys .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ prog
. theor .
phys .
suppl . _
* # 1 * ( 19#2 ) # 3 # 1#2#3 _ rev .
mod .
phys . _
* # 1 * ( 19#2 ) # 3 # 1#2#3 _ zeit .
fur physik _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ fortschr .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ z. physik _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ science _ * # 1 * ( 19#2 ) # 3 .c.#1#2#3 _
nuovo cim . _ * # 1 * ( 19#2 ) # 3 .n.c.#1#2#3 _ riv .
del nuovo cim .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ sov .
j. nucl .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ yad .
fiz .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ z. eksp .
teor .
fiz .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ z. eksp .
teor .
fiz .
pisma .
red .
_ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ jetp _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ mod .
phys .
lett . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ usp .
fiz .
naut . _
* # 1 * ( 19#2 ) # 3 # 1#2#3 _ sov .
phys .- usp._*#1 * ( 19#2 ) # 3 # 1#2#3 _ prog .
part .
nucl .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ comm .
. part .
phys . _ * # 1 * ( 19#2 ) # 3 # 1#2#3 _ int .
j. mod .
phys . _ * # 1 * ( 19#2 ) # 3 -2 cm 16.cm # 1#1 1@xmath0 # 10= -.025em0 - 0 .05em0 - 0 -.025em.0433em0 ftuv / ific-9910 + hep - ph/9902339 e. torrente - lujan . + _
ific- csic , dept .
de fisica teorica + universidad de valencia , 46100 burjassot , valencia , spain . + e - mail : [email protected]._ |
the investigation of transport phenomena in metals is very significant since we can extract a large amount of important information with regard to the electronic states of metals . in conventional metals with weak electron - electron correlation
, various transport phenomena are governed by a single parameter , namely , the quasiparticle relaxation time @xmath0 .
that is , the rta holds well in such conventional metals @xcite .
for example , resistivity @xmath6 is proportional to @xmath7 , which is proportional to @xmath8 in conventional fermi liquids .
the hall coefficient @xmath9 is independent of @xmath0 , and @xmath10 expresses the approximate carrier density @xmath11 .
the magnetoresistance @xmath12 is proportional to @xmath13 .
the thermoelectric power @xmath14 is proportional to @xmath15 for a wide range of temperatures .
the signs of @xmath9 and @xmath14 represent the type of carrier ( i.e. , electrons or holes ) .
the behaviors of these transport coefficients , which are referred to as fermi liquid behaviors , are well explained by the rta . in strongly correlated electron systems ,
however , transport coefficients frequently exhibit a prominent deviation from conventional fermi liquid behaviors . because of this , it is difficult to obtain information on the electronic states from the transport phenomena .
for example , in high-@xmath1 superconductors ( htscs ) , both @xmath9 and @xmath16 are very sensitive to temperature ; in some compounds , sign changes occur with changes in the temperature @xcite .
evidently , a simple rta does not work well for htscs .
for example , @xmath10 is considerably greater than the electron density in optimally doped or under - doped systems .
prominent `` non - fermi - liquid - like behaviors '' are also observed in other strongly correlated systems such as organic superconductors @xcite and heavy - fermion systems ( @xmath17-electron systems ) @xcite .
although these behaviors reveal essential information about the electronic states , only a little was understood about them for a long time . to obtain a significant amount of information from these wealth of treasures , we had to develop the theory of transport phenomena in strongly correlated fermi liquids .
the main objective of the present study is to investigate the non - fermi - liquid - like transport properties in htscs , which have been intensively studied as one of the central issues of htsc . thus far , two different kinds of theoretical models have been actively investigated : in the first kind of models , non - fermi - liquid ground states are assumed .
for example , anderson proposed that the elementary excitations are described by spinons and holons .
it is considered that holons , whose density is small in under - doped systems , are responsible for the anomalous transport @xcite . in the second kind of models , fermi - liquid ground states are assumed .
for example , transport coefficients were calculated based on the nearly antiferromagnetic ( af ) fermi liquid picture @xcite . in the previous studies ,
the rta has been frequently used @xcite . however , both these models can explain only a limited number of experimental facts involved in htsc : many other anomalous transport properties remain unsolved . to understand the true ground state of htsc , it is highly desirable to solve the issue of anomalous transport phenomena in htscs in a unified way .
the main aim of the present study is to explain that the _ rich variety of anomalous transport phenomena observed in htscs can be understood in a unified way in terms of the fermi liquid picture_. for this purpose , we have to take the cvc into account correctly , which is totally dropped in the rta . in interacting electron systems ,
an excited electron induces other particle - hole excitations by collisions .
the cvc represents the induced current due to these particle - hole excitations .
the cvc is closely related to the momentum conservation law , which is mathematically described using the ward identity @xcite . in fact , landau proved the existence of the cvc , which is called backflow in the phenomenological fermi liquid theory , as a natural consequence of the conservation law @xcite .
the cvc can be significant in strongly correlated fermi liquids owing to strong electron - electron scattering .
however , its effect on the transport phenomena has not been studied well until recently . in the present study
, we discuss the role of cvc in nearly af fermi liquids , such as htscs , organic metals , and heavy - fermion systems near the magnetic quantum critical point ( qcp ) .
for example , the rta for the highly anisotropic @xmath18 model can not explain the transport anomaly in htsc . we find that _ the rta is unreliable in the presence of strong af fluctuations , because the prominently developed cvc entirely modifies the rta results _ @xcite . in a fermi liquid , the transport coefficients are described by the total current @xmath19 , which is expressed as a sum of quasiparticle velocity @xmath20 and cvc @xcite . in the present study
, we find that the _ @xmath19 in nearly af metals shows anomalous @xmath21-dependence due to the cvc .
_ this is the origin of the transport anomalies in nearly af metals .
based on the microscopic fermi liquid theory , we investigate the important role of the cvc in @xmath9 , @xmath12 , @xmath14 , and the nernst coefficient ( @xmath22 ) .
furthermore , we confirm this idea by performing a numerical study based on the flex+cvc theory . in this approximation
, the coulomb interaction @xmath23 is the only fitting parameter . to demonstrate the physical meaning of cvc
, we discuss the scattering processes between quasiparticles in an isotropic model @xcite . in the rta
, it is assumed that the conductivity due to a quasiparticle at @xmath21 is proportional to the mean free path @xmath24 .
since @xmath25 in a fermi liquid @xcite , the resistivity according to the rta @xmath26 is finite .
however , this result is _ not true _ since the momentum conservation law ensures that @xmath27 in a spherical model in the absence of the umklapp process @xcite .
the failure of the rta originates from the assumption that the velocity of the excited quasiparticle disappears after scattering . here
, we explain that the correct answer ( zero resistivity ) in the absence of the umklapp processes is recovered by considering all the relevant normal scattering process as shown in fig .
[ fig : cvc - sp ] . when a quasiparticle at @xmath21 is scattered to @xmath28 after the relaxation time @xmath18 , a particle - hole pair ( at @xmath29 and @xmath30 ) should be created according to the momentum and energy conservation laws . _
the cvc represents the current conveyed by the particles at ( @xmath28 , @xmath29 ) and a hole at @xmath30 , which emerge during the scattering process .
_ therefore , the momentum and energy conservation laws , which are violated in the rta , are restored by considering the cvc @xcite .
the cvc is necessary to reproduce the zero - resistivity in the absence of the umklapp process . in the rta
, we consider an excited quasiparticle at @xmath21 ( in fig .
[ fig : cvc - sp ] ) as if it _ annihilates _ after the relaxation time .
since the rta allows such an unphysical process , the conservation law @xmath31 is violated in the rta .
since the conservation law is a very important constraint on the transport properties , the rta frequently yields unphysical results @xcite .
in later sections , we find that the cvc is significant in nearly af metals : we explain that the cvc is the origin of a variety of anomalous transport phenomena in such metals . .
after the relaxation time @xmath0 , the quasiparticle at @xmath21 collides with another quasiparticle at @xmath30 in the fermi sea . as a result of this collision ,
the two quasiparticles at @xmath28 and @xmath29 , and a quasi - hole at @xmath30 are created .
the total momentum is conserved in this process . ] in htscs , almost all the transport phenomena deviate from the conventional fermi liquid behaviors , which are referred to as the non - fermi liquid behaviors @xcite .
these anomalous transport phenomena have been studied intensively as one of the most important issues in htscs , since they offer important clues to reveal the true ground state in htscs , which has been unsolved until now . to answer this question ,
many analytical and numerical studies have been performed to determine the ground state of two - dimensional ( 2d ) systems .
for example , anderson et al . considered that the ground state of a square - lattice hubbard model is the resonating - valence - bond ( rvb ) state , where spin and charge degrees of freedom are separated @xcite . in the rvb state ,
concept of quasiparticle is not valid . according to his philosophy , possibility of the rvb state in the square - lattice @xmath32-@xmath33 model
was studied , by using the mean - field theory @xcite and the gauge theory @xcite . in
[ nonfl ] , we will introduce significant works performed in the @xmath32-@xmath33 model .
unfortunately , it is difficult to perform quantitative studies of transport coefficients using the @xmath32-@xmath33 model . .
at the same time , @xmath9 starts to increase .
below @xmath34 ( in the pseudo - gap region ) , the af fluctuations are suppressed since the strong sc fluctuations reduce the density of states at the fermi level , which is called the pseudo - gap .
in this study , the filling area is @xmath35 . ] on the other hand , the fermi liquid theory has been developed and applied to analyzing the htscs , considering that the fermi liquid ground state is realized @xcite . by using the self - consistent renormalization ( scr ) theory , moriya , takahashi and ueda explained that the correct superconducting ( sc ) order parameter @xmath36 as well as @xmath15-linear resistivity are derived from the strong af fluctuations @xcite .
based on the phenomenological spin fluctuation model , monthoux and pines performed a quantitative analysis for optimally - doped ybco @xcite .
further , bickers et al . studied the square - lattice hubbard model according to a self - consistent random - phase - approximation ( rpa ) , which is now called the fluctuation - exchange ( flex ) approximation @xcite .
these spin fluctuation theories have succeeded in reproducing various non - fermi - liquid behaviors in the normal state of htscs : for example , temperature dependence of nuclear spin - lattice relaxation rate given by nmr / nqr measurements , @xmath37 , is reproduced well . moreover , famous @xmath15-linear resistivity in htsc , @xmath38 is also explained .
( the fermi liquid behavior @xmath39 will recover at very low @xmath15 if we suppress the superconductivity . )
the flex approximation is also useful for the study of electron - doped systems @xcite .
these spin fluctuation theories successfully reproduce various non - fermi - liquid - like behaviors in htscs except for the under - doped region @xmath40 , where @xmath11 is the number of electrons per site . a schematic phase diagram of htsc is shown in fig .
[ fig : htsc - phase ] .
the fermi liquid description of htscs is still in progress @xcite .
a @xmath41-wave symmetry in the sc state was confirmed by phase - sensitive measurements @xcite , and tunneling spectroscopy @xcite .
@xmath41-wave sc state was also derived according to the third - order - perturbation theory with respect to @xmath23 @xcite . ]
regardless of the remarkable success of the fermi liquid theory during the 90s , the non - fermi - liquid - like transport phenomena observed in htscs remained unresolved for a long time .
the significant deviation from the rta results in htscs was frequently considered as a hallmark of the breakdown of the fermi liquid state .
this issue has remained as one of the most important problems in htscs for a long time . in htscs
, @xmath42 increases below @xmath43k as @xmath44 above the pseudo - gap temperature @xmath45k , and @xmath46 at @xmath34 @xcite .
the sign of @xmath9 is positive in hole - doped systems such as la@xmath47sr@xmath48cuo@xmath49 ( lsco ) @xcite , yba@xmath50cu@xmath51o@xmath52 ( ybco ) @xcite and tl@xmath50ba@xmath50cuo@xmath53 ( tbco ) @xcite ; however , it is negative in electron - doped systems such as nd@xmath47ce@xmath48cuo@xmath49 ( ncco ) @xcite and pr@xmath47ce@xmath48cuo@xmath49 ( pcco ) @xcite , even though the angle resolved photoemission ( arpes ) measurements resolved hole - like fermi surfaces @xcite .
the experimental @xmath15-dependences of @xmath9 for lsco and ncco are shown in fig .
[ fig : sato ] .
the magnetoresistance of htscs also shows strong temperature dependence as @xmath54 for a wide range of temperatures in lsco @xcite , ybco @xcite and tbco @xcite .
these results completely contradict with kohler s rule ( @xmath55const . and
@xmath56 ) which is derived using the rta for a single - band model .
interestingly , the following `` modified kohler s rule '' holds well for optimally - doped lsco @xcite , 90k and 60k ybco @xcite and tbco @xcite : @xmath57 where @xmath58 is the hall angle .
this experimental fact strongly suggests that the anomalous behaviors of the hall effect and the magnetoresistance have the same origin . below @xmath34 ,
@xmath9 decreases whereas @xmath12 increases further @xcite .
therefore , modified kohler s rule is not very well applicable for under - doped htscs at low temperatures @xcite . for a long time , anomalous transport phenomena have been considered as one of the strongest objection against the fermi liquid description of htscs .
for example , to explain eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) , anderson introduced the tomonaga - luttinger model with two types of relaxation times .
however , it is not obvious how to describe the crossover from the tomonaga - luttinger liquid state in the under - doped region to the landau - fermi liquid state with doping . on the other hand , the fermi liquid description for htsc
, starting from the well - established fermi liquid state in the over - doped region , appears to account for a wide doping range in htscs @xcite . here
, we conform to the following principle : before abandoning the fermi liquid picture in htscs , we have to verify whether the rta is really applicable in htscs or not .
we stress that the rta may be unreliable for strongly correlated fermi liquids since the cvc is not considered , regardless of its importance to satisfy the conservation laws . due to this inadequacy
, the rta sometimes yields unphysical results in correlated metals @xcite .
in later sections , we will explain that the non - fermi liquid - like behaviors in htscs can be understood based on the fermi liquid theory by including the cvc . here
, we discuss the pseudo - gap phenomena in slightly under - doped systems below @xmath45k , which is also one of the most important issues in htsc .
@xmath59 starts to decrease below @xmath34 , which means that the af fluctuations are suppressed in the pseudo - gap region @xcite .
according to arpes measurements , prominent deep pseudo - gap appears in the density of states ( dos ) below @xmath34 @xcite .
a simple spin fluctuation theory can not explain the various anomalous phenomena in the pseudo - gap region .
recent theoretical studies using the @xmath15-matrix theory @xcite and the flex+@xmath15-matrix theory @xcite have shown that the strong sc amplitude fluctuations , which are induced by the af fluctuations , are a promising candidate for the origin of pseudo - gap @xcite . in [ flex - t ] , we study transport phenomena below @xmath34 using the flex+@xmath15-matrix theory by including cvcs , and show that the various anomalous transport coefficients are well reproduced in a unified way .
the present study strongly supports the idea that the pseudo - gap phenomena in under - doped htscs are induced by the strong sc amplitude fluctuations with a @xmath36-symmetry @xcite .
we note that wide and shallow pseudo - gap in the dos ( weak pseudo - gap ) is observed by arpes even above @xmath60 , which is considered to originate from af fluctuations that appear below @xmath43 k. we will discuss the weak pseudo - gap phenomena in [ validityflex ] . during the last decade
, it has been found that in strongly correlated materials , including heavy - fermion systems and organic metals , various transport coefficients exhibit striking deviations from the fermi liquid behaviors . in particular , anomalous transport properties similar to those in htscs ( eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) ) have been observed in many systems with strong magnetic fluctuations such as @xmath4-(bedt - ttf)@xmath50x [ x = cu[n(cn)@xmath50]br @xcite , x = cu[n(cn)@xmath50]cl @xcite , x = cu(ncs)@xmath50 @xcite ] , @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 @xcite , and ce@xmath2in@xmath3 ( @xmath2=co , rh ) @xcite .
bedt - ttf is an abbreviation of bis(ethylenedithio)tetrathiafulvalene .
these experimental facts strongly suggest that the transport anomaly given by eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) is not a problem specific to htscs , but a universal property in the nearly af fermi liquids @xcite .
the study of these transport phenomena in such systems will serve to resolve the origin of the non - fermi - liquid - like behaviors in htscs .
@xmath4-(bedt - ttf ) is a layered organic compound made of bedt - ttf molecules .
the @xmath64-wave superconductivity can be realized in a wide region of the pressure - temperature ( @xmath65-@xmath15 ) phase diagram , adjacent to the af insulating states @xcite .
for example , cu[n(cn)@xmath50]cl salt at ambient pressure is an af insulator , with its nel temperature @xmath66 k. with increasing pressure , @xmath67 decreases and superconductivity appears via a weak first order transition ; the maximum @xmath1 is 13 k at 200 bar .
an effective theoretical model can be given by the anisotropic triangular lattice hubbard model at half - filling @xcite . according to this model , the phase diagram of @xmath4-(bedt - ttf ) can be well reproduced by using the flex approximation , by assuming that @xmath68 decreases by pressure @xcite .
these studies revealed that the @xmath36-wave superconductivity occurs due to the strong af fluctuations @xmath69 . in the metallic phase of @xmath4-(bedt - ttf ) , the relationships @xmath70 and @xmath71 are observed @xcite
. moreover , the magnitude of @xmath9 decreases with increasing the pressure @xcite .
these behaviors are quantitatively reproduced by the flex+cvc theory @xcite .
ce@xmath2in@xmath3 is a quasi two - dimensional ( 2d ) heavy - fermion compound @xcite . at ambient pressure , cecoin@xmath3 is a superconductor with @xmath72 = 2.3 k. the electronic specific heat coefficient @xmath73 has been measured to be 300 mj / k@xmath74mol at @xmath75 .
ceirin@xmath3 is also a superconductor at ambient pressure with @xmath72 = 0.4 k and @xmath76 680 mj / k@xmath74mol .
cerhin@xmath3 is an af metal with nel temperature @xmath67 = 3.8 k at ambient pressure . under the pressure
, cerhin@xmath3 undergoes a sc transition at @xmath77 = 2 gpa , indicating that the af quantum critical point ( qcp ) is located at or in the vicinity of @xmath77 .
the nmr relaxation rate @xmath78 measurements indicate the presence of quasi 2d af spin fluctuations in the normal state @xcite .
the measurements of the angle resolved thermal conductivity @xcite and specific heat @xcite revealed that the symmetry of the sc state exhibits @xmath64-wave symmetry .
furthermore , the fulde - ferrell - larkin - ovchinnikov ( fflo ) sc state has been observed @xcite . in ce@xmath2in@xmath3 , eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) are well satisfied for a wide range of temperatures , and the value of @xmath79 reaches @xmath80 in cerhin@xmath3 near the af qcp @xcite .
similarly , the nernst coefficient @xmath22 also takes a large positive value at low temperatures @xcite .
recently , we studied both @xmath9 and @xmath22 in a quasi 2d hubbard model , and found that they are prominently enhanced , as large as those in pure 2d systems @xcite . in ce@xmath2in@xmath3 ,
modified kohler s rule ( [ eqn : mkrexp ] ) is well satisfied for @xmath81 tesla , whereas both @xmath82 and @xmath83 are drastically suppressed by a very weak magnetic field ( @xmath84 tesla ) near the af qcp @xcite .
this surprising fact can be understood in terms of the field - dependence of the cvc . in the present article
, we assume that the fermi liquid state is realized in a wide range of the phase diagram in htscs , and the non - fermi - liquid - like behaviors are created by strong spin ( and sc ) fluctuations . according to this idea , we study the anomalous transport phenomena in htsc , by carefully analyzing the main - body effects that had been overlooked in previous studies .
especially , we intensively study the cvc based on the microscopic fermi liquid theory .
our final aim is to explain the anomalous transport phenomena in various nearly af fermi liquids _ in a unified way _
, including htsc , heavy fermions and organic metals . however , it is a nontrivial question whether the fermi liquid state is realized in two - dimensional strongly correlated electron systems near the half - filling .
in fact , the fermi liquid state seems to be broken in heavily under - doped htscs . in the infinite dimension hubbard model , the fermi liquid state with heavy mass
is realized next to the mott insulating state @xcite . in 2d hubbard models
@xcite , it was rigorously proved that the limit value of the interaction @xmath23 , below which the fermi liquid state is realized , is finite . however , in strongly correlated 2d systems , an exotic non - fermi liquid ground state may be realized next to the mott insulating state due to the strong quantum fluctuations .
that is , removal of large tracts of the hilbert space due to strong correlation effect may lead to a violation of the fermi liquid state . to study the strong correlation effect in htscs
, the @xmath32-@xmath33 model has been frequently analyzed .
the @xmath32-@xmath33 model is derived from the hubbard model ( or @xmath64-@xmath85 model ) by a canonical transformation , by excluding the double occupancy of holes to represent the strong coulomb interaction .
it is given by @xmath86 where @xmath87 represents the exclusion of the doubly occupied state , @xmath88 is the hopping integral between @xmath89 sites , and @xmath33 is the superexchange energy between the neighboring spins .
@xmath90 mev in real htscs . in the @xmath32-@xmath33 model ,
the hilbert space with high - energy state is eliminated by @xmath87 , which enables us to perform numeral calculations easier or to invent new approximations . based on the @xmath32-@xmath33 model , various versions of the new fluid
have been proposed , with novel kinds of excitation , many involving gauge theories of spin - charge separated spinons and holons @xcite . the exact diagonalization technique @xcite has been applied to the square - lattice @xmath32-@xmath33 model .
the ground state phase diagram of the @xmath32-@xmath33 model has been studied using the variational monte carlo method @xcite , and it was found that the @xmath36-wave sc state is realized in a wide range of the phase diagram . unfortunately , quantum monte carlo ( qmc ) simulation for @xmath32-@xmath33 model is difficult because of a serious negative sign problem . instead
, qmc simulation for the hubbard model with moderate value of @xmath23 @xcite have been performed intensively . on the other hand , many authors have considered non - fermi liquid ground states due to novel quantum criticalities , other than a conventional ( scr - type ) quantum criticality near the spin density wave ( sdw ) state @xcite .
for example , varma et al . proposed that the marginal fermi liquid state is realized in htscs , where the @xmath21-independent self - energy is given by @xcite @xmath91 where @xmath92 , @xmath93 is a coupling constant , and @xmath94 is a cutoff energy .
this state is not a fermi liquid since the quasiparticle renormalization factor @xmath95 vanishes logarithmically as @xmath96 . the self - energy in eq .
( [ eqn : mfl ] ) can be derived if electrons couple to the following @xmath21-independent charge and spin fluctuations that are singular at @xmath97 : @xmath98 in this model , @xmath99 , @xmath100 , and the optical conductivity is @xmath101 .
they are typical non - fermi liquid behaviors in htscs .
now , a microscopic derivation of the @xmath21-independent quantum critical fluctuations is an important issue .
varma proposed @xcite that the circulating current phase exists in the under - doped regime , and the current fluctuations near the qcp ( @xmath102 at the optimum doping ) are the origin of the marginal fermi liquid state . in the present article
, we will argue that transport anomaly near the af - qcp can happen even if the fermi liquid state ( @xmath103 ) remains intact at the qcp .
we will show that the cvc in the landau fermi liquid theory causes various striking quantum critical behaviors .
here , we discuss the functional form of the dynamical spin susceptibility @xmath104 in nearly af metals .
hereafter , we use the unit @xmath105 .
this is the most important physical quantity in such metals since it is the origin of various non - fermi liquid behaviors in htscs .
the phenomenological form of @xmath104 , which can be obtained by using nmr / nqr spectroscopy and the neutron diffraction measurement , is given by @xcite @xmath106 where @xmath107 is the antiferromagnetic ( af ) wavevector , and @xmath108 is the af correlation length .
this is referred to as the millis - monien - pines model @xcite . in htscs above the pseudo - gap temperature @xmath34 , both @xmath109 and @xmath110 are scaled by @xmath111 as follows @xcite : @xmath112 where @xmath113 , @xmath114 , @xmath115 and @xmath116 are constants . since @xmath117 in eq .
( [ eqn : parameters2 ] ) , the corresponding dynamical exponent @xmath118 is 2 .
the coefficient @xmath114 rapidly increases in the under - doped region : @xmath108 reaches @xmath119 in optimally - doped ybco , and it exceeds @xmath120 in slightly under - doped ncco just above @xmath1 @xcite .
( @xmath121 is the unit - cell length ; we put @xmath122 hereafter . )
the relationship @xmath123 ( @xmath124 ) is satisfied in the over - doped ( under - doped ) ybco . on the basis of the phenomenological @xmath104 model , the sc transition temperature was successfully reproduced by solving the strong - coupling eliashberg equation @xcite . theoretically , the relationships in eqs .
( [ eqn : parameters1 ] ) and ( [ eqn : parameters2 ] ) can be explained according to the scr theory , where we consider the renormalization of the dynamical susceptibility due to both self - energy correction and vertex correction .
this renormalization effect is referred to as the `` mode - mode coupling effect '' @xcite , which represents the destruction of the af long - range order due to thermal and quantum fluctuations .
relationships ( [ eqn : parameters1 ] ) and ( [ eqn : parameters2 ] ) are also reproduced according to the flex approximation @xcite . in this approximation ,
only the mode - mode coupling effect due to the self - energy correction is taken into account : since the large imaginary part of the self - energy reduces the dos , the spin susceptibility given by the mean - field approximation ( which is equivalent to the rpa ) is drastically suppressed in the flex approximation . as a result , the magnetic long - range order does not occur in pure 2d systems , which means that the mermin - wagner theorem is satisfied in the flex approximation @xcite .
in nearly af metals , the flex approximation and the scr theory yield similar results since the mode - mode coupling effect due to the self - energy is dominant @xcite . on the other hand ,
the mode - mode coupling effect due to the second - order vertex correction ( aslamazov - larkin term ) can be important in the nearly ferromagnetic metals .
in the main part of this study , we investigate the transport phenomena in htscs based on the fermi liquid theory , and perform a numerical study using the spin fluctuation theory . in this study , we analyze the following hubbard model : @xmath125 where @xmath23 is the coulomb interaction , and @xmath126 is the spectrum of the conduction electron . in a square lattice , @xmath126 is given by @xmath127 where @xmath128 is the creation operator of an electron with momentum @xmath21 and spin @xmath129 , and @xmath23 is the on - site coulomb repulsion .
we represent the electron filling by @xmath11 , and @xmath130 corresponds to the half - filling .
to fit the band structures given by the local density approximation ( lda ) band calculations for ybco @xcite , ncco @xcite , and lsco @xcite and by the angle resolved photoemission ( arpes ) experiments for ybco @xcite , ncco @xcite and lsco @xcite , we select the following set of parameters @xcite .
( i ) ybco ( hole - doping ) and ncco ( electron - doping ) : @xmath131 , @xmath132 , and @xmath133 .
( ii ) lsco ( hole - doping ) : @xmath131 , @xmath134 , and @xmath135 . here
, @xmath11 is smaller ( larger ) than unity in ybco and lsco ( ncco ) .
the fermi surfaces for ybco and ncco without interaction are shown in fig .
[ fig : fs - hotcold ] ( i ) .
the deformation of the fermi surface in the presence of @xmath23 has been discussed in ref .
note that the fermi surface in bi@xmath50sr@xmath50cacu@xmath50o@xmath63 ( bscco ) is similar to that in ybco @xcite . since @xmath136k in htscs , @xmath137 in the present study corresponds to @xmath138k . ) and ncco ( @xmath139 ) .
the location of the hot spots and the cold spots are shown . according to the flex approximation , the hot spot in ybco shifts to point b , by reflecting the large dos at @xmath140 . *
( ii ) * diagrammatic representation of the self - energy in the one - loop ( flex ) approximation . ] here , we study the self - energy , dynamical spin susceptibility and the cvc using the flex approximation @xcite .
the flex approximation is classified as a conserving approximation whose framework was constructed by baym and kadanoff @xcite and by baym @xcite . for this reason , we can calculate the cvc without ambiguity , by following the ward identity @xmath141 . here
, the green function and the self - energy are given by @xmath142 where @xmath143 and @xmath144 , respectively .
the self - energy in eq .
( [ eqn : self ] ) is schematically shown by fig .
[ fig : fs - hotcold ] ( ii ) .
we solve the eqs .
( [ eqn : self])-([eqn : chi0 ] ) self - consistently , choosing @xmath145 so as to satisfy @xmath146 .
the flex approximation is suitable for the analysis of nearly af fermi liquids .
many authors have applied this approximation to the square - lattice hubbard model @xcite .
although it is an approximation , the imaginary - time green function obtained from the flex approximation is in good agreement with the results obtained from the quantum monte carlo simulations for a moderate value of @xmath23 @xcite .
it has also been applied to the sc ladder compound , sr@xmath147ca@xmath148cu@xmath149o@xmath150 @xcite , and the organic sc @xmath4-(bedt - ttf ) compounds @xcite . using the flex approximation , @xmath1 was studied in various types of tight - binding models @xcite .
the flex approximation predicts that the af fluctuations are predominant in a square - lattice hubbard model near half - filling .
this result has been confirmed by renormalization group analyses , which considers both spin and charge fluctuations on the same footing by solving the parquet equation @xcite . for ( i ) ybco ( @xmath151 ) at @xmath152 and ( ii ) for ncco ( @xmath153 ) at @xmath154 , given by the flex approximation . in ybco , @xmath155 , whereas @xmath108 for ncco exceeds @xmath156 at @xmath152 .
@xmath121 is the lattice spacing .
@xcite ] ] the flex approximation can not reproduce the pseudo - gap behaviors below @xmath157k in slightly under - doped systems .
however , they are well explained by the flex+@xmath15-matrix approximation @xcite , where the self - energy correction due to strong sc fluctuations , which are induced by af fluctuations , are considered self - consistently .
we will explain about the flex+@xmath15-matrix approximation in
[ flex - t ] . in the present study
, we perform numerical studies from the over - doped region to the slightly under - doped region ( i.e. , @xmath158 or @xmath159 ) , where the flex(+@xmath15-matrix ) approximation yields reasonable results .
note that the flex approximation is inappropriate to describe the `` mott physics '' in the heavily under - doped region .
figure [ fig : flex - chi ] shows the spin susceptibility @xmath160 given by the flex approximation , both for ybco ( @xmath151 ; optimum doping ) at @xmath152 and for ncco ( @xmath153 , slightly over - doping ) at @xmath154 , respectively @xcite . since the nesting of the fermi surface is not good as shown in fig .
[ fig : fs - hotcold ] , @xmath161 is moderate even at low temperatures . in optimum ybco ,
the stoner factor @xmath162 at @xmath152 , and the af correlation length @xmath108 is approximately @xmath163 @xmath121 ( @xmath121 denotes the lattice spacing ) . on
the other hand , @xmath108 in ncco ( @xmath153 ) exceeds 10 @xmath121 at @xmath152 .
it should be noted that @xmath160 given by the rpa shows an incommensurate structure , which is inconsistent with the experimental results .
i verified that @xmath59 of the cu nuclei under @xmath164 given by the flex approximation for @xmath151 is @xmath165 [ 1/k msec ] at 200k , which is consistent with the experimental results @xcite .
one of the advantages of the flex approximation is that the mermin - wagner theorem with respect to the magnetic instability is satisfied : its analytic proof is given in the appendix a of ref .
hence , the critical region ( @xmath166 ) is stable in 2d systems since the sdw order ( @xmath167 ) is prevented by the mermin - wagner theorem .
that is , @xmath168 in 2d systems . as a result ,
the @xmath23 dependence on @xmath9 ( and @xmath169 ) given by the flex+cvc approximation is rather moderate , as shown in refs .
@xcite . .
@xmath170 is the irreducible four point vertex with respect to the particle - hole channel @xmath171 . ] in the next stage , we derive the two - particle green function @xmath172 which is indispensable for the study of transport phenomena . according to the microscopic fermi liquid theory @xcite
, @xmath172 can be obtained by the solution of the following bethe - salpeter equation : @xmath173 where @xmath174 and @xmath175 is the full four - point vertex .
@xmath176 is the irreducible four - point vertex , which is given by the fourier transformation of the ward identity in real space ; @xmath177 .
the bethe - salpeter equation is expressed by fig .
[ fig : l ] .
in later sections , we show that various linear transport coefficients are described in terms of @xmath178 .
( e.g. , see .
( [ eqn : sig - d ] ) . ) in the conserving approximation , transport coefficient obtained by @xmath172 in eq .
( [ eqn : l ] ) automatically satisfies conservation laws @xcite .
this is the reason why we refer to it as the conserving approximation .
this is a great advantage of the flex approximation for the study of transport phenomena . in the flex approximation ,
irreducible four - point vertex will be given in eq .
( [ eqn : ivc - flex ] ) in [ total - current ] .
as we have explained , important advance in htscs has been achieved by using the fermi liquid theory with strong af fluctuations @xcite .
one of the most important predictions given by these theories is the `` hot / cold - spot structure '' of the quasiparticle damping rate , @xmath179 ( @xmath180 ) .
that is , @xmath181 becomes anisotropic in the presence of af fluctuations .
this fact is very important to understand the transport phenomena in htscs .
the portions of the fermi surface at which @xmath181 takes the maximum and minimum values are referred to as _ hot spots _ and ( _ cold spots _ ) , respectively @xcite . according to the spin fluctuation theory
, the hot spots usually exist around the crossing points with the af brillouin zone ( afbz)-boundary , whereas the cold spot is at the points where the distance from the afbz - boundary is the largest .
their positions are shown in fig .
[ fig : fs - hotcold ] ( i ) .
the electronic states around the cold spots play the major role for various transport phenomena .
note that the hot spot in ybco shifts to the brillouin zone boundary ( point b ) , by reflecting the large dos at @xmath140 .
figure [ fig : del ] shows the @xmath21-dependence of @xmath182 on the fermi surface given by the flex approximation @xcite . in ybco , the hot spot shifts from point c in fig .
[ fig : fs - hotcold ] ( i ) to the bz - boundary [ point b ] , by reflecting the large dos at the van - hove singularity point @xmath140 .
therefore , the spectral weight at the fermi energy is strongly reduced around @xmath140 due to large @xmath183 , which is consistent with the arpes experiments @xcite . in ncco ,
in contrast , the hot spot and the cold spot are located at point c and point b , respectively .
we emphasize that the location of the cold spot in ncco was first predicted by the flex approximation in 1999 @xcite , and it was later confirmed by arpes @xcite .
according to the spin fluctuation theories @xcite , @xmath184 , and @xmath185 except that @xmath186 ; @xmath187 is shown in fig . [
fig : fs - hotcold ] ( i ) .
-dependence of @xmath182 on the fermi surface at various temperatures .
the cold spot in ybco ( ncco ) is point a ( b ) . ]
the critical value of @xmath23 for a spin density wave ( sdw ) transition in the rpa ( i.e. , the mean - field approximation ) is @xmath188 in lsco ( @xmath189 ) . in ybco and ncco
, @xmath190 takes a much larger value , @xmath191 for both ybco ( @xmath189 ) and ncco ( @xmath192 ) , since the nesting is not good due to the large next - nearest and third - nearest hopping integrals ( @xmath193 and @xmath194 ) . here
, we consider to put the same @xmath23 for both ybco and ncco . in ybco ,
the dimensionless coupling constant @xmath195 is large , since the saddle point is close to the fermi level . here , @xmath196 is the dos at the fermi level .
therefore , @xmath197 takes a large value , which significantly reduces the interacting dos and @xmath198 at @xmath199 @xcite .
in fact , the obtained @xmath200 is considerably smaller than unity in ybco . on the other hand ,
the reduction in @xmath201 due to @xmath197 is small in ncco since the saddle point is far below the fermi level .
for this reason , @xmath108 for ncco is much larger than that for ybco in the flex approximation .
this result is consistent with experiments .
the real - frequency green function @xmath202 is given by the analytic continuation of @xmath203 in eq .
( [ eqn : g ] ) . in a fermi liquid
, the advanced green function @xmath204 in the vicinity of @xmath205 and @xmath206 can be represented as @xmath207 where @xmath208 is the renormalization factor given by @xmath209 , and @xmath210 is the renormalized quasi - particle spectrum , which is the solution of @xmath211 .
the quasiparticle weight is given by @xmath212 the dos is expressed as @xmath213 . in the case of @xmath214 , @xmath215 for @xmath216 .
in the flex approximation , @xmath181 is obtained by the analytic continuation of eq .
( [ eqn : self ] ) : @xmath217 { \rm im}v_\q(\e+i\delta ) \rho_{\k+\q}(\e ) \label{eqn : gam - flex - true } \end{aligned}\ ] ] where @xmath218 is given by the analytic continuation of eq .
( [ eqn : def_v ] ) . in the spin fluctuation model ( millis - monien - pines model in eq .
( [ eqn : kai_qw ] ) ) , @xmath219 in eq .
( [ eqn : gam - flex - true ] ) is replaced with @xmath220 , where @xmath221 . according to refs .
@xcite , eq .
( [ eqn : gam - flex - true ] ) is approximately transformed to @xmath222 according to eq .
( [ eqn : gam - k ] ) , @xmath183 ( @xmath181 at the hot spot ) in 2d systems is given by @xcite : @xmath223 since @xmath224 , @xmath184 for any value of @xmath225 : this result is recognized in the numerical study in fig .
[ fig : del ] .
also , @xmath226 in 2d systems is obtained as @xcite : @xmath227 } , \label{eqn : g - cold1 } \\
\g_{\rm cold}&\propto & t^2 \ \ \ \mbox{for $ \pi t/2\w_{\rm sf}\ll(\xi_{\rm af } \delta k_c)^2$}. % [ low-$t$ region]}. \label{eqn : g - cold2}\end{aligned}\ ] ] according to eqs . (
[ eqn : parameters1 ] ) and ( [ eqn : parameters2 ] ) , @xmath228 is constant if @xmath229 , and it is of the order of @xmath230 in optimally - doped htscs .
therefore , @xmath185 when @xmath231 , and @xmath232 when @xmath233 .
[ here , @xmath187 represents the `` deviation from the nesting condition at the cold spot '' ; see fig . [
fig : fs - hotcold ] . ]
we comment that the almost all part of the fermi surface becomes the cold spot in the case of @xmath233 .
now , we discuss the temperature dependence of resistivity according to the scr theory @xcite , by dropping the cvc . in the scr theory ,
the resistivity is derived from the born approximation @xmath234 . in the case of @xmath235 , eq .
( [ eqn : gam - k ] ) is simplified as @xmath236 where @xmath237 .
it can also be derived directly from eq .
( [ eqn : gam - flex - true ] ) by using the relation @xmath238\e=(\pi t)^2 $ ] . using eqs .
( [ eqn : kai_qw ] ) and ( [ eqn : gam - flex ] ) , the temperature dependence of the resistivity is given by @xcite @xmath239 where @xmath64 is the dimension of the system . in deriving eq .
( [ eqn : g - born ] ) , we utilized the fact that the @xmath240-dependence of im@xmath241 is moderate . since @xmath242 near the af - qcp @xcite , @xmath243 is proportional to @xmath244 ( @xmath245 ) .
we stress that the fermi liquid behavior @xmath246 is recovered when @xmath247constant away from the af - qcp , like in over - doped systems at low temperatures . in [ resrh ]
, we will calculate the resistivity using the flex approximation ( and flex+@xmath15-matrix approximation ) , based on the linear response theory . in fig .
[ fig : rho ] , `` flex '' represents the resistivity obtained by dropping the cvc , and `` flex+cvc '' represents @xmath6 given by the flex+cvc approximation ; the latter gives the correct result .
consistently with eq .
( [ eqn : g - born ] ) , the obtained @xmath6 follows an approximate @xmath15-linear behavior in under - doped lsco and ncco , and it shows a @xmath8-like behavior in the over - doped ncco . as for the resistivity ,
the cvc is quantitatively important . in later sections
, we explain that the cvc completely changes the @xmath15-dependence of @xmath9 , @xmath12 and @xmath22 .
however , the born approximation ( eq . ( [ eqn : g - born ] ) ) gives overestimated values when @xmath248 is highly anisotropic , as pointed out by refs .
@xcite : according to the linear response theory , the correct resistivity is give by @xmath249 according to eqs .
( [ eqn : g - cold1 ] ) and ( [ eqn : g - cold2 ] ) , @xmath6 shows the @xmath15-linear behavior in the case of @xmath231 , whereas @xmath250 in the case of @xmath233 . in optimally doped or slightly under - doped ybco and lsco , @xmath6 shows an approximate @xmath15-linear behavior , which means that the relation @xmath231 is satisfied in these compounds above @xmath1 .
in fact , if @xmath233 , the dominant part of the fermi surface should be the cold spot .
however , this result is inconsistent with the arpes measurements @xcite . in electron - doped systems
, on the other hand , @xmath251 seems to be realized at low temperatures even in optimally - doped systems , since @xmath108 reaches 100 just above @xmath1 in ncco @xcite .
the relationship @xmath251 is also satisfied in the present numerical study . in this case , @xmath181 becomes highly anisotropic and @xmath252 is realized @xcite . according to eq .
( [ eqn : r - inv - g ] ) , the relationship @xmath250 holds in the close vicinity of the af - qcp ; @xmath251 .
however , rosch @xcite pointed out that the relationship @xmath253 ( @xmath254 ) holds even at the af - qcp when the quasiparticle damping rate due to impurities @xmath255 is finite : in fact , at sufficiently low temperatures where elastic scattering is dominant ( @xmath256 ) , then eq . ( [ eqn : r - inv - g ] ) becomes @xmath257 therefore , according to eq .
( [ eqn : g - born ] ) , @xmath258 ( @xmath245 ) holds near the af - qcp in the presence of impurities . here
, we explained the @xmath15-linear resistivity in htscs based on the spin fluctuation theory .
there are other theories which accounts for the @xmath15-linear resistivity . as discussed in [ nonfl ] ,
the marginal fermi liquid hypothesis was proposed by varma et al . @xcite .
also , the tomonaga - luttinger model with two types of relaxation times was proposed by anderson @xcite .
unfortunately , based on these theories , it is difficult to calculate other transport coefficients such as @xmath9 and @xmath12 . in later sections
, we will explain various anomalous transport phenomena in htscs in a unified way based on the fermi liquid theory .
before investigating the transport coefficients using the microscopic fermi liquid theory , we briefly review the relaxation time approximation ( rta ) based on the bloch - boltzmann theory @xcite .
the cvc provides correction for the rta . in later sections
, we will see that the cvc becomes crucial near the af qcp .
the boltzmann equation in a nonequilibrium steady state is expressed as @xcite @xmath259 where we put @xmath260 , and @xmath261 is the charge of electron .
@xmath262 is the distribution function in a nonequilibrium steady state , and @xmath263 and @xmath264 are the electric and magnetic fields , respectively .
@xmath265 represents the rate of change in @xmath262 due to scattering between quasiparticles , which is called the collision integral . using the scattering amplitude @xmath266 for @xmath267
, it is given by @xcite @xmath268 \label{eqn : boltz2}\end{aligned}\ ] ] where the first ( second ) term represents the outgoing ( incoming ) scattering process .
to derive the conductivity , we have to linearize the boltzmann equations ( [ eqn : boltz ] ) and ( [ eqn : boltz2 ] ) with respect to @xmath263 . in solving these equations
, we frequently apply the rta ; @xmath269 , where @xmath270 and @xmath271 is the equilibrium distribution function .
then , the linearized boltzmann equation is simplified as @xmath272 the solution of eq .
( [ eqn : bol - rta ] ) is given by @xmath273 although it is a crude approximation , the rta can successfully explain the various transport phenomena in metals with weak correlation . in the rta
, the conductivity is given by @xmath274 , where the factor 2 is attributed to the spin degeneracy .
when the @xmath21-dependence of @xmath18 is moderate , the functional form of @xmath275 under the magnetic field is @xmath276 . here , we assume @xmath277 .
then , @xmath278 . in the rta ,
the longitudinal conductivity , hall conductivity , and magnetoconductivity are given by @xmath279 where @xmath280 .
when the @xmath21-dependence of @xmath18 is moderate , @xmath281 which is called the kohler s rule .
based on the rta , hussey et al .
studied transport coefficients in heavily over - doped tl2201 @xcite and lsco @xcite .
they determined the anisotropy of @xmath18 in tl2201 by measuring the angle - dependent magnetoresistance oscillation ( amro ) along @xmath282-axis @xcite , and calculated transport coefficients using eqs .
( [ eqn : sxx - rta])-([eqn : dsxx - rta ] ) .
the derived @xmath283 agrees with the experimental hall coefficient @xmath284 in heavily over - doped samples ( @xmath285 k ) , where the hall coefficient is small and its temperature dependence is tiny .
as the doping decreases , @xmath284 increases quickly and its temperature dependence becomes prominent , whereas the doping dependence of @xmath283 is more moderate .
in optimally doped htscs , however , the rta does not work : stojkovic and pines @xcite attempted to explain the violation of kohler s rule in htscs based on the rta .
they assumed the highly anisotropic @xmath18 model ( hot / cold - spot model ) , where only the quasiparticles near the cold spot contribute toward the transport phenomena . in this model
, @xmath9 can take a large value since the `` effective carrier density for the transport phenomena '' is reduced .
they calculated @xmath18 based on the millis - monien - pines model ; the anisotropy of @xmath18 reaches 100 in the optimally doped ybco , which is too large to be consistent with the arpes measurements .
the reason for this overestimation is that self - consistency is not imposed in their calculations . in spite of the large anisotropy of @xmath18 , the obtained enhancement ratio of @xmath9 is about two according to ref .
@xcite . therefore ,
this scenario can not account for the large @xmath9 in under - doped systems .
it should be stressed that the magnetoresistance becomes 100 times greater than the experimental value when the anisotropy of @xmath18 is of the order of 100 @xcite .
therefore , the highly anisotropic @xmath18 model is not applicable for optimally - doped htscs . in the next section
, we explain the various anomalous transport phenomena in htscs `` all together '' , by considering the cvc . in the rta where @xmath286 is assumed
, the deviation from the equilibrium distribution function @xmath287 dissipates with time since @xmath288 .
this oversimplification frequently leads to serious unphysical results .
for example , in the absence of electric field , the rta predicts that @xmath287 always vanishes when @xmath289 since @xmath290 at finite temperatures .
however , @xmath287 should remain finite when the umklapp scattering process is absent , because of the momentum conservation laws @xcite .
therefore , the rta violates the conservation laws . to satisfy the conservation laws , the incoming scattering from the other states to @xmath21 has to be taken into account by solving eq .
( [ eqn : boltz2 ] ) .
this scattering process is represented by the cvc in the microscopic fermi liquid theory . here
, we intuitively discuss the cvc in nearly af metals , where the quasiparticles are scattered by strong af fluctuations with @xmath291 .
then , the momentum transfer @xmath240 in the scattering process is restricted to @xmath292 , as shown in fig .
[ fig : cvc - af ] . according to the momentum conservation law
, the quasiparticle at @xmath30 is scattered to @xmath29 as shown in fig .
[ fig : cvc - af ] ( i ) .
the current due to the quasiparticle at @xmath29 and the hole at @xmath30 is given by @xmath293 .
this current almost cancels after performing the @xmath30-summation : in fact , the current of the particle - hole pair ( @xmath294 , @xmath295 ) for @xmath296 , which is shown in fig .
[ fig : cvc - af ] ( ii ) , is @xmath297 since @xmath298 is a reciprocal lattice vector .
therefore , the cvc is given only by the quasiparticle at @xmath28 ; @xmath299 . in this case
, the conductivity does not diverge because of the existence of the umklapp processes , e.g. , the process in fig .
[ fig : cvc - af ] ( ii ) . in htscs .
* ( i ) * due to strong af fluctuations , this quasiparticle at @xmath21 is scattered to @xmath28 ( @xmath300 ) , creating a particle - hole pair ( @xmath30 , @xmath29 ) . *
( ii ) * the current of the particle - hole excitation in ( i ) is almost canceled by another particle - hole excitation ( @xmath295 , @xmath294 ) for @xmath296 .
therefore , the cvc is given by the quasiparticle at @xmath301 alone .
the process ( ii ) is the umklapp scattering . ] in nearly af metals , the total current @xmath19 is approximately parallel to @xmath302 , which implies that @xmath19 is not perpendicular to the fermi surface .
this fact is the origin of the enhancement of @xmath9 @xcite .
the anomalous @xmath21-dependence of @xmath19 becomes much more prominent near the af qcp due to the multiple backscattering of the quasiparticle between that at @xmath21 and that at @xmath301 .
[ the schematic behavior of @xmath19 is shown in fig .
[ fig : j ] . ]
thus far , we discussed only the two - body scattering process , and ignored the higher - order processes .
this simplicity is justified in good metals where @xmath303 , which is satisfied in htscs except for heavily under - doped compounds @xcite . here
, we calculate the cvc based on the microscopic fermi liquid theory . in principle
, we can also calculate the cvc based on the bloch - boltzmann theory , by analyzing eq .
( [ eqn : boltz2 ] ) using the variational principle @xcite .
however , a systematic calculation of the cvc is very difficult , particularly in the presence of a magnetic field .
therefore , we analyze the cvc based on the linear response theory since we can utilize the powerful field theoretical techniques . in the linear response theory @xcite ,
the conductivity is given by @xmath304 , \label{eqn : nakano - kubo}\end{aligned}\ ] ] where @xmath305 is the retarded current - current correlation function . here , we consider the conductivities in the presence of the uniform magnetic field .
then , the hopping integral between site @xmath306 and @xmath307 exhibits the peierls phase : @xmath308 , \label{eqn : peierls}\end{aligned}\ ] ] where @xmath309 is the original hopping integral , and @xmath310 is the vector potential at site @xmath311 .
then , the hamiltonian and the velocity operator under the magnetic field , @xmath312 and @xmath313 respectively , are given by @xmath314\end{aligned}\ ] ] here , we assume that the vector potential is given by @xmath315 , and take the limit @xmath316 at the final stage of the calculation @xcite .
( the magnetic field is @xmath317 . ) then , eq .
( [ eqn : peierls ] ) can be expanded as @xmath318 + o(a^2)$ ] .
after the fourier transformation , both the hamiltonian and the velocity operator of the order of @xmath319 are given by @xcite @xmath320 where @xmath321 , and @xmath322 .
here , @xmath323 , @xmath324 , and @xmath261 is the electron charge .
therefore , @xmath305 is given by the analytic continuation of the following function for @xmath325 @xcite : @xmath326 where @xmath327 is the matsubara frequency ; here we promise @xmath328 represents an integer and @xmath11 is a half - integer , respectively . hereafter , we ignore the spin indices to simplify expressions . according to eq .
( [ eqn : km1 ] ) , @xmath329 without the magnetic field is given by @xcite @xmath330 where @xmath331 .
@xmath332 is the green function where @xmath11 is a half - integer , and @xmath333 is the two - particle green function in eq .
( [ eqn : l ] ) .
@xmath334 is the velocity of the free electron , and @xmath335 is the three - point vertex . in the same way , kohno and yamada @xcite derived the @xmath336-linear term of @xmath337 from eq .
( [ eqn : km2 ] ) as @xmath338 \left [ \lambda_{\k x}^{n;l } \cdot \d_\nu \lambda_{\k y}^{n;l } \right ] \cdot \e_{\mu\nu z } \nonumber \\ & & + [ \mbox{6 point vc term } ] , \label{eqn : sxy - d}\end{aligned}\ ] ] where @xmath339 is an antisymmetric tensor with @xmath340 . in deriving eq .
( [ eqn : sxy - d ] ) , we used the relation @xmath341 and took the limit @xmath316 at the final stage of calculation @xcite .
several kinds of ward identities have to be correctly applied to maintain the gauge invariance @xcite .
( or @xmath342 ) . here , we put i m @xmath343 . from each region ,
@xmath344 and @xmath345 are analytically continued to become @xmath346 and @xmath347 ( @xmath306=1 - 3 ) , respectively . ] in order to derive the conductivity @xmath348 , we have to perform the analytic continuations of eqs .
( [ eqn : sig - d ] ) and ( [ eqn : sxy - d ] ) ; @xmath349 .
then , the @xmath350 summation is replaced with the integrations along the three cut lines in the complex plane in fig .
[ fig : cut ] , together with the thermal factor @xmath351 . in region 1
, @xmath344 becomes @xmath352 . in the same way , it becomes @xmath353 and @xmath354 in region 2 and 3 , respectively .
here , we stress the relation @xmath355 , see eqs .
( [ eqn : qprep1])-([eqn : qprep3 ] ) . as a result , @xmath356 is given by @xcite @xmath357 , \label{eqn : kxx - realw } \\
k_{xx}^{(i)}(\e;\w)&= & \sum_\k v_{\k x}^0 g_\k^{(i)}(\e;\w ) j_{\k x}^{(i)}(\e;\w ) , \end{aligned}\ ] ] where @xmath358 .
@xmath347 is given by the analytic continuation of @xmath345 from region @xmath306 .
it is expressed as @xmath359 where @xmath360 is given by the analytic continuation of the full four - point vertex @xmath361 in eq .
( [ eqn : l ] ) from region @xmath89 for @xmath362 . thermal factors [ such as @xmath363 and @xmath364 are included in the definition of @xmath365 .
the explicit expression for @xmath365 is given in eq .
( 12 ) of ref .
@xcite . the dc - conductivity @xmath366 is given by @xmath367 .
for example , if we take the @xmath5-derivative of the thermal factor @xmath368 associated with @xmath369 in eq .
( [ eqn : kxx - realw ] ) , we obtain @xmath370 where @xmath371 ; note that @xmath372 at sufficiently low temperatures . since @xmath373 is not singular with respect to @xmath374 , eq .
( [ eqn : analytic - cont ] ) is proportional to @xmath374 , which diverges when @xmath375 . on the other hand , if we take the @xmath5-derivative of the thermal factor in front of @xmath376 in eq .
( [ eqn : kxx - realw ] ) , the obtained term is @xmath377 . to derive the exact expression for @xmath366 with respect to @xmath378
, we also have to take the @xmath5-derivative of the thermal factor in front of @xmath369 in @xmath379 . as a result , the exact expression for @xmath366 is given by eq .
( [ eqn : analytic - cont ] ) by replacing @xmath380 with @xmath381 @xcite .
@xmath382 is called the total current since it contains the cvc discussed in previous sections .
hereafter , we denote @xmath383 and @xmath384 as @xmath385 and @xmath386 .
the analytic continuation of @xmath337 had been performed in refs .
@xcite . to summarize ,
the general expressions for @xmath366 and @xmath387 , which are exact within the most divergent term with respect to @xmath378 , are given by @xmath388 equation ( [ eqn : sigma_xx ] ) was derived by eliashberg @xcite .
equation ( [ eqn : sigma_xy ] ) was derived by fukuyama et al within the born approximation @xcite , and it was proved to be correct in fermi liquids @xcite .
note that we dropped the [ 6 point vc term ] in eq .
( [ eqn : sxy - d ] ) since they are less singular with respect to @xmath389 @xcite .
in particular , its contribution to @xmath387 vanishes in the flex approximation . in the same way
, the present author has derived the exact expressions for the magneto - conductivity @xmath390 @xcite and the peltier coefficient @xmath391 in the presence of @xmath392 @xcite . apparently , eqs .
( [ eqn : sigma_xx ] ) and ( [ eqn : sigma_xy ] ) become equal to the rta results given in eqs .
( [ eqn : sxx - rta ] ) and ( [ eqn : sxy - rta ] ) if we replace the total current @xmath393 with the quasiparticle velocity @xmath394 , that is , if we drop the cvc .
the rta had been frequently used in analyzing transport anomaly in high-@xmath1 cuprates .
however , we will explain that the neglect of the cvc frequently causes various unphysical results . according to eliashberg
, the total current @xmath395 can be rewritten as @xcite @xmath396 where @xmath397 , and we put @xmath398 in @xmath385 for simplicity .
@xmath399 is the `` irreducible '' vertex with respect to @xmath400 , which is given by @xmath401 equation ( [ eqn : bs - t ] ) is expressed in fig .
[ fig : bs - t ] ( i ) .
the cvc in the microscopic fermi liquid theory is also called the backflow in the phenomenological fermi liquid theory . ) ) for the total current @xmath19 . *
( ii ) * expression for @xmath399 in the microscopic fermi liquid theory .
the factor @xmath402 is necessary to avoid double counting . ] here , we analyze the bethe - salpeter equation ( [ eqn : bs - t ] ) : the solution of eq .
( [ eqn : bs - t ] ) is real since @xmath399 in eq .
( [ eqn : t22 ] ) is pure imaginary . since im@xmath403
, we have to extract the @xmath342-linear term of im@xmath404 , which is derived from the cut of the particle - hole pair or that of the particle - particle pair in @xmath404 @xcite . by this procedure , @xmath405 is divided into two @xmath361 s .
therefore , @xmath342-linear term of im@xmath406 is given by @xcite @xmath407 where @xmath408 is the full four - point vertex at the fermi level , which is a real function of ( @xmath409 ) .
the factor @xmath410 in front of the second term in the curly bracket in eq .
( [ eqn : img ] ) is necessary to avoid double counting .
then , @xmath399 is given by @xmath411 2{\rm im}\gamma(0,\e'-i\delta)$ ] ; it is schematically shown in fig .
[ fig : bs - t ] ( ii ) . after changing momentum variables ,
the bethe - salpeter equation ( [ eqn : bs - t ] ) is transformed into @xcite @xmath412 where we used the relation @xmath238\e=(\pi t)^2 $ ] .
@xmath413 ( @xmath414 ) are functions of ( @xmath409 ) , and they represent the forward scattering amplitude @xcite .
the expressions for @xmath413 ( @xmath414 ) at sufficiently low temperatures are given by @xcite @xmath415 since the cvc due to @xmath416 represents the hole current , the minus sign appears in eq .
( [ eqn : vccc ] ) . in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) , we dropped spin indices in @xmath361 to simplify equations : if spin indices are taken into account , @xmath417 in each equation is replaced with @xmath418 as explained in ref .
@xcite . equations ( [ eqn : vcaa])-([eqn : vccc ] ) are expressed in fig .
[ fig : cvc - diag ] ( i ) .
note that eqs .
( [ eqn : vcaa ] ) and ( [ eqn : vcbb ] ) are equivalent since @xmath408 is full antisymmetrized as a consequence of the pauli principle . in the same way ,
im@xmath419 at sufficiently low temperatures can be expressed as @xcite @xmath420 which is proportional to @xmath8 at the zero - temperature limit , if the dimension is slightly higher than two .
[ in pure 2d systems , @xmath421 . ] equation ( [ eqn : gamma ] ) is shown in fig .
[ fig : cvc - diag ] ( ii ) .
we note that @xmath181 is also given by @xmath422 . )
-([eqn : vccc ] ) .
each hatched part represents the real part of the full antisymmetrized four point vertex @xmath423 .
each line with arrow represents the imaginary part of the green function .
note that ( a ) becomes different from ( b ) if we violate the antisymmetric nature of @xmath423 in the course of approximation . *
( ii ) * diagrammatic representation of eq . ( [ eqn : gamma ] ) . ] here , we calculate the cvc in a free dispersion model in the absence of umklapp scattering according to yamada and yosida @xcite . in this case , both @xmath181 and @xmath19 on the fermi surface are isotropic , that is , @xmath424 and @xmath425 . by noticing the relationship @xmath426 and using eqs .
( [ eqn : vcaa])-([eqn : gamma ] ) , it is easy to verify that the cvc in eq .
( [ eqn : bs - abc2 ] ) is exactly given by @xmath427 in a free dispersion model .
if we put @xmath428 ( @xmath282 is a constant ) , the solution of eq .
( [ eqn : bs - abc ] ) is given by @xmath429 , which diverges when @xmath430 .
therefore , the conductivity given by eq .
( [ eqn : sigma_xx ] ) diverges , which is a natural consequence of the momentum conservation of the system @xcite . yamada and
yosida also showed that @xmath393 remains finite in the presence of the umklapp processes @xcite .
next , we dicuss the case where @xmath126 is anisotropic . then , the formal solution of eq .
( [ eqn : bs - t ] ) is given by @xmath431 where @xmath432 .
when @xmath126 is anisotropic , quasiparticle current is not conserved in the normal scattering process ; @xmath433 .
nonetheless of the fact , one can show that det@xmath434 as a result of the momentum conservation @xcite .
therefore , @xmath19 given in eq .
( [ eqn : j - sol ] ) and @xmath366 diverge in any anisotropic model if the umklapp processes are absent .
[ exactly speaking , @xmath435 is proportional to @xmath436 when the @xmath437-particle umklapp scattering processes are present @xcite . ]
we stress that the flex approximation can reproduce the divergence of the conductivity in the absence of the umklapp scatterings , if one consider the cvc correctly .
this result is assured by the fact that the ward identity @xmath141 is consistently satisfied in the flex approximation @xcite .
therefore , the flex approximation will be appropriate for the study of transport phenomena in correlated electron systems . ) .
the wavy lines represent the spin fluctuations , @xmath438 . *
( ii ) * diagrammatic representation of @xmath439 in eq .
( [ eqn : g - propto - chi ] ) .
@xmath23 is the bare coulomb interaction in the hubbard model and @xmath440 is the three - point vertex that is irreducible with respect to @xmath23 . ]
next , we consider the role of the cvc in nearly af fermi liquids in the presence of umklapp scatterings , where the electron - electron scatterings are mainly given by the strong af fluctuations with @xmath441 . in this case , we can approximate the full four - point vertex in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) as @xmath442 which satisfies the antisymmetric nature of @xmath361 that is the consequence of the pauli principle . the diagrammatic representation of eq .
( [ eqn : g - propto - chi ] ) is shown in fig .
[ fig : ubar ] ( i ) . here , @xmath443 ( @xmath444 ) represents the spin index and @xmath445 is the pauli matrix vector .
@xmath439 in eq .
( [ eqn : g - propto - chi ] ) is the effective interaction between the electrons and spin fluctuations , which is the renormalized coulomb interaction due to the irreducible three - point vertex shown in fig .
[ fig : ubar ] ( ii ) . in the present subsection , however , we analyze the cvc using the following more simplified approximation for @xmath361 : @xmath446 although eq .
( [ eqn : g - propto - chi - simple ] ) violates the pauli principle , this approximation produces the dominant cvc ( maki - thompson term ) near af - qcp . in
[ beyondone ] , we will analyze the cvc using eq .
( [ eqn : g - propto - chi ] ) , and show that the analysis based on eq .
( [ eqn : g - propto - chi - simple ] ) is justified . according to the present approximation , @xmath413 ( @xmath414 ) in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) are expressed by ( a)-(c ) in fig .
[ fig : cvc - diag - onel ] ( i ) , respectively , where wavy lines represent the spin fluctuations .
the cvcs expressed in ( a ) and ( b ) correspond to the current due to the quasiparticles at @xmath28 and @xmath29 , respectively , and the cvc in ( c ) corresponds to the current due to the quasi - hole at @xmath30 in fig .
[ fig : cvc - af ] . in the field theory , ( a )
is called the maki - thompson term , and ( b ) and ( c ) are called the aslamazov - larkin terms . as explained in fig .
[ fig : cvc - af ] , the aslamazov - larkin terms approximately disappears @xcite whereas the maki - thompson term plays an important role when @xmath447 and @xmath448 . ) , @xmath449 , @xmath450 , and @xmath451 in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) are approximated as ( a)-(c ) , respectively .
( a ) corresponds to the maki - thompson term in the flex approximation , which represents the cvc ( mt ) in fig .
[ fig : cvc - af ] .
( b ) and ( c ) correspond to the aslamazov - larkin terms , which represent the ( al1 ) and ( al2 ) in fig .
[ fig : cvc - af ] . *
( ii ) * the expression for @xmath248 according to the approximation in eq . ( [ eqn : g - propto - chi - simple ] ) . ]
according to eq .
( [ eqn : g - propto - chi - simple ] ) , @xmath417 in @xmath452 becomes @xmath453 ^ 2 $ ] after taking the summation of spin indices .
however , we replace @xmath417 with @xmath454 ^ 2 $ ] since the term ( a ) in fig . [
fig : cvc - diag - onel ] appears twice if we use eq .
( [ eqn : g - propto - chi ] ) .
then , the cvc in eq .
( [ eqn : bs - abc2 ] ) and @xmath248 given in eq .
( [ eqn : gamma ] ) near the af qcp are obtained as @xmath455 ^ 2 { \rm im}\dot{\chi}_q^0(0 ) \frac{\rho_{\k+\q}(0)}{\g_{\k+\q } } { \vec j}_{\k+\q } , \label{eqn : bs - a } \\ \g_{\k}&= & \sum_{\q } ( \pi t)^2 \frac{3{\bar u}^4}{2 } [ \chi_\q^s(0)]^2 { \rm im}\dot{\chi}_q^0(0 ) \rho_{\k+\q}(0 ) , \label{eqn : gamma - a}\end{aligned}\ ] ] where we used the relation @xmath456 . since @xmath457 ^ 2 $ ] , the bethe - salpeter equation ( [ eqn : bs - a ] ) and @xmath181 in eq .
( [ eqn : gamma - a ] ) in a nearly af fermi liquid are given by @xmath458 note that eq .
( [ eqn : gamma - oneloop ] ) is equivalent to @xmath182 in the spin fluctuation theory in eq .
( [ eqn : gam - flex ] ) . in this subsection , we have neglected the energy - dependence of the four - point vertices [ in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) ] and that of the spin susceptibilities [ in eqs .
( [ eqn : bs - a])-([eqn : gamma - oneloop ] ) ] to simplify the discussion . for this reason , eqs .
( [ eqn : bs ] ) and ( [ eqn : gamma - oneloop ] ) are appropriate only for @xmath123 where @xmath5-dependence of @xmath459 can be ignored . in the case of @xmath460 , @xmath461 in eqs .
( [ eqn : bs ] ) and ( [ eqn : gamma - oneloop ] ) should be replaced with @xmath462\e d\e = 4 t \w_{\rm sf}$ ] . for any value of @xmath225 ,
they are expressed by eqs .
( 36 ) and ( 33 ) in ref .
@xcite , respectively . here
, we approximately solve eq .
( [ eqn : bs ] ) in the case of @xmath463 and @xmath464 . here , we assume @xmath21 is close to point a in fig . [
fig : fs - hotcold ] ( i ) . because @xmath438 takes a large value only when @xmath465 , the cvc term in eq .
( [ eqn : bs ] ) can be expressed as @xmath466 , where @xmath467 for @xmath468 , and @xmath469 for @xmath470 as shown in fig .
[ fig : fs - hotcold ] ( i ) .
considering eq .
( [ eqn : gamma - oneloop ] ) , @xmath471 in the case of @xmath463 . the more detailed expression for the cvc term @xmath472
is given by @xcite @xmath473 where @xmath474 and @xmath475 is a constant .
the @xmath21-dependence of @xmath476 will be moderate if @xmath477 . by this simplification , eq .
( [ eqn : bs ] ) becomes @xmath478 and the solution is given by @xcite @xmath479 here , we examine the @xmath21-dependence of @xmath393 given in eq .
( [ eqn : j - ap ] ) : ( a ) near points a and b , @xmath19 is parallel to @xmath20 by symmetry . at point
a , @xmath480 since @xmath481 .
( b ) near point c , @xmath482 , which is nearly parallel to the afbz .
these behaviors of @xmath19 together with @xmath20 are schematically shown in fig .
[ fig : j ] ( i ) . in the present discussion
, we assumed that @xmath483 , where @xmath187 represents the deviation from the nesting condition at the cold spot in fig .
[ fig : fs - hotcold ] ( i ) .
this condition seems to be satisfied in slightly under - doped ybco and lsco above @xmath1 , as discussed in [ hotcold ] . in fig .
[ fig : fs - hotcold ] ( ii ) , we show an `` effective fermi surface '' obtained by bending the true fermi surface such that it is orthogonal to @xmath19 around the cold spot @xcite .
we discuss the role of the cvc in the hall coefficient based on the concept of the effective fermi surface in [ resrh ] . and the quasiparticle velocity @xmath20 on the fermi surface .
the cold spot is located around the point a ( b ) in hole - doped ( electron - doped ) htscs . *
( ii ) * effective fermi surface for @xmath387 . the large curvature of the effective fermi surface around the cold spot is the origin of the enhancement ( and the sign - change ) of @xmath9 . ]
when @xmath251 , one may think that the cvc is small around the cold spot , and therefore the increment in @xmath9 due to the cvc is also small . however , this expectation is not true unless @xmath459 is a step function like @xmath484 .
in fact , we verified that the cvc around the cold spot is significant even if @xmath485 is much larger than unity , by assuming a phenomenological @xmath459 model given in eq .
( [ eqn : kai_qw ] ) @xcite . in later sections
, we demonstrate that @xmath9 is prominently enhanced by the cvc using the flex approximation , which yields a realistic functional form of @xmath459 .
we will show that the cvc is important in ncco , although @xmath251 is realized at low temperatures .
finally , we comment that the aslamazov - larkin terms ( ( b ) and ( c ) in fig . [ fig : cvc - diag - onel ] ) become important when the af fluctuations are weak . in the flex approximation
, one can reproduce the divergence of the conductivity in the absence of the umklapp scatterings , only if all the cvcs ( ( a)-(c ) in fig .
[ fig : cvc - diag - onel ] ) are taken into consideration . in the previous subsection
, we analyzed the cvc by assuming eq .
( [ eqn : g - propto - chi - simple ] ) , which corresponds to the one - loop approximation for the self - energy like the flex approximation . in this approximation ,
( [ eqn : bs - a ] ) and eq .
( [ eqn : gamma - a ] ) contain the same kernel function [ @xmath486 ^ 2 { \rm
im}\dot{\chi}_q^s(0 ) \rho_{\k+\q}(0)$ ] ] as a consequence of the ward identity between the cvc and the imaginary part of the self - energy . for this reason , the factor @xmath476 in eq .
( [ eqn : j - ap ] ) approaches unity near af - qcp , which assures the anomalous @xmath21-dependence of @xmath19 near af - qcp that shown in fig .
[ fig : j ] .
however , one may be afraid that this result is an artifact due to the violation of the pauli principle . in this subsection
, we analyze the self - energy and the cvc beyond the one - loop approximation by assuming eq .
( [ eqn : g - propto - chi ] ) . here , we confirm that the cvc near the af qcp is important even if the higher - loop diagrams are taken into account .
according to eq .
( [ eqn : g - propto - chi ] ) , @xmath181 in eq .
( [ eqn : gamma ] ) is approximately given by @xmath487 where @xmath488 and @xmath489 are shown in fig .
[ fig : cvc - diag - twol ] ( ii ) .
the integrand in eq .
( [ eqn : gamma - two2 ] ) takes a large value when @xmath490 , under the condition that @xmath491 .
when @xmath492 , @xmath493 can be significant at the hot spot since @xmath494 and @xmath495 in modulo @xmath496 . ) , @xmath449 , @xmath450 , and @xmath451 in eqs .
( [ eqn : vcaa])-([eqn : vccc ] ) are approximated as ( a)-(f ) , respectively .
( a ) , ( b ) , ( d ) , and ( e ) come from @xmath452 and @xmath497 .
( c ) and ( f ) come from @xmath416 . *
( ii ) * the expression for @xmath248 according to the approximation in eq . ( [ eqn : g - propto - chi ] ) . ] in the same way , the cvc in eq . ( [ eqn : bs - abc2 ] ) is given by the six diagrams shown in fig .
[ fig : cvc - diag - twol ] . as discussed in [ cvc - oneloop ] , terms ( b ) and ( c ) are negligible in nearly af fermi liquids .
other terms are given by @xmath498 here , @xmath499 takes a large value only when @xmath490 . in ( d)-(f ) , the set of momenta ( @xmath409 ) is changed to ( @xmath500 ) due to the transformation of variables . applying the same approximation that was used in eq .
( [ eqn : jdash ] ) , the total cvc is given by @xmath501 where @xmath502 , @xmath503 and @xmath504 are positive , and they approach zero in proportion to @xmath505 near the af - qcp . then , the solution of the bethe - salpeter equation @xmath506 is given by @xmath507 [ \g_\k+\a_\k\g_\k^{(1)}+(2\a_\k'+\a_\k'')\g_\k^{(2 ) } ] } \nonumber \\ & & \times \left\ { { \vec v}_\k+\frac{\a_\k\g_\k^{(1)}+2\a_\k'\g_\k^{(2 ) } } { \g_\k+\a_\k''\g_\k^{(2 ) } } { \vec v}_{\k^ * } \right\ } .
\label{eqn : j - exact3}\end{aligned}\ ] ] when @xmath447 , eq .
( [ eqn : j - exact3 ] ) is approximately given by @xmath508 } \left\ { { \vec v}_\k+{\vec v}_{\k^ * } \right\ } .
\label{eqn : j - exact3 - 2}\end{aligned}\ ] ] near the af - qcp , @xmath509 since @xmath510 .
therefore , @xmath393 in eq.([eqn : j - exact3 - 2 ] ) shows a similar behavior to that given in eq .
( [ eqn : j - ap ] ) . as a result
, the significance of the cvc in nearly af fermi liquid is confirmed even if we assume eq .
( [ eqn : g - propto - chi ] ) , which satisfy the pauli principle . in eq .
( [ eqn : g - anti - anyloop ] ) . ]
finally , we discuss the cvc composed of the higher loop terms with respect to @xmath438 , by considering the following approximation for the full four - point vertex : @xmath511 by considering that @xmath361 mainly depends on the momentum transfer in the electron - electron scattering .
@xmath512 is given by fig .
[ fig : g - anyloop ] . using the full four - point vertex ,
the spin susceptibility is given by @xmath513 , where @xmath514 .
according to the approximation in eq .
( [ eqn : g - anti - anyloop ] ) , @xmath438 is given by @xmath515 ^ 2 { \bar \gamma}(\q ) \nonumber \\ - 2t^2\sum_{\k,\k',\e,\e ' } g_\k(\e)g_{\k+\q}(\e ) { \bar \gamma}(\k-\k ' ) g_{\k'+\q}(\e')g_{\k'}(\e ' ) \label{eqn : chi - anyloop } .\end{aligned}\ ] ] the third term of eq .
( [ eqn : chi - anyloop ] ) is expected to be smaller than the second term since the integrand takes large values only when @xmath516 .
since the @xmath240-dependence of @xmath517 is moderate , the present analysis suggests that the approximate relation @xmath518 holds for @xmath519 .
for this reason , the solution of @xmath19 obtained in this subsection , eq .
( [ eqn : j - exact3 ] ) or ( [ eqn : j - exact3 - 2 ] ) , will be valid even in higher - loop approximations . as a result , the anomalous @xmath21-dependence of @xmath393 in fig .
[ fig : j ] , which is derived in the previous section within the one - loop approximation , is expected to be realized even in the close vicinity of the af qcp where the higher - loop diagrams becomes important . in later sections ,
we present a numerical study of the transport phenomena based on the flex+cvc approximation .
although it is a one - loop approximation with respect to the spin fluctuation , the obtained numerical results will be qualitatively reliable even in the under - doped region , since the cvc due to the higher - loop diagrams do not alter the one - loop approximation results qualitatively .
here , we discuss the transport phenomena in htscs above the pseudo - gap temperature @xmath34 where the sc fluctuations are negligibly small . in the flex+cvc approximation ,
the self - energy is given by eq .
( [ eqn : self ] ) , and the vertex function in the bethe - salpeter equation , @xmath520 , is given by eq .
( [ eqn : t22 ] ) .
the total current @xmath19 is obtained by solving the bethe - salpeter equation @xcite . according to the ward identity ,
the irreducible four - point vertex is given by @xmath141 .
in the flex approximation , the self - energy is given by the convolution of @xmath521 and @xmath522 as in eq .
( [ eqn : self ] ) .
then , @xmath523 contains one maki - thompson term that is given by taking derivative of @xmath521 , and two aslamazov - larkin terms that are given by taking derivative of @xmath522 , which is composed of infinite @xmath521 s .
these three terms corresponds to ( a)-(c ) in fig .
[ fig : cvc - diag - onel ] @xcite .
considering the relation @xmath524 , we can derive @xmath525 in the flex approximation as @xmath526 \nonumber \\ & & \times g_{\k'-\q}(\e_{n'}-\w_l)g_{\k-\q}(\e_{n}-\w_l ) \nonumber \\ & - & t\sum_{\q , l}u^2\left[\frac32 ( u\chi_\q^s(\w_l)+1)^2 + \frac12 ( u\chi_\q^c(\w_l)-1)^2 -1 \right ] \nonumber \\ & & \times g_{\k'+\q}(\e_{n'}+\w_l)g_{\k-\q}(\e_{n}-\w_l ) \label{eqn : ivc - flex } .\end{aligned}\ ] ] the first term @xmath527 corresponds to the maki - thompson vertex correction .
the last two terms in eq .
( [ eqn : ivc - flex ] ) , both of which contain @xmath528 ^ 2 $ ] , are the aslamazov - larkin vertex corrections .
hereafter , we drop the aslamazov - larkin terms since they are negligible in nearly af fermi liquids ( @xmath529 ) , as we have discussed in previous sections . after the analytic continuation of the maki - thompson term @xmath530 from the region @xmath531 and
@xmath532 , the bethe - salpeter equation in the flex+cvc theory for @xmath533 is given by @xcite , @xmath534 \nonumber \\ & & \ \ \ \ \ \ \ \ \ \ \ \ \times { \rm im}v_{\q}(\e'+\i\delta ) \cdot |g_{\k+\q}(\e+\e'+i\delta)|^2 \cdot { \vec j}_{\k+\q}(\e+\e ' ) \nonumber \\ & \approx&{\vec v}_{\k } + \sum_{\q } \frac{(\pi t)^2}{2 } { \rm im}\dot{v}_{\q}(0 ) \frac{\rho_{\k+\q}(0)}{\g_{\k+\q}}{\vec j}_{\k+\q } , \label{eqn : j_numerical } \end{aligned}\ ] ] where @xmath535 . in deriving eq .
( [ eqn : j_numerical ] ) , we used the relation @xmath536\e d\e= ( \pi t)^2 $ ] , and assumed the relation @xmath235 . if we approximate @xmath537 , eq .
( [ eqn : j_numerical ] ) is equivalent to eq .
( [ eqn : bs ] ) .
[ note that @xmath538 in the flex approximation . ] the numerical solution of eq .
( [ eqn : j_numerical ] ) is shown in fig .
[ fig : numerical - j ] , whose schematic behavior is shown in fig .
[ fig : j ] ( i ) .
this singular @xmath21-dependence of @xmath539 is realized because @xmath476 in eq .
( [ eqn : jdash ] ) approaches unity for @xmath540 . as discussed in [ beyondone ] , this result is not specific to the flex approximation since the same vertex function [ @xmath520 ] appears in both ( [ eqn : bs - t ] ) and ( [ eqn : gamma ] ) in the microscopic fermi liquid theory , which is one of the ward identity . for this reason , singular @xmath21-dependence of @xmath539 in the case of @xmath447 , which is shown in fig .
[ fig : numerical - j ] , is not specific to the flex approximation , but is a universal behavior in fermi liquids near the af qcp . and
* ( ii ) * mean free path with cvc obtained by eq .
( [ eqn : j_numerical ] ) ; @xmath542 .
@xmath543 at the cold spot ( a ) increases as @xmath15 decreases .
@xmath544 attains its local minimum at the cold spot for @xmath545 , which induces increases in the nernst coefficient and magnetoresistance .
we also plot the mean free path without cvc ; @xmath546 . ] in hole - doped systems , @xmath181 attains its minimum ( maximum ) at point a ( point b ) as shown in fig .
[ fig : del ] @xcite . at low @xmath15 ,
the local minimum of @xmath547 is located at points a and b , and its maximum is located at point c : this is schematically shown in fig .
[ fig : j ] ( i ) .
further , fig .
[ fig : numerical - j ] ( ii ) shows that the mean free path with cvc , @xmath542 , attains its local minimum at point a for @xmath545 , since @xmath547 rapidly increases as @xmath21 deviates from the point a. in fact , eq .
( [ eqn : j - ap ] ) suggests that @xmath548 .
therefore , @xmath547 takes a large value near the af qcp except for point a and b. in later sections , we explain that the a local minimum structure of @xmath544 at point a becomes prominent below @xmath34 , which causes the significant increases in the nernst coefficient and magnetoresistance . in ref .
@xcite , we solved the bethe - salpeter equation by considering both maki - thompson term and the aslamazov - larkin terms , and calculated both @xmath6 and @xmath9 . as a result
, we verified numerically that the aslamazov - larkin terms are negligible in nearly af metals .
first , we analyze the conductivity @xmath366 using eq .
( [ eqn : sigma_xx ] ) . according to the approximate expression for @xmath19 given in eq .
( [ eqn : j - ap ] ) , @xmath549_{\rm cold } = \left [ \frac{{v}_{\k x}^2}{1+\a_\k } \right]_{\rm cold}\end{aligned}\ ] ] at the cold spot of ybco [ point a ] .
we have used the relationship @xmath550 at the cold spot . since @xmath551 when @xmath447 , then @xmath552 .
therefore , the resistivity @xmath6 is slightly increased by the cvc @xcite . since the effect of the cvc on @xmath6 is not large , the rta result for @xmath6 can be qualitatively justified .
figure [ fig : rho ] ( i ) shows the numerical results of @xmath6 for lsco ( @xmath553 and @xmath554 ) obtained from the flex approximation and the flex+@xmath15-matrix approximation . in the flex approximation without cvc
, @xmath6 shows an approximate @xmath15-linear behavior for @xmath555 ( @xmath556 k ) . because of the cvc
, @xmath6 is slightly enhanced for a wide range of temperatures .
moreover , @xmath6 obtained by the flex+cvc approximation shows a tiny `` kink '' structure at @xmath557 , below which af fluctuations grows prominently .
this result is consistent with the experimental results @xcite .
the kink becomes more prominent in the flex+@xmath15-matrix approximation , which will be discussed in
[ nernst ] . in optimally - doped ybco and lsco , the resistivity at 300 k is @xmath558 cm ,
and it increases to @xmath559 cm in slightly under - doped compounds ( @xmath560 ) @xcite . in fig .
[ fig : rho ] ( i ) , @xmath561 @xmath562 cm at @xmath563 @xmath564 k : the resistivity increases with @xmath23 , and @xmath565 cm for ybco with @xmath566 and @xmath189 @xcite , which is consistent with experimental values .
however , the flex+cvc method can not reproduce the huge resistivity in heavily under - doped compounds , since the self - consistency condition in the flex approximation tends to suppress @xmath181 too strongly ; see [ validityflex ] . in under - doped compounds
, one has to take account of the effect of residual disorder since the residual resistivity grows prominently .
we will explain the reason in [ imp - strong ] based on the fermi liquid theory with strong af fluctuations .
figure [ fig : rho ] ( ii ) shows the @xmath6 for ncco ( @xmath567 ) obtained from the flex approximation ( within the rta ) and the flex+cvc approximation . in the case of @xmath153 ( over - doped )
, @xmath6 shows a @xmath8-like behavior below @xmath137 @xmath568 k , which is consistent with experiments . in the case of @xmath569 ( under - doped )
, @xmath6 shows an approximate @xmath15-linear behavior when the cvc is included in the calculation .
interestingly , @xmath570 for @xmath569 below @xmath571 , since @xmath572 due to the cvc in ncco at low temperatures . for lsco obtained by the flex ( without cvc ) and flex+cvc approximations .
@xmath6 is slightly enhanced because of the cvc .
we also plot @xmath6 obtained by the flex+@xmath15-matrix ( without cvc ) and [ flex+@xmath15-matrix]+cvc approximations .
the flex+@xmath15-matrix approximation coincides with the flex approximation for @xmath573 , where sc fluctuations disappear .
@xmath574 corresponds to 800 k. * ( ii ) * @xmath6 for ncco obtained by the flex ( without cvc ) and flex+cvc approximations . ]
next , we discuss the hall coefficient . using the onsager s
relation @xmath575 , the general expression for @xmath576 in eq .
( [ eqn : sigma_xy ] ) and @xmath577 in eq .
( [ eqn : sxy - rta ] ) in 2d systems can be rewritten as @xcite @xmath578 where @xmath579 , @xmath580 and @xmath581 .
@xmath582 is the momentum tangent to the fermi surface , which is depicted in fig .
[ fig : j ] ( i ) .
note that @xmath542 represents the mean free path by considering the cvc . in deriving above equations
, we used the relation @xmath583 in 2d systems . using the relation @xmath584 , @xmath387 in a 3d system
is given by @xmath585 where @xmath586 represents the fermi surface element , and @xmath582 is the momentum tangent to the fermi surface and parallel to the @xmath587-plane .
we stress that the @xmath588-derivative of @xmath181 does not enter into the expression of @xmath387 , while it exists in expressions of @xmath589 and @xmath590 . in the expression of @xmath591
, @xmath592 represents the curvature of the fermi surface @xcite ; in both hole - doped and electron - doped htscs , @xmath593 on the fermi surface . on the other hand ,
@xmath594 exhibits strong @xmath21-dependence in nearly af metals , as shown in fig .
[ fig : j ] ( i ) ; @xmath595 is positive around point a , whereas it is negative around point b. since the cold spot is located around point a in hole - doped systems @xcite , the present study predicts that @xmath596 in hole - doped systems when the af fluctuations are strong .
[ note that the charge of electron is @xmath597 @xmath598 in the present study . ] on the other hand , @xmath599 in electron - doped systems since the cold spot in electron - doped systems is b @xcite . for an intuitive understanding of the cvc
, we introduce an `` effective fermi surface '' obtained by bending the true fermi surface such that it is orthogonal to @xmath19 around the cold spot @xcite . the effective fermi surface in htscs is shown in fig . [ fig : j ] ( ii ) .
it can be seen that the curvature of the effective fermi surface , which is equal to @xmath594 by definition , takes a large positive ( negative ) value around the cold spot in hole - doped ( electron - doped ) systems .
therefore , @xmath9 in hole - doped ( electron - doped ) systems exhibits a large positive ( negative ) value at low temperatures @xcite .
figures [ fig : rh ] ( i ) , ( iii ) , and ( iv ) show the numerical results for @xmath9 obtained from the flex approximation by including the cvc . in hole - doped systems ( @xmath600 ) , @xmath9 increases as the doping @xmath601 decreases .
on the other hand , @xmath9 for electron - doped systems ( @xmath139 ) becomes negative below @xmath602 @xmath603400 k. these results are consistent with the experimental results .
previously , negative @xmath9 in electron - doped systems had been considered to be strong evidence of the breakdown of the fermi liquid state since the curvature of the fermi surface in htscs is positive everywhere .
however , it is naturally explained in terms of the fermi liquid picture since the curvature of the effective fermi surface becomes negative around point b , as shown in fig .
[ fig : j ] ( ii ) @xcite . as shown in fig .
[ fig : rh ] ( i ) , ( iii ) , and ( iv ) , @xmath9 given by the rta ( without cvc ) decreases with decreasing @xmath15 , since the curvature of the true interacting fermi surface becomes smaller in the flex approximation at low temperatures @xcite : the deformation of the fermi surface is caused by the strong @xmath21-dependence of the self - energy , which becomes prominent when af fluctuations are strong . a tiny increment in @xmath283 in ybco below @xmath152 is caused by the strong anisotropy of @xmath248 . in the present calculation ,
@xmath604 is approximately 3 at @xmath152 figure [ fig : rh ] ( ii ) shows the @xmath605-dependence of @xmath606 and @xmath607 , where @xmath608 and @xmath609 . in both @xmath366 and @xmath387 ,
the quasiparticles around the cold spot gives the dominant contributions .
we can see that @xmath366 is slightly reduced by including the cvc . on the other hand
, @xmath387 obtained from the flex+cvc theory is considerably larger than @xmath591 , since @xmath610 at the cold spot .
@xmath611 is large only around the cold spot , and it becomes negative around the hot spot . for lsco obtained by the flex+cvc approximation .
in contrast to its prominent @xmath15-dependence , @xmath283 given by the rta ( without cvc ) shows only a weak @xmath15-dependence . @xmath9
obtained using the flex+@xmath15-matrix approximation is also shown ; it starts to decrease below @xmath612 k ( pseudo - gap behavior ) . *
( ii ) * @xmath605-dependence of @xmath613 and @xmath611 at @xmath152 . note that @xmath614 . *
( iii ) * @xmath9 for ybco obtained using the flex+cvc method . *
( iv ) * @xmath9 for ncco obtained using the flex+cvc method .
] here , we discuss the temperature dependence of @xmath9 in detail using the approximate expression for @xmath19 in eq .
( [ eqn : j - ap ] ) . since @xmath615 at the cold spot
[ point a ] because of the symmetry , @xmath616 at the cold spot . by noticing the relationships @xmath617 and @xmath618
, we obtain that @xmath619 which is proportional to @xmath620 @xcite .
in fact , fig . [
fig : numerical - j ] ( i ) shows that eq .
( [ eqn : dtdk ] ) increases as the temperature decreases . as a result
, @xmath9 behaves as @xcite @xmath621 therefore , both the sign and the @xmath15-dependence of @xmath9 in hole - doped htsc are successfully reproduced in the present approach .
finally , we discuss electron - doped systems . at the point b in fig .
[ fig : fs - hotcold ] ( i ) , @xmath622 is equal to @xmath623 for @xmath624 . then , the cvc at @xmath625 , which is given by the second term of eq .
( [ eqn : bs ] ) , is approximately proportional to @xmath626 .
therefore , @xmath627 in the simplified bethe - salpeter equation ( [ eqn : bs - ap ] ) and @xmath628 at point b. since @xmath476 rapidly increases if @xmath21 deviates from point b , @xmath629 attains a large negative value around the cold spot of ncco .
therefore , the negative sign of @xmath9 is realized by considering the cvc .
finally , we stress that we have also studied the cvc by using a widely used phenomenological af fluctuation model given in eq .
( [ eqn : kai_qw ] ) . assuming a reasonable set of parameters , we find that @xmath9 is prominently enhanced due to the cvc @xcite .
thus , the enhancement of @xmath9 in nearly af metals is not an artifact in the flex approximation , but a universal phenomena near the af qcp .
next , we study the orbital magnetoresistance @xmath12 in htscs by considering the cvc .
according to the linear response theory @xcite , the magnetoresistance is give by @xmath630 where @xmath631 denotes the conductivity without a magnetic field , and @xmath632 is the magneto - conductivity , which is always negative . to derive the magnetoresistance , we have to calculate the magneto - conductivity , which is given in eq .
( [ eqn : dsxx - rta ] ) in the rta . by performing the partial integration ,
( [ eqn : dsxx - rta ] ) for 2d system is rewritten as @xmath633 \frac1{\g_\k } , \label{eqn : mcrta}\end{aligned}\ ] ] where @xmath634 . here , we used the relation @xmath583 in 2d systems . when the fermi surface is spherical , the second term in the bracket in eq .
( [ eqn : mcrta ] ) vanishes identically . in this case ,
@xmath635 , @xmath636 and @xmath637 according to the rta , where @xmath638 and @xmath639 are the fermi momentum and the fermi velocity , respectively
. therefore , @xmath12 given in eq .
( [ eqn : mr ] ) becomes zero . except for this special case ,
the orbital magnetoresistance is always positive @xcite .
the general expression for the magneto - conductivity in fermi liquids are derived in ref .
@xcite , which is exact of order @xmath640 .
it is derived by performing the analytic continuation of eq .
( [ eqn : km ] ) for @xmath641 . this work had enabled us to calculate the magneto - conductivity along with satisfying the conservation laws . at low temperatures ,
the magneto - conductivity can be expressed by a simple form : @xmath642 ; @xmath643 where @xmath644 and @xmath645 is the irreducible four - point vertex in the particle - hole channel introduced in eq .
( [ eqn : bs - t ] ) @xcite .
@xmath646 in eq .
( [ eqn : dsxx - rta ] ) is given by eqs .
( [ eqn : mca ] ) and ( [ eqn : d ] ) , by replacing @xmath19 with @xmath20 . in the flex approximation ,
@xmath647 , as shown in eq .
( [ eqn : bs ] ) . here
, we can rewrite @xmath648 as @xmath649 \frac1{\g_\k } , \label{eqn : mca2}\end{aligned}\ ] ] where @xmath644 . according to eq .
( [ eqn : dtdk ] ) , the first term in the bracket in eq .
( [ eqn : mca2 ] ) is proportional to @xmath650 @xcite .
further , remaining terms ( the second term in eq .
( [ eqn : mca2 ] ) and @xmath651 ) are also proportional to @xmath652 for a wide range of temperatures , since the @xmath653-derivative of @xmath654 yields a factor proportional to @xmath111 @xcite . in
[ nernst ] , we will explain that @xmath544 becomes highly anisotropic below @xmath34 due to af+sc fluctuations .
for this reason , due to the second term in eq .
( [ eqn : mca2 ] ) , @xmath12 is prominently enhanced below @xmath34 in under - doped htscs .
therefore , the magnetoresistance in nearly af metals behaves as @xmath655 above @xmath34 due to the cvc since @xmath656 and @xmath224 .
equation ( [ eqn : mr - scaling ] ) is consistent with experimental results in lsco @xcite , ybco @xcite and tbco @xcite . on the other hand , in a conventional fermi liquid where @xmath108 is constant , eq .
( [ eqn : mr - scaling ] ) gives kohler s rule @xmath56 .
equations ( [ eqn : rh - scaling ] ) and ( [ eqn : mr - scaling ] ) directly imply the `` modified kohler s rule '' @xcite @xmath657 where @xmath658 .
experimentally , the modified kohler s rule is well satisfied in various htsc compounds @xcite and in ce@xmath2in@xmath3 ( @xmath2=co or rh ) @xcite , whereas the conventional kohler s rule is completely broken .
this long - standing problem had been naturally explained in terms of the fermi liquid theory .
the modified kohler s rule implies that the enhancements of @xmath9 and @xmath12 are induced by the same origin ; the cvc due to af fluctuations . for ( i ) ybco and ( ii ) lsco calculated by the flex+cvc method for @xmath659 ( 80k @xmath603 400k ) .
the modified kohler s rule ( @xmath660const . ) holds well by considering the cvc . on the other hand , @xmath661 obtained by
the rta exhibits a strong @xmath15-dependence .
@xcite ] ] figure [ fig : mkr ] shows the numerical results for ( i ) ybco and ( ii ) lsco by using the flex+cvc method .
the coefficient @xmath662 takes an almost constant value for a wide range of temperatures , only when the cvcs are correctly considered : in fig .
[ fig : mkr ] , we show that @xmath663 and @xmath664 : @xmath661 increases as @xmath15 decreases , since the anisotropy of @xmath665 increases at low temperatures . in contrast , @xmath666 decreases at low temperatures since the @xmath9 is strikingly enhanced by the cvc .
only when both @xmath387 and @xmath589 are calculated using the flex+cvc method , the modified kohler s rule @xmath667const .
is satisfied . in ybco ,
the dominant contribution to the total magneto - conductivity is given by @xmath648 , particularly by the first term in the bracket in eq .
( [ eqn : mca2 ] ) that includes @xmath668 . in the case of lsco , the second term in eq .
( [ eqn : mca2 ] ) is very important .
further , @xmath669 . in both these systems , the modified kohler s rule ( [ eqn : mkr - the ] ) is well reproduced by the flex+cvc method .
the experimental value is @xmath670 for ybco @xcite and @xmath671 for tl@xmath50ba@xmath50cuo@xmath53 @xcite .
in contrast , @xmath672 takes much larger value in lsco : @xmath673 for a slightly over - doped sample ( @xmath674 ) @xcite , and it increases ( decreases ) as the doping increases ( decreases ) @xcite .
these experimental results are reproduced by the present study and they are shown in fig .
[ fig : mkr ] . in lsco ,
the fermi surface is very close to the van - hove singularity point @xmath140 since both @xmath675 and @xmath676 in eq .
( [ eqn : band ] ) are smaller than those in other systems . since @xmath677 at @xmath140 , the anisotropy of the velocity @xmath678 on the fermi surface is larger in lsco .
hence , the second term in eq .
( [ eqn : mca2 ] ) [ or eq .
( [ eqn : mcrta ] ) ] takes a large value in lsco . in
heavily over - doped lsco at @xmath679 , where cvc is expected to be unimportant , @xmath672 exceeds 100 @xcite .
this result is consistent with recent arpes measurement @xcite , which shows that the fermi surface in lsco passes through @xmath140 in the over - doped region @xmath680 .
therefore , experimental value of @xmath672 gives us a useful measure of the anisotropy of @xmath644 ( @xmath681 in weakly correlated systems ) . in ce@xmath2in@xmath3 ( @xmath2=rh , co ) ,
@xmath672 is of order @xmath682 @xcite , which indicates that the anisotropy of @xmath683 is large in the main fermi surface .
we also discuss the thermoelectric power , @xmath14 . in htscs
, @xmath14 takes a large value in under - doped systems , and it increases as @xmath15 decreases from the room temperature . except for over - doped compounds , @xmath684 in hole - doped systems @xcite whereas @xmath685 in electron - doped systems @xcite below the room temperature .
interestingly , the peak temperature of @xmath14 is nearly equal to the pseudo - gap temperature @xmath34 in many hole - doped compounds ; in hgba@xmath50cuo@xmath686 @xcite , lsco , ybco @xcite and bi@xmath50sr@xmath50rcu@xmath50o@xmath63 ( r = ca , y , pr , dy or er ) @xcite . by neglecting the cvc ,
hildebrand et al .
calculated the thermoelectric power for ybco using the flex approximation @xcite . here , we study the thermoelectric power both for hole - doped and electron - doped systems using the flex+cvc method .
according to the linear response theory @xcite , the thermoelectric power in a fermi liquid system is given by @xmath687 where @xmath688 is the diagonal peltier conductivity ; @xmath689 .
it is given by @xcite , @xmath690 where @xmath691 is given by the analytic continuation of @xcite @xmath692 where @xmath693 and @xmath694 are charge current operator and heat current one , respectively .
the heat current is defined by the following energy conservation law ; @xmath695 .
then , @xmath694 contains complex two - body terms in the presence of electron - electron correlation .
fortunately , the ward identity allows us to rewrite the heat current as @xmath696 , where @xmath697 @xcite . by performing the analytic continuation of eq .
( [ eqn : kaxx ] ) , we can derive the exact expression of @xmath688 of order @xmath698 as @xcite @xmath699 , \label{eqn : lxx } \end{aligned}\ ] ] where @xmath385 is the total current given in eq .
( [ eqn : bs - t ] ) . at sufficiently low temperatures , @xmath688 is approximately simplified as @xmath700 where @xmath261 is the electron charge .
@xmath701 represents the momentum vertical to the fermi surface .
when @xmath702 is energy - independent , eq .
( [ eqn : lxx - lowt2 ] ) yields @xmath703 ( @xmath704 is the effective mass ) in the 2d isotropic system with @xmath705 . for ybco and ncco given by the flex+cvc approximation , by including the cvc for @xmath19 .
reflecting the reduction in af fluctuations , @xmath14 starts to decrease below @xmath706 k in the flex+@xmath15-matrix method , which is consistent with the experimental pseudo - gap behavior in @xmath14 . *
( ii ) * @xmath707 denotes the thermoelectric power given by the inside ( outside ) of the afbz
. @xmath708 .
@xcite ] ] in fig .
[ fig : s ] ( i ) , we show the numerical results for @xmath14 of ybco and ncco given by the flex+cvc method . in lsco ,
@xmath684 in the optimally - doped case ( @xmath151 ) below the room temperature , and it increases in the under - doped case . in ncco ( @xmath569 ) , @xmath685 and it takes a peak around @xmath709k .
these results are consistent with the experimental results @xcite .
the origin of the non - fermi - liquid - like behavior is the prominent @xmath710 and @xmath711-dependences of @xmath712 due to strong af fluctuations . according to the spin fluctuation theory , @xmath713 at point a in fig .
[ fig : fs - hotcold ] ( i ) because @xmath181 becomes large near the afbz .
since point a is the cold spot of ybco , eq .
( [ eqn : lxx - lowt2 ] ) becomes positive in ybco .
therefore , @xmath684 in under - doped ybco and lsco . on the other hand , @xmath685 in ncco because @xmath714 at point b. in ybco , @xmath14 slightly decreases if the cvc is included , since @xmath715 around the cold spot . in ncco , in contrast
, absolute value of @xmath14 is enhanced by the cvc . in both cases ,
the effect of the cvc on @xmath14 is smaller than that on @xmath9 .
figure [ fig : s ] ( ii ) shows the @xmath14 for ybco ( @xmath716 ) , together with @xmath717 ( @xmath718 ) which denotes the contribution from the inside ( outside ) of the afbz .
@xmath717 is positive and it deviates from a conventional behavior ( @xmath719 ) since the @xmath711-dependence of @xmath720 becomes large inside of the afbz at lower temperatures . on the other hand ,
@xmath718 is negative and it approaches zero as @xmath15 decreases . therefore , the sign of @xmath721 becomes positive below @xmath722 ( @xmath603 300 k )
. in htscs , anomalous transport phenomena ( such as the violation of kohler s rule ) had been frequently considered as strong evidence for the breakdown of the quasiparticle picture .
for example , the rta for the highly anisotropic @xmath18 model can not reproduce relationships ( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) for hole - doped systems _ at the same time for a wide range of temperatures_. furthermore , the rta can not explain the negative hall coefficient in electron - doped systems , since the curvature of the true fermi surface is positive everywhere .
the explanation for the nearly symmetric behavior of @xmath9 and @xmath14 with respect to the type of carrier doping ( shown in fig .
[ fig : sato ] ) was desirable for a long time .
these highly nontrivial transport phenomena had been one of the central issues in htscs , which should serve to elucidate the actual electric ground state of htscs . to resolve this long - standing problem
, we developed a method to calculate various transport coefficients based on the microscopic fermi liquid theory . in the rta ,
the momentum and energy transfers between the quasiparticles by scattering are not treated correctly .
therefore , the rta results frequently yield erroneous results . to overcome this defect , we study the role of the cvc in nearly af fermi liquids .
we find that the total current @xmath19 shows an anomalous @xmath21-dependence because of the cvc in nearly af metals , which gives rise to the non - fermi liquid behaviors such as @xmath723 and @xmath724 .
consequently , the modified kohler s rule @xmath725 is realized in htscs .
we also studied the @xmath14 for htscs , which increases as the temperature decreases above @xmath34 .
this experimental fact is reproduced by considering the strong @xmath710 and @xmath711-dependences of @xmath726 @xcite .
therefore , various anomalous transport phenomena in htscs above @xmath34 can be well explained by the flex+cvc method .
recently , tsukada et al . measured the @xmath9 in clean heavily over - doped lsco samples @xcite , and found that @xmath9 becomes almost temperature - independent for @xmath727 ; further , its sign smoothly changes from positive to negative between @xmath728 and @xmath729 , corresponding to the change in the curvature of the fermi surface .
hussey et al . also showed that the rta analysis is successful in heavily over - doped tl2201 @xcite and lsco @xcite .
these experimental results are consistent with the present theory : according to the present analysis , the rta works well in heavily over - doped systems since the cvc is unimportant when the af fluctuations are very weak .
this fact is well explained in ref .
@xcite . finally , we comment on other important theoretical study involving the cvc in strongly correlated systems . in the interacting electron gas model ,
the electron cyclotron frequency @xmath94 is given by @xmath730 , where @xmath311 is the bare electron mass . in 1961
, kohn proved that @xmath94 is unchanged by electron - electron interaction due to the consequence of the angular momentum conservation law , which is called the `` kohn s theorem '' @xcite .
the kohn s theorem is also proved by the phenomenological fermi liquid theory , by correctly considering the cvc @xcite .
kanki and yamada studied this problem based on the microscopic fermi liquid theory , and found that @xmath94 is influenced by the electron - electron correlation if the umklapp processes are present @xcite .
the umklapp processes also reduces the drude weight in @xmath731 @xcite and the penetration depth in the sc state @xcite .
moreover , the effect of the vertex correction on the raman spectroscopy in htsc was studied @xcite .
-matrix induced by the af fluctuations . *
( i ) * self - energy @xmath732 in the flex+@xmath15-matrix approximation . *
( iii ) * dos given by the flex+@xmath15-matrix approximation .
the pseudo - gap appears below @xmath733 due to the strong sc fluctuations . *
( iv ) * im@xmath734 and im@xmath735 given by the flex+@xmath15-matrix approximation at @xmath152 . for comparison ,
im@xmath736 given by the flex approximation at @xmath152 is also shown . ] in under - doped htscs , a deep pseudo - gap in the dos emerges at the chemical potential below @xmath737200 k. the origin of the pseudo - gap has been a central issue with regard to htscs . the pseudo - gap does not originate from the spin fluctuations since the strength of the spin fluctuations decrease below @xmath34 @xcite . according to recent arpes measurements
@xcite , the @xmath21-dependence of the pseudo - gap coincides with that of the @xmath36-wave sc gap function .
this pseudo - gap structure in the quasiparticle spectrum @xmath738 starts to appear around @xmath533 below @xmath34 .
below @xmath1 , the quasiparticle spectrum shows sharp peaks at the edge of the sc gap ( @xmath739 ) since the inelastic scattering is reduced by the sc gap . at the same time , a dip - hump structure is induced around @xmath740 by the strong resonance peak in im@xmath741 at @xmath742 mev in ybco and bi@xmath50sr@xmath50cacu@xmath50o@xmath63 ( bscco ) .
we note that @xmath743 and @xmath744 .
however , the overall gap - like structure in the quasiparticle spectrum remains unchanged around @xmath1 .
motivated by these experimental facts , various strong - coupling theories of sc fluctuations have been studied to reproduce the pseudo - gap phenomena @xcite .
several groups developed the `` flex+@xmath15-matrix theory '' , where the normal self - energy correction induced by the strong sc fluctuations , @xmath745 , has been self - consistently included into the flex approximation .
the sc `` amplitude '' fluctuations are given by the @xmath15-matrix ( particle - particle scattering amplitude ) induced by the af fluctuations , which prominently increase below @xmath34 @xcite .
the @xmath15-matrix with respect to the af fluctuations , which is responsible for the thouless instability for the @xmath36-channel , is given by @xmath746 which is shown in fig .
[ fig : flex - t ] ( i ) .
@xmath747 is given in eq .
( [ eqn : def_v ] ) . in the flex+@xmath15-matrix approximation , the green function and the self - energy
are given by @xmath748^{-1 } , \label{eqn : g - flext } \\
\sigma_\k^{\rm aff}(\e_n)&= & \mbox{eq .
( \ref{eqn : self } ) } , \nonumber \\
scf}(\e_n)&= & t\sum_{\q , l } g_{-\k+\q}(-\e_n+\w_l ) t_{\k,\k;\q}(\e_n,\e_n;\w_l ) , \label{eqn : self - scf}\end{aligned}\ ] ] where @xmath749 and @xmath750 are shown in fig .
[ fig : flex - t ] ( ii ) . in the self - consistent flex+@xmath15-matrix approximation
, we have to solve eqs .
( [ eqn : self ] ) , ( [ eqn : def_v ] ) , and ( [ eqn : tpg])-([eqn : self - scf ] ) self - consistently .
unfortunately , it is very difficult to calculate the @xmath15-matrix in eq .
( [ eqn : tpg ] ) since a lot of memory in computer is required .
fortunately , eq . ( [ eqn : tpg ] ) can be simplified by considering only the @xmath36-channel and drop other pairing channels as @xmath751 , where @xmath752 @xcite .
this approximation is expected to be reasonable when the temperature is close to @xmath1 in htscs .
then , the @xmath21-dependence of @xmath750 becomes @xmath753 . in the present approximation , @xmath34
is defined as the temperature below which @xmath750 takes considerable values .
@xmath34 is slightly greater than the @xmath754 in the flex approximation . in a pure 2d system , @xmath755 since the flex+@xmath15-matrix theory satisfies the mermin - wagner - hohenberg theorem .
@xmath1 becomes finite when weak three - dimensionality is assumed . sufficiently above @xmath34 , where @xmath756 , the flex+@xmath15-matrix theory is equivalent to the flex approximation .
the obtained dos and the self - energy are shown in fig .
[ fig : flex - t ] ( iii ) and ( iv ) , respectively .
below @xmath757 , a large pseudo - gap structure appears in the flex+@xmath15-matrix approximation , due to the cooperation of the real part and the imaginary part of @xmath750 @xcite .
the af fluctuations are suppressed below @xmath60 in accordance with the pseudo - gap in the dos @xcite .
for this reason , im@xmath758 in the flex+@xmath15-matrix approximation is much smaller than im@xmath759 in the flex approximation , as recognized in fig .
[ fig : flex - t ] ( iv ) .
below @xmath1 , the pseudo - gap in the flex+@xmath15-matrix approximation smoothly changes to the sc gap @xcite . here
, we discuss transport phenomena in the pseudo - gap region . in ybco
@xcite , bi2201 and bi2212 @xcite , @xmath9 shows a maximum around the pseudo - gap temperature @xmath34 , at which @xmath59 shows the maximum value . also , the peak temperature of @xmath14 is nearly equal to @xmath34 in hgba@xmath50cuo@xmath686 @xcite , lsco , ybco @xcite and bi@xmath50sr@xmath50rcu@xmath50o@xmath63 ( r = ca , y , pr , dy and er ) @xcite . as discussed in ref .
@xcite , these behaviors are naturally understood since both of them are strikingly enhanced by the af fluctuations above @xmath60 .
therefore , _ both @xmath9 and @xmath14 should decrease below @xmath34 in accordance with the reduction in the af fluctuation .
_ this behavior is easily reproduced by the flex+@xmath15-matrix method @xcite .
the obtained @xmath6 , @xmath9 and @xmath14 for hole - doped systems by using the [ flex+@xmath15-matrix]+cvc method are shown in figs .
[ fig : rho ] ( i ) , [ fig : rh ] ( i ) , and [ fig : s ] ( i ) , respectively . in this approximation ,
@xmath760 in eq .
( [ eqn : bs - t ] ) is given by @xmath761 { \rm im}\ { v_{\k-\k'}(\e-\e ' ) - t_{\k,\k;\k-\k'}(\e,\e;\e-\e ' ) \}$ ] , where @xmath762 represents the @xmath15-matrix in eq .
( [ eqn : tpg ] ) .
the detailed method of calculation is explained in ref .
@xcite . as shown in figs .
[ fig : rh ] ( i ) and [ fig : s ] ( i ) , both @xmath9 and @xmath14 start to decrease below @xmath763 ( @xmath603160 k ) in the flex+@xmath15-matrix approximation .
these results are consistent with experiments @xcite .
figure [ fig : rho ] ( i ) shows the resistivity given by the flex approximation and the flex+@xmath15-matrix approximation , both of which are given by including cvcs . as pointed out in ref .
@xcite , @xmath6 given by the flex+cvc method shows a tiny `` kink '' structure at @xmath557 because of the cvc .
this kink structure at @xmath764 is experimentally observed in lsco @xcite and ybco @xcite .
the kink becomes more prominent in the [ flex+@xmath15-matrix]+cvc method since the inelastic scattering is reduced by the formation of the pseudo - gap .
as shown in fig .
[ fig : flex - t ] ( iv ) , im@xmath758 decreases prominently below @xmath60 , in accordance with the emergence of im@xmath765 . as a result , @xmath6 is approximately proportional to @xmath8 below @xmath34 . to summarize , @xmath6 , @xmath9 and @xmath14
are suppressed in the pseudo - gap region in the [ flex+@xmath15-matrix]+cvc method . according to the flex+@xmath15-matrix theory , the sc `` amplitude '' fluctuations is the origin of the pseudo - gap formation .
other than this scenario , various mechanisms of pseudo - gap formation have been proposed .
for example , kivelson and emery proposed the sc `` phase '' fluctuation scenario @xcite : they considered that the amplitude of the sc order - parameter develops in the pseudo - gap region , whereas the global phase coherence is absent .
moreover , several hidden - order scenarios , such as @xmath64-density wave formation @xcite , have been proposed .
to elucidate the correct scenario , anomalous transport phenomena in the pseudo - gap region are significant and they severely constrain the theories . in the following ,
we show that the scenario involving sc `` amplitude '' fluctuations enables us to understand the anomalous transport phenomena below @xmath34 in a unified way . for lsco obtained by the [ flex+@xmath15-matrix]+cvc ; the modified kohler s rule @xmath766 holds above @xmath60 .
below @xmath767 k , modified kohler s rule is violated since @xmath12 increases quite drastically that is caused by the cvc due to af+sc fluctuations . *
( ii ) * @xmath22 for lsco obtained by the [ flex+@xmath15-matrix]+cvc ; the abrupt increase in @xmath22 below @xmath34 is caused by the cvc due to af+sc fluctuations .
experimental data @xmath768 is cited from @xcite . *
( iii ) * @xmath22 for ncco is obtained by using the flex+cvc approximation .
experimental data @xmath768 is cited from @xcite .
@xcite ] ] in contrast to @xmath9 and @xmath14 , the nernst coefficient @xmath22 ( @xmath769 ; off - diagonal thermoelectric power under a magnetic field ) prominently increases in the pseudo - gap region by approximately 100 times than that in conventional metals @xcite .
moreover , in the pseudo - gap region , the magnetoresistance @xmath12 increases faster than @xmath770 @xcite .
therefore , the modified kohler s rule in eq .
( [ eqn : mkr - the ] ) violates below @xmath34 .
such nontrivial behaviors of @xmath22 and @xmath12 have attracted considerable attention as a key phenomenon closely related to the origin of the pseudo - gap . here , we study the nernst coefficient . according to the linear response theory @xcite
, @xmath22 is given by @xmath771/h_z , \label{eqn : nu1}\end{aligned}\ ] ] where @xmath772 , and @xmath773 is the off - diagonal peltier conductivity under the magnetic field @xmath392 . in conventional metals with simple fermi surfaces ,
@xmath22 is small because of an approximate cancellation between the first and the second terms in eq.([eqn : nu1 ] ) , which is known as the sondheimer cancellation .
( sondheimer cancellation is exact only when @xmath774 . ) in htscs , however , sondheimer cancellation is totally violated since @xmath590 is considerably enhanced .
that is , @xmath590 is the origin of giant nernst coefficient in lsco below @xmath60 and in ncco below @xmath775 @xcite . in this study , we investigate @xmath590 due to the quasiparticle transport , and find that @xmath590 is prominently enhanced below @xmath34 if we include the cvc caused by sc and af fluctuations . as a result , @xmath776 takes a large value below @xmath34 .
moreover , we show that all the transport anomalies below @xmath34 are understood as the quasiparticle transport phenomena in a unified way .
note that yip showed that the maki - thompson - type cvc due to sc fluctuations disappears in @xmath64-wave superconductors when the inelastic scatterings are negligible @xcite .
however , this is inappropriate for htsc since inelastic scatterings are much larger than elastic scatterings .
we find that the maki - thompson type cvc is significant in htsc .
we note that ussishkin et al .
studied @xmath22 induced by short - lived cooper pairs , which is independent of the quasiparticle lifetime and order @xmath777 @xcite .
it will be observed in the close vicinity of @xmath1 .
however , quasiparticle transport is dominant in in good metals where @xmath778 .
( here , @xmath779 is the mean free path . ) in htscs , @xmath780 corresponds to @xmath781 cm if we omit the @xmath21-dependence of @xmath782 .
therefore , quasiparticle transport will be dominant in slightly under - doped htsc ( @xmath783 cm above @xmath1 ) for a wide range of temperatures .
hereafter , we show that @xmath590 due to quasiparticle transport is strongly enhanced by the cvc . according to the microscopic fermi liquid theory , @xmath590
is given by the correlation function between the heat current and the charge current in the presence of @xmath784 .
therefore , @xmath590 is given by eq .
( [ eqn : km2 ] ) by replacing @xmath785 with @xmath786 . after the analytic continuation
, the exact expression for @xmath590 of the order @xmath374 is given by @xcite @xmath787 where @xmath788 , @xmath789 is the quasiparticle heat velocity , and @xmath790 is the total heat current @xcite .
we stress that the cvc term in eq .
( [ eqn : q ] ) vanishes if we omit the energy dependence of @xmath760 .
this fact means that the heat cvc is small and thus @xmath791 in general cases @xcite . in the flex+@xmath15-matrix approximation , @xmath760 is given by @xmath761 { \rm im}\ { v_{\k-\k'}(\e-\e ' ) - t_{\k,\k;\k-\k'}(\e,\e;\e-\e ' ) \}$ ] .
in contrast to @xmath9 and @xmath14 , the nernst coefficient @xmath22 @xcite and @xmath12 @xcite in lsco rapidly increase below @xmath34 .
these experimental facts are also reproduced by the [ flex+@xmath15-matrix]+cvc method as shown in fig .
[ fig : nu ] ( i ) and ( ii ) . since @xmath9 decreases whereas @xmath12 increases drastically below @xmath34 , the plot of @xmath12 as a function of @xmath792 forms an `` inverse s - shape '' .
therefore , the modified kohler s rule is completely violated below @xmath34 .
this result is very similar to the experimental results provided in ref .
next , we discuss @xmath22 in ncco using the flex ( not the flex+@xmath15-matrix ) approximation since sc fluctuations is considered to be absent in ncco .
the obtained @xmath22 is shown in fig .
[ fig : nu ] ( iii ) ; it takes a large value only when the cvc due to the af fluctuations are considered . in ncco , @xmath22 increases gradually as @xmath15 decreases , and it starts to decrease below the maximum temperature of 120 k. the obtained behaviors of @xmath22 are semiquantitatively consistent with the experimental results @xcite .
therefore , the present numerical study can explain the experimental behaviors of @xmath22 and @xmath12 in htscs both above and below @xmath60 . here
, we explain a theoretical reason why @xmath22 is enhanced in both lsco and ncco : @xmath793 in eq .
( [ eqn : a ] ) can be rewritten as @xmath794 where @xmath795 , @xmath796 , and @xmath797 is the mean free path with cvc . in the rta ,
the heat current is given by @xmath798 . since @xmath799 ,
the second term is absent in the rta . if one includes the cvc
, @xmath800 is not parallel to @xmath19 because the cvc for @xmath801 is usually small ( i.e. , @xmath802 ) as we will discuss below @xcite .
therefore , we obtain @xmath803 . in this case , the second term gives rise to an enhancement of @xmath22 if @xmath804 is highly anisotropic around the cold spot .
figure [ fig : numerical - j ] ( ii ) shows @xmath804 in lsco given by the flex+cvc approximation .
we can see that @xmath805 takes a large value near the cold spot below @xmath152 .
( note that @xmath806 just at point a. ) this is the origin of increment in @xmath807 in fig .
[ fig : nu ] ( ii ) . as shown in fig .
[ fig : nu ] ( iii ) , @xmath807 for ncco is much larger than @xmath807 for lsco : one reason is that the anisotropy of @xmath808 in ncco is greater since @xmath809 in ncco .
another reason is that @xmath810 in ncco is larger than that in lsco . for these reasons
, flex+cvc method can explain the huge @xmath22 observed in ncco .
next , we consider the increment of @xmath22 in the pseudo - gap region in lsco using the flex+@xmath15-matrix method , which is shown as @xmath811 in fig .
[ fig : nu ] ( ii ) .
the cvc due to sc fluctuations ( maki - thompson term ) represents the acceleration of quasiparticles caused by the short - lived cooper pairs .
since htscs are @xmath36-wave superconductors , the cvc due to the sc fluctuations magnifies @xmath812 except at the nodal point ( point a in fig .
[ fig : j ] ) .
in fact , when the ginzburg - landau correlation length is longer than @xmath108 , the total current in the flex+@xmath15-matrix approximation is approximately given by @xcite @xmath813 where @xmath814 has been introduced in eq .
( [ eqn : jdash ] ) .
@xmath815 represents the enhancement factor due to sc fluctuations .
the anisotropy of @xmath816 becomes considerably prominent below @xmath60 since im@xmath817 increases whereas im@xmath758 decreases below @xmath60 as shown in fig .
[ fig : flex - t ] ( iv ) . for this reason , the second term in eq .
( [ eqn : a2 ] ) takes a large value in the neighborhood of the point a , as explained in ref .
thus , [ flex+@xmath15-matrix]+cvc method can explain the rapid increment in @xmath22 in lsco below @xmath60 .
if the af fluctuations are absent , @xmath818 even if the cvc due to sc fluctuations are taken into account
. therefore , to explain the enhancement in @xmath22 in the pseudo - gap region , we have to include cvcs due to both the af fluctuations and sc fluctuations .
in contrast , @xmath9 decreases below @xmath34 in proportion to @xmath111 since the second term in eq .
( [ eqn : a2 ] ) is absent if we replace @xmath800 with @xmath19 .
finally , we explain why the heat cvc is usually small .
as we have explained , the cvc represents the current of the fermi sea that is transfered from the excited quasiparticles by the electron - electron scattering .
the charge cvc can be large because of the momentum conservation law .
however , the heat current @xmath819 is not conserved even in the free dispersion model , that is , @xmath820 even if @xmath821 .
for this reason , the heat cvc is small in general , and therefore the thermal conductivity @xmath4 is finite even if the umklapp process is absent : in a 3d free dispersion model , @xmath822 within the second - order perturbation theory with respect to @xmath23 @xcite .
the heat cvc is also small in nearly af metals : since the direciton of the the heat cvc due to the quasiparticle at @xmath28 in fig .
[ fig : cvc - af ] ( the mt term ) depends on the sign of @xmath823 , the heat cvc becomes small due to cancellation after the @xmath240-summation for @xmath465 . in conclusion , we studied @xmath6 , @xmath9 , @xmath12 , @xmath14 and @xmath22 in htscs using the flex+@xmath15-matrix approximation .
the results are shown in figs .
[ fig : rho ] ( i ) , [ fig : rh ] ( i ) , [ fig : s ] ( i ) , and [ fig : nu ] ( i)-(ii ) .
we can explain that ( i ) @xmath9 and @xmath14 start to decrease below @xmath34 , ( ii ) @xmath6 starts to deviate from the @xmath15-linear behavior at @xmath775 , and it is approximately proportional to @xmath8 below @xmath34 .
moreover , ( iii ) @xmath22 and @xmath824 further increase below @xmath34 .
therefore , the present study gives us a unified understanding of the various anomalous transport phenomena in both below and above @xmath60 .
this study provides strong evidence that the sc fluctuations are predominant in the pseudo - gap region .
the striking increment in @xmath22 below @xmath34 is a cooperative phenomenon between the d - wave sc fluctuations and af fluctuations .
we stress that the rta can not explain the enhancement of @xmath22 since the second term in eq .
( [ eqn : a2 ] ) vanishes identically in the rta .
further , a very large nernst signal appears also in electron - doped systems , which starts to increase in proportion to @xmath825 below @xmath826 k. this experimental fact is reproduced by the cvc due to the af - fluctuations , even in the absence of sc fluctuations . in [ nernst ] , we found that the origin of rapid increment in @xmath22 is the second term of eq .
( [ eqn : a2 ] ) , which emerges only when the cvc is included . in ncco
( and ce@xmath2in@xmath3 ) , @xmath22 starts to increase below @xmath775 because of the cvc due to strong af fluctuations ; @xmath827 . in lsco , on the other hand , increment in @xmath22 below @xmath775 is small since @xmath108 is of order 1 ; @xmath22 in lsco starts to increase rapidly below @xmath828 , with the aid of the strong sc fluctuations . here ,
we comment on the transport phenomena in nd - doped lsco , la@xmath829nd@xmath830sr@xmath148cuo@xmath49 , which shows the static stripe ordered phase at @xmath831 k @xcite . in this compound
, @xmath6 increases whereas @xmath9 decreases monotonically below @xmath832 , which indicates that the system becomes one - dimensional due to the stripe order .
this experimental fact might tempt us to consider that the stripe order is the origin of the reduction in @xmath9 in other htscs in the pseudo - gap region .
however , the stripe order scenario can not explain the reduction in @xmath6 nor increment in @xmath22 below @xmath60 .
on the other hand , the overall transport phenomena in htscs can be explained by considering the cvcs due to af+sc fluctuations . in this section
, we explained that the pseudo - gap behaviors in @xmath6 and @xmath9 originate from the reduction of af fluctuations , and the pseudo - gap behavior of @xmath22 reflects the increment of sc fluctuations .
the pseudo - gap temperature determined by @xmath6 , @xmath833 , is larger than @xmath834 and @xmath835 determined by @xmath9 and @xmath22 , respectively .
recently , impurity ( zn ) effect on the pseudo - gap behavior in ybco was studied @xcite , and found that @xmath833 and @xmath834 are independent of impurities , whereas @xmath835 decreases with zn - doping in parallel to @xmath1 .
it is an important future problem to show whether these experimental facts can be explained by the flex+@xmath15-matrix approximation or not . in
[ imp ] , we will show that the cvc due to fluctuations is usually sensitive to the impurities .
therefore , to reconcile this problem , we have to study the impurity effect on the cvc seriously .
it should be noted that @xmath22 takes a very large value in the vortex - liquid state above @xmath836 in a clean 2d sample , reflecting the high mobility of the vortices .
therefore , @xmath22 is frequently used as a sensitive probe for the mixed state .
based on his observation , ong _ et al .
_ proposed that spontaneous vortex - antivortex pairs emerge in under - doped systems below @xmath34 , and they govern the transport phenomena in the pseudo - gap region @xcite
. however , this assumption seems to contradict with the other transport coefficients ; for example , the flux - flow resistance does not appear below @xmath34 .
in optimally - doped ybco at 95 k , derived from faraday rotation . the closed circles and squares represent experimental values of re@xmath837 and im@xmath837 , respectively . [ ref .
@xcite ] ] in previous sections , we discussed the dc transport phenomena in nearly af metals , and found that the cvc produces various non - fermi - liquid - like behaviors . in principle , the ac transport phenomena can yield further useful and decisive information about the electronic status .
unfortunately , the measurements of the ac transport coefficients are not common because of the difficulty in their observations , except for the optical conductivity @xmath838 measurements .
fortunately , drew s group has performed intensive measurements of the ac hall coefficient @xmath839 in ybco @xcite , bscco @xcite , lsco @xcite , and pcco @xcite.they found that the @xmath5-dependence of @xmath837 in htsc shows amazing non - fermi - liquid - like behaviors , which have been a big challenge for researchers for a long time . here , we show that this crucial experimental constraint is well satisfied by the numerical study using the flex+cvc method . in the rta , both @xmath838 and @xmath840 in a single - band model will follow the following `` extended drude forms '' : @xmath841 where @xmath842 is the mass - enhancement factor and @xmath843 is the @xmath5-dependent damping rate in the optical conductivity , which is approximately given by @xmath844 for small @xmath5 . according to the spin fluctuation theory @xcite , @xmath845 , which is observed by the optical conductivity measurements .
the @xmath5-dependence of @xmath118 is important in heavy - fermion systems ( @xmath846 at @xmath533 ) , whereas it will not be so important in htsc since @xmath847 is rather small .
expressions ( [ eqn : edxx ] ) and ( [ eqn : edxy ] ) are called the `` extended drude form '' . within the rta ,
the ac - hall coefficient is independent of @xmath5 even if the @xmath5-dependence of @xmath118 is considered : @xmath848 very interestingly , drew s group has revealed that @xmath837 in htsc decreases drastically with @xmath5 : as shown in fig .
[ fig : drew ] , @xmath849 shows a peak at @xmath850 in optimally - doped ybco @xcite .
moreover , @xmath849 is as large as @xmath851 for @xmath852 , as a consequence of the kramers - kronig relation between @xmath851 and @xmath849 .
such a large @xmath5-dependence of @xmath9 can not be explained by the rta , even if one assume an arbitrary ( @xmath21,@xmath5)-dependence of the quasiparticle damping rate @xmath853 .
therefore , the ac - hall effect severely constrains the theories involving the normal state of htscs . and ( ii ) @xmath837 for lsco ( @xmath189 , @xmath854 ) given by the flex+cvc method .
@xmath855 corresponds to @xmath856@xmath857 . because of the cvc , @xmath858 at low frequencies .
@xcite ] ] recently , we studied both @xmath838 and @xmath840 in htsc using the flex+cvc method , by performing the analytic continuation of eqs .
( [ eqn : sig - d ] ) and ( [ eqn : sxy - d ] ) using the pade approximation @xcite .
since the @xmath5-dependence of the cvc is correctly considered , the obtained @xmath838 and @xmath840 satisfy the @xmath17-sum rule very well @xmath859 and @xmath860 ; the relative error is less than 2.5% @xcite .
note that a useful @xmath17-sum rule for the hall angle has been derived in ref .
the obtained numerical results are shown in fig .
[ fig : achall ] .
we find that ( a ) @xmath861 follows eq .
( [ eqn : edxx ] ) , by substituting @xmath862 .
therefore , the reduction in the drude weight in @xmath838 due to the cvc is small within the flex+cvc method .
however , we stress that the @xmath17-sum rules for both @xmath863 and @xmath864 do not hold since the cvc is ignored @xcite .
( b ) @xmath865 is a nearly real constant of the order @xmath866 since @xmath867 ^ 2 \propto \s_{xy}^{\rm rta}(\w)$ ] is well satisfied .
( c ) @xmath840 obtained using the flex+cvc method completely deviates from the extended drude forms in eq .
( [ eqn : edxy ] ) .
the effect of cvc on @xmath840 is very large for the far - infrared region @xmath868 , since the cvc due to af fluctuations is strong for @xmath869 .
therefore , @xmath840 and @xmath837 obtained using the flex+cvc method show striking @xmath5-dependences as shown in fig .
[ fig : achall ] . in fig .
[ fig : achall ] ( i ) , the peak frequency of im@xmath840 , @xmath870 , is six times less than that of im@xmath838 , @xmath871 , which is consistent with experiments . this fact can not be explained by using the extended drude form . here
, we analyze a qualitative @xmath5-dependence of @xmath840 by considering the @xmath5-dependence of the cvc . the bethe - salpeter equation for @xmath872
is given by @xmath873 which is equivalent to eq .
( [ eqn : bs - t ] ) when @xmath533 .
a simplified bethe - salpeter equation for @xmath533 is given in eq .
( [ eqn : bs - ap ] ) . here , we extend eq .
( [ eqn : bs - ap ] ) to the case of @xmath872 . by noticing the relationship @xmath874 , eq .
( [ eqn : bs - aw ] ) is simplified for @xmath875 as @xmath876 where @xmath877 . in deriving eq .
( [ eqn : bs - apw ] ) , we assumed that the @xmath711-dependence of @xmath878 is small for @xmath879 .
the solution of eq .
( [ eqn : bs - apw ] ) is given by @xmath880 , \label{eqn : j - apw } \end{aligned}\ ] ] which is equivalent to eq .
( [ eqn : j - ap ] ) when @xmath533 .
similar to the derivation of eq .
( [ eqn : dtdk ] ) , we get @xmath881 as a result , @xmath840 for @xmath882 is approximately given by @xmath883 if we omit the cvc ( i.e. , @xmath627 ) , eq . ( [ eqn : s - xyw ] ) is equivalent to the extended drude form in eq .
( [ eqn : edxy ] ) . according to eq .
( [ eqn : s - xyw ] ) , the solution of re@xmath884 is given by @xmath885 , which is much smaller than the solution obtained from the rta ; @xmath886
. this analytical result can be recognized in the numerical result shown in fig .
[ fig : achall ] . as a result
, the striking @xmath5-dependence of @xmath837 in htscs for @xmath887 can be explained by including the cvc .
-dependence of @xmath888 in bscco at 300k .
open squares and diamonds represent real and imaginary part of theoretical result given by the flex+cvc method . *
( ii ) * @xmath888 in optimally - doped bscco at 950 @xmath857 .
the solid dots represent the result of the flex+cvc method .
@xcite ] ] furthermore , the effect of cvc is still large even for the infrared region @xmath889 , which is consistent with the experimental results .
this fact requires the satisfaction of the @xmath17-sum rule , @xmath890 , since the dc @xmath387 is significantly enhanced .
schmadel et al .
@xcite observed @xmath840 in optimally doped bscco for a wide range of @xmath5 ( @xmath891 ) , and found that @xmath840 follows a `` simple drude form '' ; @xmath892 , where @xmath5-dependence of @xmath893 is much smaller than @xmath843 and @xmath894 .
the experimental @xmath5-dependence of @xmath888 as well as theoretical result are shown in fig .
[ fig : drew2 ] ( i ) . moreover , fig .
[ fig : drew2 ] ( ii ) shows that re@xmath895 depends on @xmath15 sensitively .
therefore , @xmath893 is independent of @xmath5 in the infrared region . in terms of the rta
, this experimental result highly contradicts the fact that @xmath896 is independent of @xmath15 for @xmath897 ; @xmath898 .
this mysterious behavior is well reproduced by the flex+cvc method , as indicated by the solid dots in fig .
[ fig : drew2 ] ( ii ) .
this numerical result suggests that the effect of the @xmath5-dependence of @xmath899 on the ac - hall conductivity , by which @xmath864 deviates from the simple drude form , approximately cancels with that of the cvc .
interestingly , the hall angle @xmath900 follows an approximate simple drude form in both ybco @xcite and lsco @xcite @xmath901 with an @xmath5-independent constant @xmath902 ( @xmath903 ) .
[ exactly speaking , both @xmath387 and @xmath904 can not follow the simple drude forms at the same time when @xmath366 exhibits an extended drude form .
] this experimental fact has also been theoretically reproduced @xcite .
therefore , the anomalous ( @xmath5 , @xmath15)-dependences of @xmath840 and @xmath837 in htscs can be semiquantitatively explained by the flex+cvc method a microscopic theory without any fitting parameters except for @xmath23 . at the present stage ,
the flex+cvc method is the only theory with the capability to explain the anomalous ac and dc transport coefficients in a unified way .
we briefly comment on the carrier - doping dependence of @xmath905 : the experimental value of @xmath906 for optimally - doped ybco ( @xmath907 ) is 0.15 [ @xmath908 and it increases to 0.3 [ @xmath908 in slightly under - doped ybco and bscco ( @xmath909 )
. this experimental doping dependence can be quantitatively reproduced by using the flex+cvc method for @xmath910 @xcite . however
, this method can not explain the large experimental value @xmath911 [ @xmath908 in heavily under - doped ybco ( @xmath755 ) @xcite .
this experimental fact will give us an important hint to understand the essential electronic properties in heavily under - doped htscs .
recently , the infrared hall coefficient in pcco ( @xmath912 ) was measured for @xmath913 @xmath857 @xcite .
the observed @xmath5-dependences for @xmath914 can not be explained by simple extended drude models .
they show that the experimental results can be understood by assuming a sdw gap @xmath915 @xmath857 .
it is an important future problem to study the role of the cvc in the infrared hall coefficient in pcco .
in earlier sections , we studied the transport phenomena in nearly af metals without randomness . by including cvcs , we succeeded in explaining the anomalous dc and ac transport phenomena in a unified way , in both hole- and the electron - doped htscs .
however , we have neglected the impurity or disorder effects on the transport phenomena , although they are prominent in real under - doped htscs . in fact , the residual resistivity due to disorder increases drastically as the system approaches the half - filling @xcite . moreover , stm / sts measurements revealed that the electronic states in under - doped htscs are highly inhomogeneous at the nanoscale , reflecting the random potential induced by the disordered atoms outside of the cuo@xmath50 plane @xcite .
these experimental facts are reproduced by assuming that the sc pairing potential is strongly influenced by the impurity potential @xcite further , the anomalous transport phenomena near the af qcp are sensitive to randomness . for example , within the born approximation , the cvc due to electron - electron interaction vanishes at zero temperature since only elastic scattering exists at @xmath97 .
in fact , the cvc term in eq .
( [ eqn : bs ] ) vanishes at @xmath97 if we replace @xmath916 with @xmath917 .
however , the born approximation is applicable only when the impurity potential is weak .
in fact , we will show that a `` strong '' local impurity potential in under - doped htscs induces drastic and widespread changes in the electron - electron correlation around the impurity site .
thus , the impurity effect in htscs depends on the strength of the impurity potential .
hereafter , we study the nontrivial impurity effects in nearly af fermi liquids .
first , we study the effect of `` weak local impurities '' on the transport phenomena within the born approximation , where the quasiparticle damping rate due to impurity scattering is given by @xmath918 where @xmath919 is the impurity density , @xmath920 is the impurity potential , and @xmath196 is the dos at the chemical potential .
then , the total quasiparticle damping rate @xmath921 is given by @xmath922 where @xmath923 represents the quasiparticle damping rate due to inelastic scattering , which is given by eq .
( [ eqn : gamma ] ) in the fermi liquid theory or by eq .
( [ eqn : gam - flex - true ] ) in the flex approximation .
( here , we ignored the self - energy correction given by the cross terms between @xmath23 and @xmath920 .
this simplification is not allowed for a large @xmath920 , as we will show in [ imp - strong ] . ) also , the cvc due to local impurities vanishes identically within the born approximation .
therefore , in the case of @xmath917 , the bethe - salpeter equation ( [ eqn : bs ] ) is changed to become @xmath924 where @xmath925 .
thus , an approximate solution of eq .
( [ eqn : bsimp ] ) is @xmath926 in the absence of impurities , @xmath393 exhibits singular @xmath21-dependence because @xmath927 approaches one . in the case of @xmath928 , on the other hand , @xmath929 and therefore @xmath930 . as a result ,
the cvc is strongly suppressed by high density `` weak local impurities '' , when the elastic scattering is comparable to the inelastic scattering due to af fluctuations .
-dependence of ( i ) @xmath6 and ( ii ) @xmath9 for ncco in the presence of elastic scattering due to weak local impurities . in the case of @xmath931
, @xmath9 becomes positive at low temperatures since the cvc is reduced by elastic scattering . ]
-dependence of cot@xmath932 for lsco for @xmath933 .
cot@xmath934 is realized well , and @xmath282 is proportional to @xmath255 . ] in ref .
@xcite , we have pointed out that the enhancement of @xmath9 due to the cvc is easily suppressed by the weak impurities .
figure [ fig : born ] shows the impurity effects on @xmath6 and @xmath9 obtained by the flex+cvc method for electron - doped systems . as expected , @xmath9 is suppressed by a small amount of impurities , although the induced residual resistivity is small .
( note that @xmath935 corresponds to 250 @xmath936 cm . ) as a result , @xmath9 becomes positive at low temperatures in the presence of impurities , which is consistent with the experimental behavior of @xmath9 in pcco for @xmath937 @xcite .
finally , we consider the case where the impurity potential is widespread ( nonlocal ) . in this case , the impurity potential @xmath938 is momentum - dependent , and the kernel of the bethe - salpeter equation in eq .
( [ eqn : bsimp ] ) , @xmath939 , is replaced with @xmath940 .
when @xmath938 is large only for @xmath941 ( forward scattering ) , both the residual resistivity and the reduction in @xmath9 due to the impurities will be small . in htscs , it is considered that impurities outside of the cuo@xmath50 plane causes the forward impurity scattering @xcite .
figure [ fig : born2 ] shows the impurity effects on cot@xmath932 obtained by the flex+cvc method for lsco .
the relationship cot@xmath934 holds well , and @xmath942 .
such a parallel shift of cot@xmath904 by impurity doping is observed in various hole - doped htscs @xcite .
now , we discuss the effect of `` strong local impurities '' in htscs . according to a recent lda study @xcite , a zn atom introduced in the cuo@xmath50 plane of htscs induces a large positive potential ( less than 10ev ) , and
the potential radius is only @xmath943 . in htscs
, however , zn doping causes a nontrivial widespread change of the electronic states . in zn - doped ybco compounds
, site - selective @xmath944y nmr measurements revealed that both local spin susceptibility @xcite and staggered susceptibility @xcite are prominently enhanced around the zn site , within a radius of the af correlation length @xmath108 .
the same result was obtained by the @xmath945li knight shift measurement in li - doped ybco compounds @xcite , and by the @xmath946cu nmr measurement in zn - doped ybco compounds @xcite . moreover
, a small concentration of zn induces a huge residual resistivity , which is significantly greater than the @xmath947-wave unitary scattering limit @xcite .
these nontrivial impurity effects were frequently considered as evidence for the breakdown of the fermi liquid state in under - doped htscs . up to now , many theorists have studied this important issue .
the single - impurity problem in a cluster @xmath32-@xmath33 model has been studied by using the exact diagonalization method @xcite .
when the number of holes is two , both af correlations and electron density increase near the impurity site .
this problem was also studied by using the extended gutzwiller approximation @xcite .
although these methods of calculation are founded , the upper limit of the cluster size is rather small .
moreover , these studies are restricted to @xmath97 .
in this section , we study a single - impurity problem in a large size square lattice ( say @xmath948 ) hubbard model based on the nearly af fermi liquid theory @xcite : @xmath949 where @xmath920 is the local impurity potential at site @xmath950 .
it is a difficult problem since we have to consider two different types of strong interactions ( @xmath23 and @xmath920 ) on the same footing .
moreover , the absence of translational symmetry severely complicates the numerical analysis . to overcome these difficulties
, we developed the @xmath951 method a powerful method for calculating the electronic states in real space in the presence of impurities @xcite .
the @xmath951 method is applicable for finite temperatures since the thermal fluctuation effect is taken into account appropriately .
based on the @xmath951 method , we successfully explain the nontrivial impurity effects in htscs _ in a unified way _ without assuming any exotic mechanisms . in the @xmath951 method ,
the real - space spin susceptibility @xmath952 is given by @xmath953 here , @xmath954 is given by solving the following dyson equation ; @xmath955 where @xmath956 , and @xmath957 is the real - space green function in the flex approximation without the impurity potential @xmath920 .
the solution of eq .
( [ eqn : gi ] ) is @xmath958 where the position of the impurity potential is @xmath950 .
method in the presence of impurities ( @xmath959 ) for the impurity potential @xmath960 .
a single impurity induces a large curie - term @xmath961 , which is consistent with the experimental results . *
( ii ) * the local susceptibility @xmath962 and staggered one @xmath963 given by the @xmath951 method .
the inset describes the state in which both local and staggered susceptibilities are enhanced around the impurity site . ] in fig .
[ fig : gvi - chi ] , we show the numerical results for spin susceptibilities , @xmath964 , for @xmath960 .
surprisingly , a nonmagnetic impurity induces a huge curie - like component in the uniform susceptibility ; @xmath965 .
this result explains the long - standing experimental problem @xcite .
the obtained value of @xmath966 is @xmath967 , which is close to the experimental value @xmath968 in yba@xmath50cu@xmath51o@xmath969 ( @xmath970 ) @xcite and in lsco ( @xmath971 ) @xcite .
we also emphasize that both the local and staggered susceptibilities get enhanced around the impurity site within a radius of about 3 @xmath972 at @xmath152 . here
, we discuss the physical reason for this drastic impurity effect . in the flex approximation , the af order ( in the mean - field level ) is suppressed by the self - energy correction ( see [ sf - theory ] ) , which represents the destruction of the long - range order due to thermal and quantum fluctuations . in the flex approximation , the reduction in the dos due to the large quasiparticle damping rate @xmath973 renormalizes the spin susceptibility . around the impurity site
, the self - energy effect due to electron - electron correlation is expected to be smaller .
in fact , quantum fluctuation is reduced near the vacant site in the @xmath974 heisenberg model @xcite .
for this reason , in the @xmath951 method , the af correlations are prominently enhanced around the impurity site due to the reduction of thermal and quantum fluctuations .
the examples of the cross terms between @xmath23 and @xmath920 for susceptibility are shown in fig .
[ fig : gvi - diag ] ( i ) .
the same physics occurs in quantum spin systems with vacancies ; af spin correlations are enhanced around a vacancy , due to the reduction in quantum fluctuations @xcite . prior to the @xmath951 method
, the single impurity problem in the hubbard model had been studied using the rpa @xcite , assuming a nonlocal impurity potential ; @xmath975 for the onsite and @xmath976 for the site adjacent to the impurity atom .
the obtained result strongly depends on the value of @xmath976 : when @xmath977 , the enhancement of susceptibility is tiny or absent although @xmath978 for the zn impurity atom in the cuo@xmath50-plane according to the lda study @xcite . on the other hand , in the @xmath951 method , the considerable enhancement of the susceptibility is realized even when @xmath977 ( local impurity potential case ) , since @xmath109 is renormalized by the self - energy correction in the flex approximation : then , @xmath109 is easily enhanced when the self - energy effect is reduced by introducing an impurity .
therefore , we have to use the @xmath951 method for investigating the impurity problem in htscs . in the presence of a single - impurity potential @xmath920 in the hubbard model .
they are cross terms between @xmath920 and @xmath23 .
the cross represents @xmath920 , and the large circle with the shadow represents the self - energy given by the flex approximation for @xmath979 . * ( ii ) * the local dos given by the @xmath951 method : it decreases around the impurity site within the radius of about @xmath980 , due to the nonlocal `` effective complex impurity potential '' , @xmath981 . ]
next , we discuss the transport phenomena in the presence of dilute impurities according to the @xmath951 method @xcite .
the enhanced susceptibilities due to the impurity induce the `` additional self - energy correction @xmath982 '' around the impurity site .
if the area of @xmath983 is large , a large residual resistivity will be induced by the non-@xmath947-wave scattering channels . in the @xmath951 method , @xmath984 is given by @xmath985 where @xmath986 is the self - energy given by the flex approximation without the impurity potential ( @xmath979 ) .
the green function @xmath987 is given by solving the following dyson equation in real space : @xmath988 in the @xmath951 method , we solve eqs .
( [ eqn : self - gvi ] ) and ( [ eqn : ds - gvi ] ) self - consistently . the local dos given by the @xmath951 method in real space
@xmath989 is shown in fig .
[ fig : gvi - diag ] ( ii ) .
the local dos decreases around the impurity site within the radius of approximately @xmath990 ( @xmath121 is the lattice spacing ) , since the quasiparticle lifetime is very short due to the large im@xmath991 .
we verified that the radius of im@xmath991 increases as the filling number @xmath11 approaches unity , in proportion to @xmath108 .
as explained in ref .
@xcite , we should not solve @xmath992 in eq .
( [ eqn : self - vi ] ) self - consistently , since the feedback effect on @xmath992 introduced by iteration is canceled by the vertex correction for @xmath992 that is absent in the flex approximation . to derive the resistivity in the presence of impurities , we have to obtain the @xmath32-matrix , @xmath993 , which is defined as @xmath994 . the expression of @xmath993 in the case of @xmath995 is derived in ref .
@xcite . using the @xmath32-matrix ,
the quasiparticle damping rate due to the impurity is given by @xcite @xmath996 where @xmath997 .
@xmath919 is the density of the impurities . in the case of @xmath998 ,
the total quasiparticle damping rate is @xmath999 , where @xmath1000 .
the resistivity for @xmath998 is approximately given by @xmath1001 .
figure [ fig : gvi - rho ] ( i ) shows @xmath21-dependences of @xmath1002 ( @xmath1003 ) and @xmath1004 .
we emphasize that @xmath1002 exhibits strong @xmath21-dependence that is similar to the @xmath21-dependence of @xmath1005 .
this result suggests that the effective impurity potential @xmath984 is nonlocal . as a result ,
the structure of the `` hot spot '' and the `` cold spot '' , is not smeared out by the strong non - magnetic impurities , although it will be smeared out by the weak local impurities .
this finding strongly suggests that the enhancement of the hall coefficient near the af qcp , which is induced by the strong cvc around the cold spot @xcite , does not decrease due to the strong impurities .
moreover , the enhancement of @xmath1006 due to @xmath919 will also enhance @xmath9 .
experimentally , @xmath9 for under - doped ybco and bi@xmath1007sr@xmath1008la@xmath1009cuo@xmath53 are approximately independent of the doping of zn and other non - magnetic impurities @xcite , whereas @xmath9 for lsco decreases with zn doping @xcite . for @xmath1003 and @xmath1004 for ybco and lsco at @xmath152 .
@xmath21-dependent @xmath1002 obtained using the @xmath951 method will be verified by the arpes measurements for zn - doped compounds . *
( ii ) * @xmath6 in ybco ( @xmath189 ) for @xmath1010 given by the @xmath951 method . at lower temperatures ,
an insulating behavior is observed in the close vicinity of the af qcp . in htscs , @xmath137 and @xmath935
correspond to 400k and 250@xmath1011 cm , respectively . *
( iii ) * the filling dependence of @xmath1012 per @xmath919 at @xmath152 given by the @xmath951 method .
note that @xmath1013 $ ] in a 2d electron - gas model due to strong local impurities . ]
figure [ fig : gvi - rho ] ( ii ) shows the resistivity @xmath6 for ybco with @xmath10140.02 for @xmath1015 , and @xmath1016 .
the obtained impurity effect is the most prominent when @xmath1017 .
as @xmath15 decreases , nonmagnetic impurities cause a `` kondo - like upturn '' of @xmath6 below @xmath1018 , reflecting an extremely short quasiparticle lifetime near the impurities .
in other words , _ the considerable residual resistivity is caused by the large scattering cross section of the `` effective impurity potential '' , @xmath991 _ , as shown in fig .
[ fig : gvi - diag ] ( ii ) .
note that @xmath1018 increases with @xmath919 .
this amazing result strongly suggests that the insulating behavior of @xmath6 observed in under - doped lsco @xcite and ncco @xcite is caused by the residual disorder in the cuo@xmath50 plane . as shown in fig .
[ fig : gvi - rho ] ( iii ) , the parallel shift of the resistivity at finite temperatures due to impurities ( @xmath1019 ) grows drastically in the under - doped region ; @xmath1019 exceeds the @xmath947-wave unitary scattering limit in the 2d electron gas model , @xmath1020 .
this result effectively explains experimental carrier - doping dependence of @xmath1019 @xcite .
the enhancement of the residual resistivity is also observed near the af qcp in heavy - fermion systems such as ceal@xmath51 @xcite and cecu@xmath3au @xcite , and in the organic superconductor @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 @xcite . in these compounds , @xmath1019 quickly decreases with pressure , as the distance from the af qcp increases . in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 , @xmath1019 under 2
gpa is six times smaller than the value at 0.5 gpa .
such a large change in @xmath1019 is difficult to be explained by the pressure dependence of the dos .
in fact , according to the @xmath32-matrix approximation , @xmath1021 .
the residual resistivity in 2d free dispersion model ( @xmath1022 ) is @xmath1023 .
therefore , @xmath1019 is given by @xmath1024 in 2d free dispersion model .
note that the renormalization factor @xmath118 does not appear in the expression of @xmath1019 . in case of @xmath1025 , eq .
( [ eqn : rimp - t ] ) gives the @xmath947-wave unitary scattering value ; @xmath1026 . in case of weak impurity scattering where @xmath1027 ( born limit ) , we obtain the relation @xmath1028 , which will decrease under pressure since @xmath1029
. however , pressure dependence of @xmath1019 in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 seems too strong to be explained by the born approximation .
we comment that the increment in @xmath1019 due to charge fluctuations is discussed in some heavy - fermion systems @xcite . in summary
, the present study revealed that a single impurity strongly influence the electronic states in a wide area around the impurity site near the af qcp . using the @xmath951 method , the characteristic impurity effects in under - doped htscs
are well explained in a unified way , without assuming any exotic non - fermi - liquid ground states .
we successfully explain the nontrivial impurity effects in htscs _ in a unified way _ in terms of a spin fluctuation theory , which strongly suggests that the ground state of htscs is a fermi liquid .
we expect that the novel impurity effects in other metals near the af qcp , such as heavy - fermion systems and organic metals , will be explained by the @xmath951 method .
the validity of the @xmath951 method is verified in ref .
@xcite based on the microscopic fermi liquid theory .
recently , we extended the @xmath951-flex method to be applicable even in below @xmath1 , and analyzed the hubbard cluster model given by eq .
( [ eqn : gvi - model ] ) .
we found that both local and staggered susceptibilities are enhanced around the impurity site even in the sc state @xcite .
this result is consistent with @xmath1030o nmr measurement for ybco with zn impurities @xcite .
interestingly , it is also reproduced by analyzing the hubbard model together with the @xmath64-wave pairing interaction using the mean - field approximation @xcite . in the normal state , on the other hand , the enhancement of the local and staggered susceptibilities can not be explained qualitatively within the mean - field approximation or rpa @xcite .
it is an important future issue to explain the sensitive impurity effects on the sc state in htscs @xcite based on the _ repulsive hubbard model _ , using the @xmath951-flex method .
thus far , we showed that the cvc due to strong af+sc fluctuations induces various anomalous transport phenomena in htscs . to validate this idea , we have to study the various nearly af systems other than htscs . in general ,
the electronic structure of heavy - fermion systems and organic metals are very sensitive to the pressure .
therefore , the distance from the af qcp can be easily changed by applying the pressure , without introducing disorders in the compounds .
this is a great advantage with respect to investigating the intrinsic electronic states near the af qcp , free from the disorder effects .
a useful theoretical review for heavy - fermion systems near the af qcp is given in ref .
recently , detailed measurements of the transport phenomena under pressure have been performed in the heavy - fermion superconductor ce@xmath2in@xmath3 ( @xmath2=co or rh ) and in the organic superconductor @xmath4-(bedt - ttf ) .
they exhibit striking non - fermi - liquid - like behaviors as observed in htscs eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) .
hereafter , we explain the experimental and theoretical studies on the transport phenomena in these systems . in ( a ) cerhin@xmath3 for @xmath1031 gpa , and in ( b ) cecoin@xmath3 for @xmath1032 .
@xcite ] ] plotted as a function of @xmath1033 for cerhin@xmath3 .
@xcite ] ] ce@xmath2in@xmath3 is a quasi 2d heavy - fermion superconductor with @xmath1034k . according to the angle - resolved measurements of thermal conductivity @xcite and specific heat @xcite ,
the symmetry of the sc gap is @xmath64-wave .
figure [ fig : nakajima ] shows the temperature dependence of @xmath9 in cecoin@xmath3 and cerhin@xmath3 in the limit of @xmath1035 . in cerhin@xmath3 , the critical pressure under which the af qcp is realized is @xmath1036gpa . at @xmath1037 ,
@xmath1038 reaches 50 , whereas the magnitude of @xmath9 rapidly decreases as the pressure is increased .
a similar pressure dependence of @xmath9 is observed in cecoin@xmath3 , where @xmath1039 is slightly below the ambient pressure . in both these compounds
, @xmath9 is inversely proportional to @xmath15 at higher temperatures .
moreover , as shown in fig .
[ fig : nakajima2 ] , the modified kohler s rule given in eq .
( [ eqn : mkr - the ] ) is well satisfied in cerhin@xmath3 @xcite for over four orders of magnitude .
furthermore , this is well satisfied in cecoin@xmath3 @xcite .
therefore , both @xmath9 and @xmath12 in ce@xmath2in@xmath3 shows anomalous behaviors which are similar to htscs .
this experimental fact strongly suggests that the cvc due to the strong af fluctuations is the origin of the anomalous transport phenomena .
now , we discuss the magnetic field dependence of the transport coefficients .
interestingly , @xmath1040 in ce@xmath2in@xmath3 near @xmath1039 is easily suppressed only by a small magnetic field , as shown in the inset of fig .
[ fig : nakajima ] . at the same time , @xmath1041 is also significantly suppressed .
therefore , the relationships @xmath1042 and @xmath1043 are satisfied only below @xmath1044 tesla near the qcp @xcite .
these behaviors can not be attributed to the orbital effect ( i.e. , the cyclotron motion of conduction electrons ) since @xmath1045 is satisfied only when @xmath1046 tesla and @xmath1047 k @xcite . the condition @xmath1048 is also recognized from the relationship @xmath1049 for @xmath1050k and @xmath1051 tesla , as shown in fig .
[ fig : nakajima2 ] .
when @xmath1052 is satisfied , we can safely expand nakano - kubo formula given in eq .
( [ eqn : nakano - kubo ] ) with respect to the vector potential , as we did in [ flex - cvc ] - [ imp ] . then , the following relationships derived in [ flex - cvc ] @xmath1053 are expected to be valid for ce@xmath2in@xmath3 . here , the factors @xmath111 and @xmath1054 in eqs .
( [ eqn : rhoh ] ) and ( [ eqn : drr0 ] ) , respectively , come from the cvcs .
therefore , experimental striking non - linear behaviors of @xmath9 and @xmath12 with respect to @xmath392 should originate from the field - dependence of @xmath1055 : the suppression of @xmath108 due to the magnetic field results in reducing both eqs .
( [ eqn : rhoh ] ) and ( [ eqn : drr0 ] ) .
since the correlation length is sensitive to the outer parameters in the vicinity of the qcp @xcite , the anomalous sensitivity of @xmath1056 and @xmath12 to the magnetic field in ce@xmath2in@xmath3 originates from the field - dependence of the cvc .
recently , @xmath1056 and @xmath12 in pcco were measured in magnetic field up to 60 t , and it is found that both of them shows striking non - linear behaviors with respect to @xmath392 @xcite .
their behaviors will be explained by the field - dependence of the cvc . even if the field dependence of cvc is prominent , the modified kohler s rule in eq . (
[ eqn : mkr - the ] ) should be satisfied for a wide range of the magnetic field strength , since both @xmath12 and @xmath1057 are proportional to @xmath1058 if the preset theory of cvc is correct .
in fact , the modified kohler s rule is well satisfied for @xmath1059k and @xmath1060 tesla in cerhin@xmath3 as shown in fig .
[ fig : nakajima2 ] , regardless of the fact that the conventional kohler s rule is violated only for 0.1 tesla .
this fact strongly suggests that both @xmath387 and @xmath589 are enhanced by the same origin , namely , the cvc due to the af fluctuations .
therefore , the anomalous transport phenomena in cecoin@xmath3 are consistently described by the theory of cvc in nearly af fermi liquids .
the nernst signal @xmath22 in cecoin@xmath3 is also very anomalous @xcite .
below 20k , @xmath22 starts to increase approximately in proportion to @xmath825 , exhibiting anomalously large values ( @xmath1061v / kt ) below 4k .
this behavior is very similar to that of @xmath22 in electron - doped htsc , whose @xmath15-dependence and the magnitude are well reproduced by the cvc due to strong af fluctuations .
it reaches @xmath1062v / kt in optimally - doped ncco . considering that @xmath22 is proportional to @xmath1063 , the experimental relations @xmath1064v / kt@xmath1065v / kt@xmath1066 and @xmath1067 are naturally understood as the quasiparticle transport phenomena .
thus , the giant nernst signal in cecoin@xmath3 is expected to be caused by the cvc due to strong af fluctuations .
although the spin susceptibility @xmath1068 in ce@xmath2in@xmath3 has a quasi 2d structure , the anisotropy of resistivity is only two ( @xmath1069 ) at low temperatures . in the dynamical - mean - field - theory ( dmft ) @xcite , which is believed to effectively describe the electronic properties in 3d systems
, the cvc vanishes identically .
therefore , it is highly desired to confirm whether the cvc can be really significant in 3d systems or not .
recently , we performed the numerical study for 3d hubbard model with the conduction electron spectrum is @xmath1070 , where @xmath126 is given in eq .
( [ eqn : band ] ) .
the obtained @xmath6 , @xmath9 and @xmath22 in quasi 2d hubbard model are shown in fig .
[ fig : onari ] ( i)-(iii ) .
since the unit cell lengths in ce@xmath2in@xmath3 are @xmath1071 and @xmath1072 , @xmath935 and @xmath1073 correspond to 300 @xmath936 cm and @xmath1074 m@xmath1075 , respectively .
we find that both @xmath9 and @xmath22 are strikingly enhanced due to the cvc even if @xmath1076 . in cecoin@xmath3
, @xmath9 starts to increase below @xmath1077k , which corresponds to @xmath1078 in fig .
[ fig : onari ] ( ii ) . at @xmath152 ( @xmath6038k for ce@xmath2in@xmath3 ) ,
@xmath1079 for @xmath1080 .
then , the extrapolated value of @xmath9 at 2k is 20 .
as shown in fig .
[ fig : onari ] ( iii ) , the obtained @xmath22 is drastically enhanced by the cvc , which is consistent with experiments . to derive the correct absolute value of @xmath22
, we have to take account of the fact that @xmath22 is proportional to the mass - enhancement factor @xmath847 . according to the de haas - van alphen measurement , @xmath1081 in cecoin@xmath3 .
since the mass - enhancement factor in the present flex approximation is @xmath1082 , we describe the experimental mass - enhancement factor as @xmath1083 , where @xmath1084 is the mass enhancement factor which can not be described by the flex approximation in the hubbard model .
[ even in the flex approximation , relatively large @xmath1085 is obtained in the periodic anderson model , which is an effective model for heavy fermion systems @xcite .
we will comment on this fact in [ validityflex ] .
] we present @xmath1086 in fig .
[ fig : onari ] ( iii ) , where @xmath807 is given by the flex approximation , by using the relation @xmath1087 nv@xmath1088kt .
recently , we extended the flex approximation to reproduce appropriate results under a finite magnetic field , and calculated the field dependence of @xmath9 and @xmath22 .
it was suggested that both these quantities were rapidly suppressed by the magnetic field near af qcp , reflecting the reduction of af fluctuations .
the obtained result is in good agreement with the experimental results . , *
( ii ) * @xmath9 and * ( iii ) * @xmath22 for the 3d hubbard model ( @xmath1089 ) given by the flex+cvc method .
broken lines represent the numerical results for @xmath1090 , @xmath1091 and @xmath1092 . in these parameters , the stoner factors @xmath1093 is 0.995 at @xmath152 .
full lines represents the numerical results for @xmath23s with which @xmath1094 is satisfied at @xmath152 . *
( iv ) * fermi surfaces for @xmath1095 . [ ref .
@xcite ] ] in cecoin@xmath3 , @xmath9 exhibits a peak at @xmath1096k at ambient pressure , and @xmath1097 increases with pressure [ see fig . [
fig : nakajima ] ] .
this is different from the pseudo - gap behavior in under - doped htscs since @xmath1097 in cecoin@xmath3 increases as the system goes away from the af - qcp . in particular , @xmath9 of cerhin@xmath3 at @xmath1037 maintains its increasing trend just above @xmath1 .
these behaviors can be understood as the effect of the weak ( local ) residual disorders , as we have discussed in [ rh - imp - weak ] : as shown in fig .
[ fig : born ] ( ii ) , @xmath1098 starts to decrease at lower temperatures when @xmath1099 .
the peak temperature @xmath1097 increases with @xmath1100 . in cecoin@xmath3 ,
the resistivity at @xmath1101 is @xmath1102 cm @xcite for @xmath1103 gpa . on the other hand , in cerhin@xmath3 at @xmath1104 gpa , @xmath1105 cm because of the large inelastic scattering . since @xmath1100 in cecoin@xmath3 and
that in cerhin@xmath3 are expected to be similar , @xmath1106 in eq .
( [ eqn : j - ap - imp ] ) will be smaller in cecoin@xmath3 .
that is , reduction of cvc due to impurities is more prominent in cecoin@xmath3 .
therefore , we can explain the different behaviors of @xmath9 in cerhin@xmath3 and cecoin@xmath3 at low temperatures as the effect of residual disorders .
one may ascribe the temperature dependence of @xmath9 in ce@xmath2in@xmath3 to the multiband effect .
for example , a sign change in @xmath9 can occur if a hole - like fermi surface and an electron - like fermi surface coexist and their mean free paths have different @xmath15-dependences .
however , it is very difficult to explain the relation @xmath46 in a multiband model based on the rta .
in fact , in order to explain the pressure dependence of @xmath9 in cecoin@xmath3 , one has to assume that a small fermi surface governs the transport phenomena at 0gpa ( near af qcp ) , whereas a large fermi surface should consequently govern the transport phenomena under 2.5gpa .
the same drastic change in the electronic states should occur by applying a magnetic field @xmath1107tesla .
this unnatural assumption is not true since the other transport coefficients can not be explained at all , as discussed in ref .
@xcite in detail .
in particular , the elegant modified kohler s rule plot for cerhin@xmath3 shown in fig . [
fig : nakajima2 ] _ for over four orders of magnitude _ is strong evidence that the cvc is the origin of the anomalous transport phenomena , and the anomalous transport phenomena are mainly caused by a single large fermi surface with heavy quasiparticles . since the enhancements of both @xmath9 and @xmath12 originate from a small portion of the fermi surface ( i.e. , the cold spot ) as shown in fig .
[ fig : rh ] ( ii ) , the multiband effect is unimportant in ce@xmath2in@xmath3 . therefore , the modified kohler s rule will be realized near the af qcp even in multiband systems like ce@xmath2in@xmath3 . in usual heavy - fermion compounds
, @xmath9 exhibits a curie - like behavior above the coherent temperature @xmath775 due to the large anomalous hall effect , which is caused by the angular momenta of the @xmath17-orbitals as we will discuss in [ ahe - kw ] . for @xmath1108 , @xmath9 due to the ahe is proportional to @xmath1109 . in heavy - fermion compounds near the af qcp , on the other hand , both the ordinary hall coefficient @xmath1110 and the anomalous hall coefficient @xmath1111 can show large temperature dependences .
paschen et al .
extracted @xmath1110 from the experimental hall coefficient in ybrh@xmath50si@xmath50 , and discussed the critical behavior of @xmath1110 near af qcp , which is realized under the magnetic field @xcite .
in contrast , @xmath9 in ce@xmath2in@xmath3 ( @xmath2=co , rh ) is almost constant above 50 k as shown in fig .
[ fig : nakajima ]
. therefore , anomalous hall effect in ce@xmath2in@xmath3 is very small , and therefore @xmath1112 as explained in detail in ref .
this fact is a great advantage to study the anomalous @xmath15-dependence of the ordinary hall effect in ce@xmath2in@xmath3 . in cecoin@xmath3 ,
novel kinds of critical behaviors are observed near @xmath1113 tesla for @xmath1114 k , where prominent increment in the effective mass was observed .
this phenomenon is referred to as a `` field - induced qcp '' .
one possible origin will be the field - induced sdw state that is hidden in the sc state .
we will discuss this important future problem in
[ validityflex ] in more detail . and ( ii ) @xmath1115 for @xmath4-(bedt - ttf)@xmath50cu(ncs)@xmath50 under pressure .
@xcite ] ] the measurements of @xmath9 under pressure have been intensively performed in @xmath4-(bedt - ttf)@xmath50x .
figure [ fig : taniguchi ] shows the temperature dependence of @xmath9 and @xmath1115 for x = cu(ncs)@xmath50 under homogeneous pressure . here ,
the af correlations are the strongest at the lowest experimental pressure ( 0.19gpa ) . as the pressure is increased , the af fluctuations get reduced and a conventional fermi liquid state is realized . at the same time , @xmath9 is reduced and exhibits a constant value .
the observed @xmath9 is independent of pressure at room temperature .
therefore , the origin of the increment in @xmath9 below 100 k can not be related to deformation of the fermi surface under pressure .
note that similar increment in @xmath9 is observed in x = cu[n(cn)@xmath50]cl , which has a single elliptical fermi surface @xcite .
because the observed @xmath15 dependences of @xmath9 and @xmath1115 are very similar to the observations involving htscs and in ce@xmath2in@xmath3 , the cvc is expected to play a significant role in @xmath4-(bedt - ttf)@xmath50x . as discussed in [ hotcold ] , the resistivity in 2d systems in the presence of af fluctuations is @xcite , @xmath1116 this relationship is reliable when @xmath123 , which is satisfied in optimally- or over - doped htscs .
the @xmath15-dependence of @xmath1117 is given in eq .
( [ eqn : parameters2 ] ) .
we consider eq .
( [ eqn : ky ] ) is realized in x = cu(ncs)@xmath50 since the af fluctuations are not so prominent . since @xmath9 is proportional to @xmath111 as shown in eq .
( [ eqn : rh - scaling ] ) , we obtain @xmath1118 figure [ fig : taniguchi ] ( ii ) shows that eq .
( [ eqn : ha ] ) is well satisfied below 80k , except at @xmath1119 gpa .
[ the thermal contraction of the sample might modify this relationship at the lowest pressure . ]
the success of the scaling relationships ( [ eqn : rh - scaling ] ) , ( [ eqn : ky ] ) and ( [ eqn : ha ] ) is strong evidence that the enhancement of @xmath9 in @xmath4-(bedt - ttf)@xmath50x is caused by the cvc due to af fluctuations . in many @xmath4-(bedt - ttf)@xmath50x compounds ,
the phase transition between the mott insulating phase and the metallic ( sc ) phase induced by pressure is weak first order .
therefore , the af fluctuations are not so strong even in the vicinity of the af insulating phase . as a result ,
the enhancement of @xmath9 in @xmath4-(bedt - ttf)@xmath50x is much smaller than that in ce@xmath2in@xmath3 . on the contrary ,
carrier doped ( 11% doping ) @xmath4-type superconductor @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 ( @xmath1120 k at ambient pressure ) exhibits very strong af fluctuations , and therefore the critical pressures is expected to be slightly below 0 kbar . according to nmr measurement , @xmath1121 $ ] increases with decreasing temperature above @xmath1122 k , in proportion to @xmath1123 with @xmath1124 k @xcite .
the @xmath65-@xmath15 phase diagram and transport properties of @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 are carefully measured by taniguchi et al .
@xcite : it was found that @xmath1 shows two - peak structure under pressure . also , the power @xmath11 in @xmath1125 increases from 1 to 2 with increasing pressure , which indicates that the af fluctuations are suppressed under the pressure .
taniguchi et al . also measured the @xmath9 in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 under pressures , and found that @xmath9 exhibits a curie - weiss temperature dependence ; @xmath1126 where @xmath1127 @xcite . at 0.19gpa where af fluctuations are strong
, @xmath1128 reaches 10 , whereas the enhancement of @xmath9 is totally suppressed by 1 gpa .
the behavior of @xmath9 in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 is very similar to the observation in ce@xmath2in@xmath3 .
@xmath1129 decreases with pressure , and its extrapolated value to 0 gpa ( @xmath1130 k ) coincides with the weiss temperature of @xmath59 at ambient pressure .
this is strong evidence that the hall coefficient in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 is proportional to @xmath111 , due to the cvc induced by af fluctuations .
it is noteworthy that the residual resistivity @xmath1019 in @xmath4-(bedt - ttf)@xmath49hg@xmath62br@xmath63 drastically decreases with pressure @xcite . as we have discussed in
[ imp - strong ] , the large residual resistivity near 0 gpa is expected to be given by the enlarged effective impurity potential due to the many - body effect .
note that @xmath9 in @xmath4-(bedt - ttf)@xmath50cu[n(cn)@xmath50]br changes to be negative under high pressures and low temperatures , when the electron - electron correlation becomes very weak .
the origin of this experimental fact is considered to be the lattice deformation with the long - period structure inherent in this compound @xcite .
-@xmath15 phase diagram for the anisotropic triangular lattice hubbard model [ @xmath1131 given by the flex approximation .
the large ( small ) dots represent the @xmath64-wave @xmath1 ( @xmath67 ) . *
( ii ) * @xmath9 given by the flex+cvc method .
[ refs . @xcite and @xcite ] ] here , we analyzed the anisotropic triangular lattice hubbard model at half - filling , which is a theoretical effective model for @xmath4-(bedt - ttf)@xmath50x .
figure [ fig : rh - kappa ] ( i ) shows the numerical results for the @xmath23-@xmath15 phase diagram given by the flex approximation . here ,
@xmath1132 and @xmath1133 are the nearest neighbor integrals of the triangular lattice ; @xmath1134 is for two of three axes and @xmath193 is for the remaining axis , respectively . in many @xmath4-(bedt - ttf)@xmath50x compounds ,
then , the corresponding fermi surface is hole - like and ellipsoidal around the @xmath361-point .
hereafter , we substitute @xmath1136 k. then , the af ordered state appears at @xmath113724k when @xmath23 is as large as the bandwidth ( @xmath1138 ) . as @xmath23 decreases , @xmath67 decreases and the @xmath36-wave sc appears next to the af phase .
the obtained @xmath23-@xmath15 phase diagram well explains the experimental @xmath65-@xmath15 phase diagram ( kanoda diagram @xcite ) , since the increment in @xmath65 corresponds to the reduction in @xmath1139 .
figure [ fig : rh - kappa ] ( ii ) also shows the @xmath15-dependence of @xmath9 for @xmath1140 .
we see that @xmath9 increases at low temperatures due to the cvc and it becomes smaller as @xmath23 decreases . at half - filling , obtained enhancement of @xmath9 is not so large since the af fluctuations in this model are rather weak due to the strong geometric frustration ( @xmath1141 ) .
therefore , a consistent understanding of the @xmath65-@xmath15 phase diagram as well as the @xmath65-dependence of @xmath9 are achieved by using the flex+cvc method .
in the present study , we investigated the mechanism of non - fermi - liquid - like transport phenomena in strongly correlated fermi liquids in the presence of strong magnetic fluctuations .
this problem was first realized in the study of htsc , which has been one of the central issues in htsc .
for example , the relaxation time approximation ( rta ) for the highly anisotropic @xmath18 model can not reproduce relationships ( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) at the same time , even for a narrow range of temperatures .
recent experiments have been revealed that these anomalous transport properties are not a special phenomena in htsc , but a universal phenomena in fermi liquids near af qcp . we discussed two typical examples ce@xmath2in@xmath3 ( @xmath2=co or rh ) and @xmath4-(bedt - ttf)@xmath50x .
the ratio @xmath1142 in cerhin@xmath3 reaches @xmath80 , which is much larger than the ratio in htscs . here
, we studied this long - standing problem by developing the microscopic fermi liquid theory . in the rta ,
the momentum and energy conservation laws in quasiparticles scattering are violated , as explained in [ intro ] and [ cvc ] . to overcome this defect , we take account of the cvc to satisfy the conservation laws .
the total current @xmath19 , which is responsible for the transport coefficients , is given by the summation of the quasiparticle velocity @xmath20 and the cvc @xmath1143 . in interacting electron systems
, an excited quasiparticle induces other particle - hole excitations by collisions .
the cvc represents the current due to these particle - hole excitations .
since the cvc is caused by quasiparticle interactions , it can be significant in strongly correlated fermi liquids .
however , the effects of cvc in htscs have not been studied in detail until recently .
since the cvc is totally ignored in the rta , the rta frequently yields unphysical results in strongly correlated systems . in the presence of strong af fluctuations , we find that * ( a ) * quasiparticle damping rate @xmath1144 becomes anisotropic , and the portion of the fermi surface with small @xmath181 ( cold spot ) governs the transport phenomena @xcite . at the same time , * ( b ) * the total current @xmath539 becomes highly anisotropic due to the prominent cvc @xcite .
( a ) and ( b ) should occur simultaneously since both of them are caused by the same origin ; highly anisotropic incoherent scattering due to af fluctuations .
that is , both ( a ) and ( b ) are induced by the strong backward scattering between @xmath21 and @xmath1145 ( @xmath1146 ) , as shown in fig .
[ fig : fs - hotcold ] ( i ) .
mathematically , @xmath179 and the cvc are closely connected by the ward identity near af qcp , that is , the same vertex function @xmath520 appears in eqs .
( [ eqn : gamma ] ) and ( [ eqn : bs - t ] ) . the facts ( a ) and ( b ) naturally explain the enhancement of @xmath9 , @xmath12 , and @xmath22 in nearly af metals in a unified way .
as shown in fig .
[ fig : j ] ( ii ) , the large curvature of the effective fermi surface around the cold spot gives rise to the enhancement of @xmath9 . in
[ beyondone ] , we proved the significance of the cvc beyond the one - loop ( such as the flex ) approximation .
it is also very important to demonstrate to what extent the cvc can reproduce the aspects of anomalous transport properties based on a standard spin fluctuations theory . for this purpose ,
we study @xmath9 , @xmath12 , @xmath14 and @xmath22 based on the flex ( or flex+@xmath15-matrix ) approximation by including cvcs .
the obtained results semiquantitatively reproduce the various experimental facts in slightly under - doped htscs .
the present study strongly suggests that _ the striking deviation from the fermi liquid behaviors ( such as eqs .
( [ eqn : t - linear - rho])-([eqn : mkrexp ] ) ) are ubiquitous in strongly correlated metals near the af qcp , not specific to htscs . _
note that the increment in @xmath9 is also observed in non - sc metals near af phase , such as v@xmath1147o@xmath51@xcite and @xmath1148bi@xmath148ru@xmath50o@xmath1149 ( @xmath1150 = sm or eu)@xcite .
the main results in the present study are as follows : * ( i ) * hall coefficient : due to the cvc , @xmath723 .
the large curvature of the effective fermi surface around the cold spot is the origin of the enhancement of @xmath9 .
in particular , we can explain that the @xmath9 for ncco and pcco are negative even above @xmath1151 k , since the curvature of the `` effective fermi surface '' at point b in fig .
[ fig : j ] is negative , although the curvature of the true fermi surface is positive . *
( ii ) * magnetoresistance : @xmath1152 due to the cvc , which means that the conventional kohler s rule @xmath1153 is violated . on the other hand , the modified kohler s rule @xmath1154 is realized since the same factor @xmath1054 appears in both sides . * ( iii ) * thermoelectric power : for @xmath1155 , @xmath1156 increases as @xmath15 decreases since the @xmath711-dependence of @xmath726 becomes prominent when the af fluctuations are strong .
further , @xmath684 ( @xmath685 ) in many hole - doped ( electron - doped ) systems . *
( iv ) * nernst coefficient : due to the cvc , @xmath22 gradually increases in ncco below the room temperature . with regard to lsco , a rapid increase in @xmath22 in the pseudo - gap region
is well explained by the flex+@xmath15-matrix approximation , due to the cvc induced by the strong af+sc fluctuations . * ( v ) * ac hall effect : the strong @xmath5-dependence of @xmath837 appears due to the frequency dependence of the cvc .
both @xmath15- and @xmath5-dependences of @xmath837 are well reproduced by the flex+cvc method . *
( vi ) * impurity effect : the strong cvc near the af qcp is suppressed by the _ weak _ local impurities . for _ strong _ local impurities , the residual resistivity exceeds the @xmath947-wave unitary scattering value since the local and staggered susceptibilities are strongly enhanced around the impurity sites .
_ results ( i)-(vi ) are given by the same mechanism the hot / cold spot structure and the singular @xmath21-dependence of the total current @xmath19 in nearly af metals , _ which is shown in fig .
[ fig : j ] .
we emphasize that the abovementioned results are not derived from a special defect of the flex approximation .
they are also reproduced by another spin fluctuation theory .
in fact , kanki and kontani @xcite calculated @xmath19 based on the millis - monien - pines model in eq .
( [ eqn : kai_qw ] ) , by assuming a realistic set of parameters for optimally - doped ybco .
the obtained @xmath19 represents the characteristic @xmath21-dependence in nearly af metals as shown in fig .
[ fig : j ] .
moreover , we proved the significance of the cvc near af qcp beyond the one - loop ( such as the flex ) approximation in [ beyondone ] . as a result
, it is general that @xmath9 is prominently enhanced by the cvc when the af fluctuations are strong , independent of the types of the spin fluctuation theories . in this study
, we explained that the cvc plays a significant roles in nearly af metals , based on the microscopic fermi liquid theory . to illustrate this theoretical idea , we performed numerical studies using the flex+cvc method in [ flex - cvc ] - [ imp ] .
since the flex is a conserving approximation , we can perform a reliable study of the transport phenomena by including cvcs . as we discussed in
[ flex ] , the flex approximation can explain characteristic electronic properties in optimally - doped htsc .
for example , the appropriate spin fluctuation @xmath459 and the hot / cold spot structure of quasiparticle damping rate @xmath181 are reproduced satisfactorily . although the flex approximation can not explain the `` strong pseudo - gap behavior '' below @xmath45 k , the flex+@xmath15-matrix approximation can reproduce various transport anomalies below @xmath34 in slightly under - doped systems as discussed in [ flex - t ] . in the flex+@xmath15-matrix approximation
, we take account of the strong sc fluctuations that are induced by af fluctuations .
the success of the flex+@xmath15-matrix method is strong evidence that sc fluctuations are the origin of strong pseudo - gap . here
, we discuss the applicable scope of the flex approximation .
first , we discuss the `` weak pseudo - gap behavior '' below @xmath826 k , where the dos @xcite , the knight shift @xcite and the uniform susceptibility @xcite decrease inversely with the growth of af fluctuations .
the weak pseudo - gap in the dos is shallow and wide in energy .
although weak pseudo - gap behaviors are reproduced by the flex approximation , the obtained gap behaviors are too moderate .
[ since the weak pseudo - gap affects transport phenomena only slightly , flex+cvc approximation can explain anomalous transport phenomena in htsc .
] this failure of the flex will originate from the fully self - consistent determination of the self - energy and the green function .
in fact , although a large weak pseudo - gap in the dos is reproduced at the first iteration stage , it vanishes in the course of iteration .
mathematically , additional self - energy correction introduced by iteration ( i.e. , the feedback effect ) should be canceled by the vertex correction in the self - energy that is absent in the flex approximation ( e.g. , @xmath1157 in fig .
[ fig : cvc - diag - twol ] ( ii ) ) .
the same result is reported in the @xmath1158 approximation : this is a first principle calculation for the self - energy , which is given by the convolution of the green function @xmath521 and the screened interaction @xmath1159 within the rpa .
although the descriptions of the bandwidth reduction and satellite structure in the quasiparticle spectrum are satisfactory in a partially self - consistent @xmath1158 method , they are smeared out in the fully self - consistent @xmath1158 due to the feedback effect by iteration @xcite . to produce a weak pseudo - gap successfully ,
it is better to derive the self - energy by performing ( i ) fully self - consistent calculation with vertex corrections or ( ii ) partially ( or no ) self - consistent calculation . along the lines of method ( i )
, schmalian and pines calculated the self - energy with all the vertex corrections by applying a high - temperature approximation @xcite . using a similar technique , fujimoto calculated the pseudogap phenomena due to quasistatic sc fluctuations @xcite .
it is noteworthy that the fourth - order - perturbation theory with respect to @xmath23 well reproduces the weak pseudo - gap since the vertex corrections for the self - energy are considered @xcite .
this method also reproduces the difference of @xmath1 between ybco and lsco @xcite , and the @xmath85-wave sc state in sr@xmath50ruo@xmath49 @xcite . along the lines of method ( ii ) , vilk and tremblay
proposed a two - particle self - consistent ( tpsc ) method @xcite , where full self - consistency is not imposed on the self - energy .
the renormalization of the spin susceptibility is described by introducing the effective coulomb interaction @xmath1160 ( @xmath1161 ) that is determined so as to satisfy the sum rule for @xmath459 .
this philosophy has been applied to extend the dynamical mean field theory ( dmft ) by including self - energy correction due to spin fluctuations @xcite .
these theories can explain the weak pseudo - gap in the dos .
however , they are not suitable for the study of the transport phenomena since they are not conserving approximations . for this reason
, we use the flex approximation in the numerical study for the transport phenomena .
since the flex approximation tends to underestimate the anisotropy of @xmath18 due to the overestimated feedback effect , the enhancements of @xmath9 and @xmath1162 given by the flex+cvc method might be underestimated quantitatively .
finally , we present important future problems .
since the flex ( or flex+@xmath15-matrix ) approximation is a one - loop approximation , it does not work satisfactorily in under - doped systems .
however , the analysis in [ beyondone ] have shown that the enhancement of @xmath9 due to the cvc is not an artifact of the flex approximation , but will be universal near the af - qcp beyond the one - loop approximation .
therefore , relations @xmath1163 and @xmath1164 are expected to hold in under - doped htscs .
it is an important challenge to perform a numerical study of the cvc using the higher - loop approximation , by taking the vertex corrections for the self - energy .
then , we will be able to judge clearly whether the striking transport anomaly in heavily under - doped systems can be understood based on the fermi liquid theory , or the idea of non - fermi liquid theory explained in [ nonfl ] has to be employed . in under - doped systems , small amount of impurities drastically changes the electronic states of the system as discussed in
[ imp - strong ] .
therefore , to understand transport phenomena in under - doped systems , we have to develop the theory of transport phenomena in the presence of disorder . in [ cemin ]
, we presented a numerical study of the transport coefficients in the quasi 2d and 3d hubbard models using the flex+cvc approximation , and explained the experimental results in ce@xmath2in@xmath3 .
however , in real heavy fermion systems , the fermi surfaces is much more complicated . as discussed in [ cemin ] , the huge @xmath9 @xmath1165 ) in ce@xmath2in@xmath3 can not be accounted for a simple rta in the multiband system
. however , we have to study the multiband system for a more realistic study .
in fact , we recently studied the 2d and 3d periodic anderson model , which is a two - band model for a heavy fermion system , and found that a relatively large mass - enhancement factor @xmath1166 is obtained in the flex approximation @xcite . since @xmath6 , @xmath9 and @xmath12
are independent of @xmath842 , we could discuss their critical behaviors using the hubbard model as we have explained in
[ cemin ] .
since @xmath14 and @xmath22 are proportional to @xmath842 , on the other hand , we have to take account of the experimental @xmath842 if one compare the theoretical results with experimental values . finally , we comment on the `` field - induced qcp '' in cecoin@xmath3 near @xmath1113 tesla , where prominent increment in the effective mass @xmath1167 was inferred experimentally .
paglione et al .
@xcite measured the in - plane resistivity near @xmath1168 , and derived the relation @xmath1169 , where @xmath1170 is the coefficient of the @xmath8-term in the resistivity . according to the kadowaki - woods relation @xmath1171 ( see eq .
( [ eqn : kw ] ) ) , @xmath1167 diverges at @xmath1172 .
this fact strongly suggests that the field - induced qcp occurs at @xmath1173 tesla in cecoin@xmath3 .
the prominent increment of @xmath1167 around the field - induced qcp is also confirmed by thermoelectric transport measurements ( @xmath14 and @xmath22 ) @xcite .
interestingly , a similar field - induced qcp is also realized in heavily over - doped bi2212 ( @xmath1174 k ) @xcite . up to now
, the origin of the critical behavior is unknown .
one possible origin will be the ( field - induced ) sdw state that is hidden in the sc state . in cecoin@xmath3 , tanatar et al .
@xcite measured the anisotropy of the resistivity near @xmath1168 , and found that the in - plane resistivity ( @xmath1175 ) is linear - in-@xmath15 above 4 k , whereas it decreases more quickly below 4 k. the observed @xmath15-dependence of @xmath1175 looks similar to the theoretical result in fig .
[ fig : onari ] ( i ) .
in contrast , inter - plane resistivity ( @xmath1176 ) shows a complete @xmath15-linear dependence from 25 mk to 16 k , which might suggest that the cvc is unimportant for @xmath1176 .
moreover , the wiedemann - frantz law ( @xmath1177 ; @xmath4 being the thermal conductivity ) is strongly violated only for the inter - plane direction .
this strong anisotropy in transport may be difficult to explain by the current theory .
since the transport phenomena under high magnetic field is outside of the scope of the present study , this is an important future problem of transport phenomena . up to now
, electronic properties in strongly correlated metals under high magnetic field have not been studied sufficiently .
this issue is an important future challenge .
we should note that the @xmath1168-line in cecoin@xmath3 is a weak first - order line for @xmath1178 @xcite .
therefore , exactly speaking , the field - induced qcp in cecoin@xmath3 can not be a true qcp . according to the arpes measurements for under - doped bi2212 @xcite and lsco @xcite
, the intensity of the spectrum @xmath1179 on the fermi surface at @xmath533 takes very small value around the hot spot below the pseudo - gap temperature @xmath34 . as a result
, @xmath1180 has finite intensity only around the cold spot , which is frequently called the `` fermi arc '' .
the length of the fermi arc diminishes as the temperature decreases , particularly below @xmath34 .
references @xcite have shown that the pseudo - gap is a precursor of the sc gap .
based on this observation , @xmath15-matrix theory @xcite and the flex+@xmath15-matrix theory @xcite have been proposed .
moreover , recent arpes measurements @xcite suggests that the reduction in @xmath1180 around the hot spot in the fermi arc state is not due to a `` density order '' .
the temperature dependence of the fermi arc in under - doped systems is qualitatively reproduced by the flex+@xmath15-matrix approximation , where no hidden order is assumed .
figure [ fig : arc ] shows @xmath1180 given by the flex and flex+@xmath15-matrix approximations for lsco ( @xmath189 ) at @xmath1181 k. the intensity of @xmath1180 at around @xmath140 is suppressed in the flex approximation .
the fermi arc structure becomes more prominent in the flex+@xmath15-matrix approximation , because the quasiparticle damping rate due to strong @xmath36-wave sc fluctuations , which is given in eq .
( [ eqn : self - scf ] ) in the flex+@xmath15-matrix approximation , takes large values around @xmath140 . in [ flex - t ]
, we explained that anomalous transport phenomena below @xmath60 are well explained by the flex+@xmath15-matrix approximation , by considering the cvc due to af+sc fluctuations . even above @xmath60 , the weak pseudo - gap due to the af fluctuations
is widely observed by arpes measurements , which reflects the hot / cold - spot structure of @xmath18 .
therefore , a kind of the fermi arc structure exists even above @xmath60 . for lsco at @xmath152 given by the flex approximation .
@xmath1180 is shown only when @xmath1182 .
the contour is shown on the basal plane . *
( ii ) * @xmath1180 given by the flex+@xmath15-matrix approximation .
the contour shows a `` fermi arc structure '' . ]
the fermi arc structure is also observed in under - doped ncco : in the arpes measurements @xcite , the intensity of @xmath1180 for a portion of the fermi surface near the afbz ( see fig .
[ fig : fs - hotcold ] ) is strongly suppressed below @xmath1151 k by the sdw order .
below @xmath67 , @xmath9 is negative by reflecting the small electron - like fermi surface around @xmath140 , which is caused by the reconstruction of the fermi surface due to the sdw order .
surprisingly , @xmath9 remains negative even above @xmath67 , although sdw - induced fermi surface reconstruction is absent .
moreover , no anomaly in @xmath9 is observed at @xmath67 .
this highly nontrivial fact is explained by the fermi liquid theory by considering the cvc , as shown in fig .
[ fig : rh ] ( iv ) . according to the arpes measurements @xcite , pseudo - gap state due to af fluctuations
is observed even above @xmath67 .
therefore , a fermi arc state appears to be realized in electron - doped systems .
when a distinct fermi arc structure exists , the `` effective carrier density at the fermi level ( @xmath1183 ) '' reduces in proportion to the length of the fermi arc . since @xmath1184 in the rta , it is sometimes claimed that the enhancement of @xmath9 in htscs can be explained by the reduction in @xmath1183 due to the emergence of the fermi arc .
this idea seems to be valid in heavily under - doped compounds @xcite . however , this idea is not true for slightly under - doped systems since @xmath9 starts to decrease below @xmath34 , whereas the length of the fermi arc shrinks due to sc fluctuations .
the reduction in @xmath9 in the pseudo - gap region is explained by the reduction in the af fluctuation , as explained in
[ flex - t-1 ] . , which was explained in fig .
[ fig : j ] ( ii ) , is also described .
the smallness of the effective fermi surface is the origin of the enhancement of @xmath9 above @xmath60 .
below @xmath60 , the effective fermi surface approaches the true fermi surface since the cvc due to af fluctuations are reduces .
therefore , @xmath9 decreases below @xmath60 . ] here , we discuss the hall coefficient in the presence of the `` fermi arc structure due to strong af fluctuations '' , by taking the cvc into consideration .
the cases we consider correspond to the hole - doped systems above @xmath60 , and the electron - doped systems above @xmath67 .
figure [ fig : arc2 ] shows the fermi arc structure due to the weak pseudo - gap formation and the `` effective fermi surface '' that is defined to be perpendicular to the total current @xmath19 , as introduced in fig .
[ fig : j ] ( ii ) .
( note that the `` effective fermi surface '' is not visible in the arpes measurements . ) within the rta , @xmath1185 and @xmath1186 , where @xmath1187 is the length of the cold spot ( fermi arc ) .
@xmath1188 as we have discussed , @xmath387 is proportional to the curvature of the effective fermi surface at the cold spot . in the presence of strong af fluctuations above @xmath34
, @xmath1189 rotates approximately by @xmath1190 around the cold spot , as shown in fig .
[ fig : arc2 ] .
therefore , the curvature of the effective fermi surface at the cold spot in fig .
[ fig : arc2 ] is proportional to @xmath1191 . as a result , due to the cvc above @xmath34 , @xmath1192 approximately .
the large curvature of the effective fermi surface around the cold spot is the origin of the enhancement of @xmath9 . in conclusion ,
when the fermi arc emerges , the magnitude of @xmath9 is much larger than the value obtained by the rta that is given in eq .
( [ eqn : arc1 ] ) .
we note that if the distinct fermi arc structure due to the highly anisotropic @xmath18 is realized , it should lead to a huge magnetoresistance within the rta : in fact , @xmath589 diverges when the @xmath21-dependence of the mean free path @xmath1193 is not continuous on the fermi surface ; in this case , the second term in eq .
( [ eqn : mcrta ] ) diverges as pointed out in refs .
@xcite . therefore , the modified kohler s rule given by eq .
( [ eqn : mkrexp ] ) can not be satisfied . in summary , a unified understanding of anomalous transport phenomena in htsc
can not be achieved by the rta even if the fermi arc structure is taken into account . in previous sections
, we studied single - band models , and showed that the cvc induces various striking non - fermi - liquid - like behaviors in the presence of af fluctuations .
for example , @xmath9 shows strong temperature dependence due to the cvc . in multiband systems , however , @xmath9 can exhibit temperature dependence or sign change within the rta , if a hole - like fermi surface and an electron - like fermi surface coexist and their relaxation times have different @xmath15-dependences .
there are many such examples even in conventional metals . in this study
, we did not discuss the multiband effect , since the magnitude @xmath42 will remain @xmath1194 in this mechanism .
thus , the multiband mechanism is impossible to explain the huge @xmath42 observed in various systems near the af qcp .
however , a kind of multiorbital effect causes a huge hall coefficient in @xmath64- and @xmath17-electron systems , which is known as the `` anomalous hall effect ( ahe ) '' . in the presence of the ahe ,
the hall resistivity is given by @xmath1195 where @xmath1196 is the magnetic field and @xmath1197 is the ordinary hall coefficient .
@xmath1198 is the anomalous hall coefficient , which is ( generally ) proportional to the magnetization @xmath2 . in general ,
@xmath1197 and @xmath1198 are even functions of @xmath1196 and @xmath2 . in ferromagnets , the second term @xmath1199 takes a finite value even if @xmath1200 .
study of the ahe due to multiband effect was initiated by karplus and luttinger @xcite , and by luttinger @xcite .
they found that the anomalous hall conductivity ( ahc ) @xmath1201 is independent of the resistivity @xmath6 .
this karplus - luttinger - term is called the `` intrinsic ahe '' because it exists even in systems without impurities . later
, smit presented a mechanism of `` extrinsic ahe '' @xcite : he found that the spin - polarized electrons are scattered asymmetrically around an impurity site in the presence of spin - orbit interaction .
the ahc due to this skew - scattering mechanism is linearly proportional to @xmath6 . in paramagnetic heavy - fermion systems ,
the observed hall coefficient is given by @xmath1202 where @xmath1203 and @xmath1204 is the uniform magnetic susceptibility . in heavy - fermion systems ,
the ahe due to the second term of eq .
( [ eqn : ahe - hf ] ) takes large value since the uniform susceptibility @xmath1205 is widely enhanced due to the strong coulomb interaction . due to the ahe
, @xmath1206 starts to increase with increasing temperature from 0k , and it turns to decrease after exhibiting its maximum value around the coherent temperature @xmath1207 .
the maximum value of @xmath1208 is more than one order of magnitude greater than @xmath1209 m@xmath1075 .
the ahe can not be derived from the rta since the interband hopping of electrons is important .
[ in the ahe , the effect of cvc is not expected to be crucial .
] experimentally , @xmath1198 is positive in usual ce and u based heavy - fermion systems @xcite . here
, we discuss the ahe in heavy - fermion systems .
bandstructure of ce- and yb - based heavy - fermion system is described by the following orbitally degenerate periodic anderson model @xcite : @xmath1210 where @xmath1211 ( @xmath1212 ) is the creation operator of the conduction electron ( @xmath17-electron ) with @xmath1213 ( @xmath1214 ) .
@xmath17-orbital degeneracy is @xmath1215 .
@xmath1216 ( @xmath1217 ) is the spectrum for conduction electrons ( @xmath17 electrons ) . in the case of ce - compound ( @xmath1218 ) , the complex @xmath282-@xmath17 mixing potential
is given by @xmath1219 , where @xmath1220 is the spherical harmonic function , in the case of yb - compound ( @xmath1221 ) , @xmath1222 .
note that the relation @xmath1223 holds .
the green function for @xmath282 electron is given by @xcite @xmath1224 where @xmath1225 , @xmath1226 , @xmath1227 , and @xmath1228 is the renormalization factor . since
the quasiparticle energy @xmath210 satisfies re@xmath1229 , the equation for @xmath210 is given by @xmath1230 here , we derive the quasiparticle energy by analyzing eq .
( [ eqn : qp ] ) : at zero temperature where @xmath1231 , heavy quasiparticles band is formed due to @xmath282-@xmath17 hybridization . below @xmath1207 ,
the quasiparticles band is @xmath1232 where @xmath1233 ( @xmath1234 ) represents the lower ( upper ) quasiparticles band .
@xmath1235 grows monotonically as temperature increases , and @xmath1236 at the coherent temperature @xmath1207 .
when @xmath1237 , @xmath282-@xmath17 hybridization is prohibited by @xmath1238 .
we summarize the electronic states in heavy - fermion systems : * ( i ) * below @xmath1207 , fermi liquid state with heavy quasiparticles is realized due to @xmath282-@xmath17 hybridization
. mass enhancement factor @xmath842 and uniform susceptibility @xmath1239 are constant .
the heavy quasiparticle bandwidth @xmath1240 is approximately given by @xmath1241 , where @xmath1242 is the @xmath282-electron bandwidth . * ( ii ) * above @xmath1207 , @xmath282-@xmath17 hybridization ceases , and localized @xmath17-electrons causes curie - weiss susceptibility .
experimentally , the temperature of maximum resistivity @xmath1243 is larger than @xmath1207
. in u@xmath51ni@xmath51sn@xmath49 plotted as functions of ( i ) @xmath8 and ( ii ) @xmath1109 .
the relation @xmath1244 is satisfied well , as predicted by the theory of ahe @xcite .
@xcite ] ] in paramagnetic heavy - fermion systems , the hall coefficients takes huge values due to the ahe @xcite . in the early stage ,
coleman et al @xcite and fert and levy @xcite developed theories of extrinsic ahe : they studied the extrinsic mechanism based on the @xmath17-electron impurity anderson models with @xmath64-orbital channels , and predicted the relation @xmath1245 above @xmath1207 when the @xmath64-orbital phase - shift is finite . on the other hand , kontani and yamada studied the intrinsic ahe due to multiband effect based on the @xmath1218 periodic anderson model , which a the model in the strong limit of spin - orbit interaction @xcite . in this model
, the anomalous velocity originates from the angular momentum of the localized @xmath17-electrons .
that is , the ahe originates from the transfer of angular momentum of the @xmath17-electron to the conduction electron .
based on the linear - response theory , they derived a large intrinsic ahc : @xmath1246 } = 5.8\times 10^{-9 } ( n\mbox{[cm]$^{-3}$})^{1/3}/{\tilde e}_f \mbox{[k]}$ ] where @xmath1247 is the renormalized @xmath17-level measured from the fermi energy .
@xmath1248 gives approximate quasiparticle bandwidth .
the present study predicts that @xmath1249 in ce ( yb ) based heavy - fermion systems since @xmath1250 .
this prediction is consistent with the experimental results @xcite . if we substitute @xmath1251k and @xmath1252@xmath1253 , @xmath1254}$ ] , which is a typical experimental value .
the predicted temperature dependence of @xmath1255 is @xmath1256 @xcite .
therefore , @xmath1257 shows the following cross - over behavior : @xmath1258 \ \ \ \ \ \
\mbox { : below $ t_{\rm coh}$ ( $ \g^*<|{\tilde e}_f|$ ) } , \label{eqn : ahe1 } \\ & & r_h^{\rm ahe } \propto \chi \ [ \propto 1/t ] \ \ \ \ \ \ \mbox { : above $ t_{\rm coh}$ ( $ \g^*>|{\tilde
e}_f|$)}. \label{eqn : ahe2}\end{aligned}\ ] ] in typical ce - based heavy - fermion systems , @xmath1259 and @xmath1260 k. below @xmath1207 , eq .
( [ eqn : ahe1 ] ) is proportional to @xmath1109 since @xmath1261 is constant for @xmath1262 .
figure [ fig : ahe ] shows the @xmath9 in u@xmath51ni@xmath51sn@xmath49 ( @xmath1263mj / k@xmath74 ) , where the relation @xmath1264 holds well below @xmath1265k .
note that the extrinsic type ahe @xcite and the intrinsic type ahe @xcite can coexist .
however , the relation @xmath1266 suggests that the intrinsic - type ahe is dominant at least below @xmath1207 . above @xmath1207 , the ahe due to interband transition
is suppressed when @xmath282-@xmath17 mixing is prohibited .
therefore , eq . ( [ eqn : ahe2 ] ) is independent of @xmath6 .
this _ coherent - incoherent crossover of intrinsic ahe _ was first theoretically derived in ref .
this crossover behavior is not restricted to heavy - fermions systems , but also observed in various transition ferromagnets @xcite . in heavy - fermion systems near the af qcp
, @xmath1110 shows strong temperature dependence due to the cvc . to extract @xmath1110 from the observed hall coefficient
, we have to seriously consider the @xmath15-dependence of @xmath1111 . very fortunately
, ahe is vanishingly small in ce@xmath2in@xmath3 @xcite .
therefore , we could perform reliable analysis of the ordinary hall effect in ce@xmath2in@xmath3 as discussed in
[ cemin ] , without the necessity of subtracting the ahe .
note that the ahe vanishes when the crystal - field splitting of the @xmath17-levels is much larger than @xmath1207 .
this may be the reason for the small ahe in ce@xmath2in@xmath3 .
recently , theory of the intrinsic ahe in ( ferromagnetic ) @xmath64-electron systems has been developed based on realistic multiorbital tight - binding models @xcite .
it was revealed that a large anomalous velocity emerges in general multiorbital @xmath64-electron systems because of the inter - orbital hopping .
this is the origin of the large ahe in @xmath64-electron systems , which is very similar to the origin of the ahe in @xmath17-electron systems @xcite .
the intrinsic ahes for fe @xcite and in srruo@xmath51 @xcite were also calculated based on the lda band calculations . here , we present an intuitive explanation of the ahe due to the multiorbital effect @xcite . in a multiorbital system , a conduction electron acquires the `` effective aharonov - bohm phase factor '' due to @xmath64-atomic angular momentum with the aid of the spin - orbit interaction and the inter - orbital hoppings , which is responsible for the hall effect . the intuitive explanation based on the tight - binding model , which is given in fig .
[ fig : she ] , which represents a square lattice ( @xmath1267)-orbital tight - binding model with @xmath118-component of the spin - orbit interaction @xmath1268 @xcite , which is a simplified model for ( ca , sr)@xmath50ruo@xmath49 : this compound shows large ahe under the magnetic field @xcite .
its magnitude is comparable with the large ahe in @xmath17-electron systems ( such as upt@xmath51 ) . in fig .
[ fig : she ] , @xmath1269 represents the interorbital hopping integral between next nearest sites .
now , let us consider the motion of a down - spin electron along a triangle of half unit cell : an electron in the @xmath1270-orbital can transfer to @xmath1271-orbital and vise versa using the spin - orbit interaction @xmath1272 , where @xmath1273 . by combination of angular dependence of interorbital hopping and the spin - orbit interaction ,
any clockwise ( anti - clockwise ) motion along any triangle path with spin - orbit interaction causes the factor @xmath1274 ( @xmath1275 ) .
this factor can be interpreted as the aharonov - bohm phase factor
@xmath1276 @xmath1277 $ ] , where @xmath1278 represents the `` effective magnetic flux '' in the half unit cell .
we revealed the fact that the effective magnetic flux , which is inherent in multiorbital @xmath64-electron systems , causes huge ahe in various transition metals . , @xmath1271)-orbital tight - binding model with @xmath118-component of the spin - orbit interaction @xmath1268 .
the effective magnetic flux for up - electrons is negative .
both @xmath1279 and @xmath1280 are given by the linear combination of @xmath1281 . since @xmath1282 in the present basis ,
the rotation operator @xmath1283 is equal to @xmath1284 for @xmath1285 .
therefore , @xmath1286 and @xmath1287 . ] in many @xmath64-electron ferromagnets , intrinsic ahe ( @xmath1288 ) seems to be observed experimentally .
in particular , asamitsu et al .
measured ahe in various transition - metal ferromagnets , and found out a universal crossover behavior from @xmath1288 to @xmath1289 @xmath1290 as the resistivity increases , around @xmath1291 cm .
recent theoretical study in ref .
@xcite has revealed that the experimental result is explained by the following coherent - incoherent crossover @xmath1292 where @xmath1293 represents the minimum bandsplitting near the fermi level .
@xmath1294 ev in usual transition metal ferromagnets .
it is apparent that eqs .
( [ eqn : ahe3 ] ) and ( [ eqn : ahe4 ] ) correspond to eqs .
( [ eqn : ahe1 ] ) and ( [ eqn : ahe2 ] ) for heavy - fermion systems .
[ note that @xmath1295 . ] therefore , a conventional behavior @xmath1288 is violated in high - resistivity metals since the interband particle - hole excitation , which is the origin of the ahe , is suppressed when @xmath702 is larger than bandsplitting energy @xmath1293 .
there is a simple intuitive explanation for this coherent - incoherent crossover of intrinsic ahe : when @xmath73 is sufficiently small , the intrinsic hall conductivity is proportional to the lifetime of the interband particle - hole excitation : @xmath1296 @xcite . in the high - resistivity regime where @xmath1297 , the shc decreases drastically with @xmath73 since the interband excitation is suppressed when the quasiparticle lifetime @xmath1298 is shorter than @xmath1296 . according to the above discussion
, _ coherent - incoherent crossover of intrinsic ahe _ should be universal and is widely observed in various multiorbital @xmath85- , @xmath64- , and @xmath17-electron systems .
it is noteworthy that the large anomalous velocity also induces a sizable spin hall effect ( she ) in paramagnetic multiorbital systems @xcite .
the she is the phenomena that an applied electric field induces a spin current @xmath1299 in a transverse direction .
recently , she attracts great attention due to its fundamental interest and its potential application in spintronics .
karplus and luttinger @xcite showed that an applied electric field induces a spin - dependent transverse current in the presence of spin - orbit interaction .
this mechanism causes the ahe in ferromagnetic metals , and the she in paramagnetic metals .
recently , the spin hall conductivity ( shc ) in pt had been observed @xcite , and found that the shc in pt is @xmath1300 times larger than that observed in @xmath11-type semiconductors .
this experiment fact has attracted considerable attention to the study of the she in transition metals .
recently , we propose a new mechanism for the giant she originate from the @xmath64-orbital degrees of freedom , which is absent in semiconductors @xcite . in a multiorbital system
, a conduction electron acquires the `` effective aharonov - bohm phase factor '' due to @xmath64-atomic angular momentum , as explained in fig .
[ fig : she ] .
we have studied the she s in sr@xmath50ruo@xmath49 @xcite , which is the first theoretical study of the she in @xmath64-electron systems , in pt @xcite , and in various @xmath1301 and @xmath1302 transition metals @xcite .
it is found that the explanation in fig .
[ fig : she ] seems to capture the characteristics of she in @xmath64-electron systems .
in contrast , the `` dirac monopole mechanism '' @xcite is appropriate when massless dirac corn dispersion exists close to the fermi level like in semiconductors .
the present mechanism of she is also different from that in rashba - type 2d electron - gas model due to momentum - dependent soi @xcite .
thus , giant she due to atomic orbital degrees of freedom is ubiquitous in various @xmath85- , @xmath64- , @xmath17-electron systems . finally , we discuss the multiorbital effect on the kadowaki - woods ( kw ) ratio @xmath1303 , where @xmath1170 is the coefficient of the @xmath8 term in the resistivity , and @xmath73 is the coefficient of the @xmath15-linear term of the electric specific heat .
experimentally , the kw ratio in ce- and u - based heavy - fermion compounds shows an approximate universal value @xmath1304 [ @xmath936 cm ( mol@xmath1305k / mj)@xmath74 ] , which is known as the kw relation @xcite .
although it was believed to be universal in heavy - fermion systems for a long time , recent experimental activities have revealed that the kw relation is strongly violated in many yb - based compounds .
recently , the author had derived a generalized kw relation that is applicable for systems with general @xmath17-orbital degeneracy @xmath1307 for ce- and yb - based compounds @xcite and for sm- and er - based compounds @xcite . by considering the material dependence of @xmath1307
, the mystery of the failure of the kw relation was resolved . and @xmath1308 .
* ( b ) * a scaling relation between @xmath1309 and @xmath1310 @xcite .
@xmath1309 corresponds to @xmath1311 , and @xmath1312 ( @xmath1313 is a constant of order @xmath230 ) . [
@xcite ] ] here , we analyze the kw ratio in terms of the dmft ( @xmath1314-limit ) @xcite , which is believed to be useful in heavy - fermion systems that is not close to the af - qcp . by using the dmft
, we can utilize the strong - coupling fermi liquid theory for the impurity anderson model developed by yamada and yosida @xcite : using the ward identity , they derived the exact wilson ratio @xmath1315 for @xmath1316 in the strong coupling limit , that is , @xmath1317 and the @xmath17-electron charge susceptibility is zero ( kondo model limit ) . here ,
@xmath1318 and @xmath1319 represent the non - interacting values . using the dmft , @xmath702 and im@xmath1320 in the pam ( in eq .
( [ eqn : pam ] ) is given by @xcite @xmath1321 in the strong coupling limit .
here , @xmath1322 is the avogadro constant , and @xmath1323 is the local four - point vertex .
@xmath1324 is the dos for @xmath17-electron per channel ; @xmath1325 is the total dos at the fermi level .
when @xmath1216 in eq .
( [ eqn : pam ] ) is a free dispersion , the conductivity is given by @xmath1326 where @xmath1327 is the density of quasiparticles that form the conduction band , and @xmath121 is the unit cell length . according to eqs .
( [ eqn : g - kw])-([eqn : s - kw ] ) , we find the scaling properties @xmath1328 and @xmath1329 .
therefore , we obtain the `` grand kw relation '' that is valid for any @xmath1330 @xcite : @xmath1331 } \label{eqn : kw}.\end{aligned}\ ] ] where we put @xmath1332 , @xmath1333 jk@xmath1334 , and we assumed @xmath1335 cm . without crystal field
splitting of the @xmath17-level , @xmath1336 for ce@xmath1337 and sm@xmath1337 ions , and @xmath1338 for yb@xmath1337 and er@xmath1339 ions .
therefore , the previous kw relation turned out to be valid only when @xmath1316 ( kramers doublet case due to strong crystal field splitting ) . a similar universal relation @xmath1340 ( @xmath14 is the seebeck coefficient )
was recently found @xcite .
the characteristics of the electronic state in heavy - fermion systems are ( i ) large mass - enhancement and ( ii ) small charge susceptibility since the @xmath17-electron is almost localized . the grand kw relation ( [ eqn : kw ] ) is derived only by imposing these constraints on the microscopic fermi liquid theory .
this fact illustrates a remarkable advantage of the fermi liquid theory for the analysis of strongly correlated systems .
figure [ fig : gkw ] ( i ) shows the experimental verification of eq .
( [ eqn : kw ] ) for various heavy - fermion compounds , where @xmath1307 in each compound was determined by the temperature dependence of @xmath1261 and the inelastic neutron scattering @xcite .
tsujii s study confirmed that @xmath1341 in many ce - based compounds , where crystal - field splitting is larger than the renormalized fermi energy @xmath1240 . on the other hand , @xmath1342 in many yb - based ones , where crystal - field splitting is smaller than @xmath1240 .
torikachvili et al . also found that other yb - based heavy - fermion systems ybt@xmath50zn@xmath1343 ( t = fe , co , ru , rh , os , or ir ) follows the grand kw relation shown in eq .
( [ eqn : kw ] ) @xcite . when @xmath1307 is larger than the number of conduction band @xmath1344 , almost unhybridized @xmath17-levels exist near the fermi level , except when the crystal - field splitting is very large .
this situation induces a huge anomalous hall coefficient @xcite as well as the large van vleck susceptibilities @xcite in kondo insulators and in singlet sc heavy - fermion systems .
we shortly discuss the absence of the multiband effect on the kw relation : in a usual heavy - fermion compound , there are one or two ( relatively ) large fermi surfaces composed of heavy quasiparticles , and several ( relatively ) small fermi surfaces composed of light - quasiparticles . in the presence impurities , @xmath1345 as explained in eq .
( [ eqn : rho - imp ] ) . therefore ,
heavy quasiparticles on the large fermi surfaces give the dominant contribution to both the specific heat and the @xmath1170-term . as a result
, the grand - kw relation is universally realized even in multiband systems .
moreover , in many heavy - fermions , there is evidence that the heavy quasiparticles on the large fermi surfaces give the dominant contribution to the conductivity .
for example , @xmath1346 seems to be realized in many heavy - fermions away from af - qcps , although @xmath1347 should be realized when small fermi surfaces are the most conductive .
figure [ fig : gkw ] ( ii ) shows the mid - infrared peak energies @xmath1309 of the optical conductivity @xmath731 in various heavy - fermion systems @xcite . according to the fermi liquid theory , @xmath1348 is satisfied in the pam , where @xmath522 is the @xmath282-@xmath17 mixing potential and @xmath1349 is the mass enhancement factor .
okamura confirmed that the relation @xmath1350 is universally satisfied in various heavy - fermion compounds , as shown in fig .
[ fig : gkw ] ( ii ) .
this scaling relation is well satisfied regardless of @xmath1307 .
their study established the validity of the pam and the fermi liquid theory in various heavy - fermion compounds .
the author would like to thank k. yamada , k. ueda , k. kanki , h. kino , s. onari and y. tanaka for collaborations on the theory of the transport phenomena in strongly correlated systems .
thanks are also due to y. matsuda and y. nakajima for close discussions and collaborations on ce@xmath2in@xmath3 and other heavy - fermion systems .
further , i would like to thank h.d
. drew and d.c .
schmadel for the useful discussions on ac hall effect in htscs .
i am grateful to h. taniguchi for the discussion on organic metals , particularly for his kind offer of providing the figures used in fig .
[ fig : taniguchi ] .
the work on magnetoresistance was supported by the ministry of education , science , sports , and culture , japan , during my stay in augsburg university .
the author was supported by a grant - in - aid for the encouragement of young scientists under the contract nos . 18740197 , 15740198 and 13740202 from the ministry of education , culture , sports , science , and technology ( mext ) of japan .
this work has been also supported by grants - in - aid for scientific research in priority areas `` skutterudites '' and `` anomalous quantum material '' from mext .
999 j. takeda , t. nishikawa , and m. sato , physica c * 231 * ( 1994 ) 293 ; t. nishikawa , j. takeda and m. sato , j. phys .
* 63 * ( 1994 ) 1441 .
hwang , b. batlogg , h. takagi , h.l .
kao , j. kwo , r.j .
cava , j.j .
krajewski , and w.f .
peck , jr .
* 72 * ( 1994 ) , 2636 .
peng xiong , gang xiao , and x.d .
wu : phys .
b * 47 * ( 1993 ) , 5516 .
chien , z.z .
wang , and n.p .
ong : phys .
* 67 * ( 1991 ) , 2088 .
y. kubo , y. shimakawa , t. manako , and h. igarashi : phys .
b * 43 * ( 1991 ) , 7875 .
p. fournier , x. jiang , w. jiang , s. n. mao , t. venkatesan , c. j. lobb and r. l. greene : phys .
b56 * ( 1997 ) 14149 .
y. dagan , m.m .
qazilbash , c.p .
hill , v.n .
kulkarni , and r.l .
greene , phys .
* 92 * ( 2004 ) 167001 .
y. abe , k. segawa and y. ando , phys .
b * 60 * , r15055 ( 1999 ) .
y. hanaki , yoichi ando , s. ono , j. takeya phys .
b * 64 * , 172514 ( 2001 )
. g. xiao , p. xiong , and m. z. cieplak , phys .
b * 46 * , 8687 ( 1992 ) .
a. malinowski , marta z. cieplak , s. guha , q. wu , b. kim , a. krickser , a. perali , k. karpi?ska , m. berkowski , c. h. shang , and p. lindenfeld , phys .
b * 66 * , 104512 ( 2002 ) .
k. katayama , t. nagai , h. taniguchi , k. satoh , n. tajima and r. kato , j. phys .
76(2007 ) suppl .
v. sushko , n. shirakawa , k. murata , y. kubo , n.d .
kushch , e.b .
yagubskii : synth . met .
* 85 * ( 1997 ) 1541 .
h. taniguchi , unpublished .
k. murata , m. ishibashi , y. honda , n.a .
fortune , m. tokumoto , n. kinoshita and h. anzai : solid state comm .
* 76 * ( 1990 ) 377 .
h. taniguchi , unpublished .
h. taniguchi , unpublished .
y. nakajima , k. izawa , y. matsuda , s. uji , t. terashima , h. shishido , r. settai , y. onuki and h. kontani : j. phys .
* 73 * ( 2004 ) 5 .
y. nakajima , h. shishido , h. nakai , t. shibauchi , k. behnia , k. izawa , m. hedo , y. uwatoko , t. matsumoto , r. settai , y. onuki , h. kontani , and y. matsuda : j. phys .
* 76 * ( 2007 ) 027403 .
t. namiki , h. sato , h. sugawara , y. aoki , r. settai and y. onuki : preprint . s. paschen , t. luhmann , s. wirth , p. gegenwart , o. trovarelli , c. geibel , f. steglich , p. coleman , q. si : nature * 432 * ( 2004 ) 881 .
anderson : phys .
lett . * 67 * ( 1991 ) 209 ; p.w .
anderson , _ the theory of superconductivity in the high - tc cuprates _ ( princeton series in physics ) .
r. hlubina and t. m. rice , phys .
b * 51 * , 9253 - 9260 ( 1995 ) .
b.p . stojkovic and d. pines , phys . rev .
b * 55 * , 8576 ( 1997 ) . l.b .
ioffe and a.j .
millis : phys .
b * 58 * ( 1998 ) 11631 .
y. yanase and k. yamada : j. phys .
soc . jpn .
* 68 * ( 1999 ) 548 .
l. landau : sov .
jetp * 3 * ( 1957 ) 920 ; ibid * 5 * ( 1957 ) 101 ; ibid * 8 * ( 1959 ) 70 .
p. nozires and d. pines : _ theory of quantum liquids_. p. nozires : _ theory of interaction fermion systems _
( benjamin , new york , 1964 ) .
abricosov , l.p . gorkov and i.e. dzyaloshinskii : _ methods of quantum field theory in statistical physics _ ( dover , new york , 1975 )
. h. kontani and k. yamada : j. phy .
* 74 * ( 2005 ) 155 .
h. kontani , k. kanki and k. ueda : phys .
b * 59 * ( 1999 ) 14723 .
k. kanki and h. kontani : j. phys .
* 68 * ( 1999 ) 1614
. h. kontani : phys .
* 70 * ( 2001 ) 1873 .
h. kontani : phys .
* 70 * ( 2001 ) 2840 .
h. kontani , phys . rev
* 89 * ( 2002 ) 237003 .
h. kontani : phys .
b * 64 * ( 2001 ) 054413 .
h. kontani , phys . rev .
b * 67 * ( 2003 ) 014408
. g. m. eliashberg : sov .
jetp * 14 * ( 1962 ) , 886 .
langer : phys . rev .
* 120 * ( 1960 ) 714 . k. yamada and k. yosida : prog .
phys . * 76 * ( 1986 ) 621 : h. kohno and k. yamada : prog .
* 80 * ( 1988 ) 623 .
h. fukuyama , h. ebisawa and y. wada : prog .
theor . phys .
* 42 * ( 1969 ) 494 .
s. fujimoto , h. kohno and k. yamada , phys .
* 60 * ( 1991 ) 2724 .
h. maebashi and h. fukuyama : phys .
soc . jpn .
* 66 * ( 1997 ) 3577 ; ibid * 67 * ( 1998 ) 242 .
see , e.g. , y. iye , in _ physical properties of high temperature superconductors , _ edited by d.m .
ginsberg ( world scientific , singapore , 1992 ) vol .
3 . see , e.g. , k. asayama , y. kitaoka , g .- q .
zheng , and k. ishida : prog .
nmr spectroscopy 28 ( 1996 ) 221 .
y. suzumura , y. hasegawa and h. fukuyama : j. phys .
soc . jpn . * 57 * ( 1988 ) 2768 ; g. kotliar : phys . rev .
b * 37 * , 3664 ( 1988 ) .
p. a. lee , n. nagaosa , and x .-
wen , rev .
phys . * 78 * , 17 ( 2006 ) .
y. yanase , t. jujo , t. nomura , h. ikeda , t. hotta and k. yamada : physics report * 387 * ( 2003 ) 1 .
k. yamada : _ electron correlation in metals _
( cambridge univ .
press 2004 ) .
t. moriya , y. takahashi , and k. ueda : j. phys .
jpn * 59 * ( 1990 ) 2905 , k. ueda , t. moriya , and y. takahashi : j. phys .
solids * 53 * ( 1992 ) 1515 .
t. moriya and k. ueda : adv
. physics * 49 * ( 2000 ) 555 .
k. miyake and o. narikiyo , j. phys .
* 63 * ( 1994 ) 3821 .
p. monthoux and d. pines : phys .
b * 47 * ( 1993 ) 6069 . n.e .
bickers and s.r .
white : phys .
b * 43 * ( 1991 ) 8044 .
k. kuroki , r. arita , and h. aoki : phys .
b * 60 * , 9850 ( 1999 ) .
h. yoshimura and d.s .
hirashima : j. phys .
jpn * 73 * ( 2004 ) 2057 .
p. coleman , c. ppin , q. si and r. ramazashvili : j.phys . : condens .
matter * 13 * ( 2001 ) r723 . s. sachdev : _ quantum phase transitions _ , ( cambridge university press , cambridge uk 1999 ) .
a. v. chubukov , d. pines and j. schmalian , _ the physics of superconductors , vol
. 1 _ , ed .
bennemann and j. b. ketterson , ( springer , berlin , 2002 ) ; a. abanov , a. v. chubukov and j. schmalian , adv .
* 52 * ( 2003 ) 119 and references therein .
d. manske , _ theory of unconventional superconductors : cooper - pairing mediated by spin excitations _
( springer 2004 , berlin ) . m. sigrist and t. m. rice , j. phys . soc .
* 61 * ( 1992 ) 4283 .
d. a. wollmann , d. j. van harlingen , w. c. lee , d. m. ginsberg and a. j. leggett , phys .
* 71 * ( 1993 ) 2134 .
c. c. tsuei , j. r. kirtley , c. c. chi , l. s. yu - jahnes a. gupta , t. shaw , j. z. sun , and m. b. ketchen , phys .
lett . * 73 * ( 1994 ) 593 .
s. kashiwaya and y. tanaka , rep .
* 63 * ( 2000 ) 1641 .
t. hotta , j. phys .
* 63 * ( 1994 ) 4126 .
see , e.g. , n. p. ong : _ physical properties of high temperature superconductors _ , edited by d.m .
ginsberg ( world scientific , singapore , 1992 ) , vol.2 .
armitage , d.h .
lu , c. kim , a. damascelli , k.m .
shen , f. ronning , d.l .
feng , p. bogdanov , and z .- x .
shen : phys .
* 87 * ( 2001 ) 147003 ; n. p. armitage , f. ronning , d. h. lu , c. kim , a. damascelli , k. m. shen , d. l. feng , h. eisaki , z .- x .
shen , p. k. mang , n. kaneko , m. greven , y. onose , y. taguchi , and y. tokura : phys .
lett . * 88 * , 257001 ( 2002 ) .
j. m. harris , y. f. yan , p. mati , n. p. ong , p. w. anderson , t. kimura and k. kitazawa : phys .
* 75 * ( 1995 ) 1391 .
t. kimura , s. miyasaka , h. takagi , k. tamasaku , h. eisaki , s. uchida , k. kitazawa , m. hiroi , m. sera , and n. kobayashi : phy .
b * 53 * ( 1996 ) 8733 .
a. w. tyler , y. ando , f. f. balakirev , a. passner , g. s. boebinger , a. j. schofield , a. p. mackenzie , and o. laborde : phys .
b * 57 * ( 1998 ) r728 .
xu , y. zhang , n.p .
ong , cond - mat/9903123 .
y. ando and t. murayama : phys .
b * 60 * ( 1999 ) r6991 .
yasushi abe , yoichi ando , j. takeya , h. tanabe , t. watauchi , i. tanaka , h. kojima , phys .
b * 59 * , 14753 ( 1999 ) .
f. f. balakirev , i. e. trofimov , s. guha , m.z .
cieplak , and p. lindenfeld , phys .
b * 57 * , r8083 ( 1998 ) .
h. yasuoka , t. imai and t. shimizu , _ strong correlation and superconductivity _ ( springer - verlag , berlin 1989 ) , p. 254 . w. w. warren , r. e. walstedt , g. f. brennert , r. j. cava , r. tycko , r. f. bell and g. dabbagh , phys .
* 62 * ( 1989 )
1193 . m. takigawa , a. p. reyes , p. c. hammel , j. d. thompson , r. h. heffner z. fisk and k. c. ott , phys .
b * 43 * ( 1991 ) 247 .
y. itoh , h. yasuoka , y. fujiwara , y. ueda , t. machi , i. tomeno , k. tai , n. koshizuka and s. tanaka , j. phys .
* 61 * ( 1992 ) 1287 .
m. h. julien , p. carretta and m. horvati@xmath1351 , phys .
* 76 * ( 1996 ) 4238 .
h. ding , t. yokoya , j. c. campuzano , t. takahashi , m. randeria , m. r. norman , t. mochiku , k. kadowaki and j. giapintzakis , nature .
* 382 * ( 1996 ) 51 .
a. g. loeser , z. x. shen , d. s. dessau , d. s. marshall , c. h. park , p. fournier and a. kapitulnik , science . * 273 * ( 1996 ) 325 .
m. r. norman , h. ding , m. randeria , j. c. campuzano , t. yokoya , t. takeuchi , t. takahashi , t. mochiku , k. kadowaki , p. guptasarma and d. g. hinks , nature .
* 392 * ( 1998 ) 157 .
a. damascelli , z. hussain and z .- x .
shen , rev .
* 75 * ( 2003 ) 473 .
q. chen , i. kosztin , b. jank ' o and k. levin : phys .
lett * 81 * ( 1998 ) 4708 ; q. chen , i. kosztin , b. janko , and k. levin : phys .
b * 59 * , 7083 ( 1999 ) ; j. maly , b. janko , and k. levin : phys . rev .
b * 59 * , 1354 ( 1999 ) .
t. dahm , d. manske and l. tewordt : europhys . lett .
* 55 * ( 2001 ) 93 .
a. kobayashi , a. tsuruta , t. matsuura and y. kuroda : j. phys .
* 70 * ( 2001 ) 1214 ; a.kobayashi , a. tsuruta , t. matsuura and y. kuroda : j. phys .
* 71 * ( 2002 ) 1640 ; a. kobayashi , a. tsuruta , t. matsuura and y. kuroda : j. phys .
* 68 * ( 1999 ) 2506 .
y. yanase and k. yamada : j. phys .
* 70 * ( 2001 ) 1659 .
k. kanoda : physica c * 282 - 287 * ( 1997 ) 299 . h. kino and h. fukuyama : j. phys
. soc . jpn .
* 65 * ( 1996 ) 2158 .
h. kino , and h. kontani : j. phys .
. jpn . * 67 * 3691 ( 1998 ) .
h. kondo and t. moriya : j. phys .
. jpn . * 67 * 3695 ( 1998 ) .
j. schmalian : phys .
* 81 * ( 1998 ) 4232 .
h. kontani and h. kino : phys .
b * 63 * ( 2001 ) 134524 . c. petrovic , p.g .
pagliuso , m.f .
hundley , r. movshovich , j.l .
sarrao , j.d .
thompson , z. fisk and p. monthoux : j. phys . condens .
matter * 13 * ( 2001 ) l337 .
h. hegger , c. petrovic , e.g. moshopoulou , m.f .
hundley , j.l .
sarrao , z. fisk , and j.d .
thompson : phys .
* 84 * ( 2000 ) 4986 . c. petrovic , r. movshovich , m. jaime , p.g .
pagliuso , m.f .
hundley , j.l .
sarrao , z. fisk , and j.d .
thompson : europhys .
* 53 * ( 2001 ) 354 .
zeng , k. tanabe , t. mito , s. kawasaki , y. kitaoka , d. aoki , y. haga and y. onuki : phys .
* 86 * ( 2001 ) 4664 .
y. kawasaki , s. kawasaki , m. yashima , t. mito , g .-
q zheng , y. kitaoka , h. shishido , r. settai , y. haga and y. onuki : j. phys .
* 72 * ( 2003 ) 2308 .
k. izawa , h. yamaguchi , y. matsuda , h. shishido , r. settai and y. onuki : phys .
* 87 * ( 2001 ) 057002 .
y. matsuda , k. izawa and i. vekhter , j. phys .
: condens .
matter * 18 * ( 2006 ) r705 .
r. movshovich , m. jaime , j.d .
thompson , c. petrovic , z. fisk , p.g .
pagliuso and j.l .
sarrao : phys .
* 86 * ( 2001 ) 5152 .
h. aoki , t. sakakibara , h. shishido , r. settai , y. onuki , p. miranovic and k. machida : j phys . :
condens matter * 16 * ( 2004 ) l13 .
radovan , n.a .
fortune , t.p .
murphy , s.t .
hannahs , e.c .
palm , s.w . tozer and d. hall : nature * 425 * ( 2003 ) 51 , a. bianchi , r. movshovich , c. capan , p.g .
pagliuso , j.l .
sarrao : phys .
lett . * 91 * ( 2003 ) 187004 , t. watanabe , y. kasahara , k. izawa , t. sakakibara , y. matsuda , c.j .
van der beek , t. hanaguri , h. shishido , r. settai and y. onuki : phys .
b * 70 * ( 2004 ) 020506 , k. kakuyanagi , m. saitoh , k. kumagai , s. takashima , m. nohara , h. takagi and y. matsuda : phys
* 94 * ( 2005 ) 047602 ; y. matsuda and h. shimahara , j. phys .
* 76 * ( 2007 ) 051005 .
r. bel , k. behnia , y. nakajima , k. izawa , y. matsuda , h. shishido , r. settai and y. onuki : phys .
* 92 * ( 2004 ) 217002 ; note that the sign of @xmath22 in cecoin@xmath3 is _ positive _ according to the definition of the present paper [ eq . ( [ eqn : nu1 ] ) ] .
see also ref .
. s. onari , h. kontani and y. tanaka : phys .
b * 73 * ( 2006 ) .
a. georges , g. kotliar , w. krauth , and m. j. rozenberg : rev .
* 68 * ( 1996 ) 13 .
g. kotliar and d. vollhardt : physics today * 57 * ( 2004 ) 53 .
j. feldman , d. lehmann , h. knorrer and e. trubowitz : _ constructive physics , ed .
v. rivasseau _ , ( springer , berlin , 1995 ) ; t. tohyama , phys . rev .
b * 70 * , 174517 ( 2004 ) .
e. dagotto , rev .
66 * , 763 ( 1994 ) .
t. yanagisawa , s. koike , and k. yamaji , phys .
b * 64 * , 184509 ( 2001 ) .
h. yokoyama and m. ogata , j. phys .
* 65 * ( 1996 ) 3615 ; h. yokoyama and m. ogata and y. tanaka , j. phys .
soc . jpn .
* 75 * ( 2006 ) 114706 .
assaad and m. imada , j. phys .
* 65 * ( 1996 ) 189 .
r. preuss , w. hanke , c. grober , and h. g. evertz , phys .
lett . * 79 * , 1122 - 1125 ( 1997 ) c. m. varma , p.b .
littlewood , s. schmitt - rink , e. abrahams , and a.e .
ruckenstein , phys .
lett . * 63 * , 1996 ( 1989 ) . c. m. varma , phys .
b * 55 * , 14554 ( 1997 ) .
a.j . millis , h. monien , and d. pines : phys .
b * 42 * , 167 ( 1990 ) .
v. barzykin and d. pines : phys .
b * 52 * ( 1995 ) 13585 . e.m .
motoyama , g. yu , i.m .
vishik , o.p .
vajk , p.k .
mang , and m. greven , nature * 445 * , 186 ( 2007 ) .
t. moriya : * spin fluctuations in itinerant electron magnetism * ( springer - verlag , berlin , 1985 ) .
t. dahm and l. tewordt : phys .
b * 52 * ( 1995 ) 1297 .
t. takimoto and t. moriya : j. phys .
* 67 * ( 1994 ) 3570 .
s. koikegami , s. fujimoto and k. yamada , j. phys .
* 66 * ( 1997 ) 1438 .
h. kontani and k. ueda : phys .
* 80 * ( 1998 ) 5619 .
s. wermbter : phys .
b * 55 * ( 1997 ) r10149 .
d. manske , i. eremin and k.h .
bennemann : phys .
b * 67 * ( 2003 ) 134520 .
h. kontani and m. ohno : phys .
b * 74 * , 014406 ( 2006 ) ; h. kontani and m. ohno : j. mag .
mag . mat . *
310 * ( 2007 ) 483 .
j. yu , s. massidda , a.j .
freeman and d.d .
koelling : phys .
a * 122 * 203 .
s. massidda , n. hamada , j. yu , and a.j .
freeman : physica c * 157 * ( 1989 ) 571 .
freeman and j. yu : physica b * 150 * ( 1988 ) 50 .
watson - yang , j. yu and a.j .
freeman : phys .
a * 120 * ( 1987 ) 489 .
campuzano , g. jennings , m. faiz , l. beaulaigue , b.w .
veal , j.z .
liu , a.p .
paulikas , k. vandervoort , h. claus , r.s .
list , a.j .
arko and r.j .
bartlett : phys .
* 64 * ( 1990 ) , 2308 .
a. ino , c. kim , m. nakamura , t. yoshida , t. mizokawa , a. fujimori , z .- x .
shen , t. kakeshita , h. eisaki , and s. uchida , phys .
b * 65 * , 094504 ( 2002 ) .
t. tanamoto , h. kohno , and h. fukuyama : j. phys .
jpn * 61 * ( 1992 ) 1886 .
m. lindroos , s. sahrakorpi , v. arpiainen , r.s .
markiewicz , and a. bansil , j. phys .
solids * 67 * ( 2006 ) 244 .
g. baym and l.p .
kadanoff : phys . rev . * 124 * ( 1961 ) , 287 .
g. baym : phys . rev .
* 127 * ( 1962 ) , 1391 .
h. kino , h. kontani and t. miyazaki , j. phy .
. jpn . * 73 * ( 2004 ) 25 ; ibid * 75 * ( 2006 ) 104702 . s. onari , r. arita , k. kuroki , and h. aoki , phys .
b * 68 * , 024525 ( 2003 ) . c. honerkamp and m. salmhofer , phys . rev .
lett . * 87 * , 187004 ( 2001 ) .
halboth and w. metzner , phys .
b * 61 * , 7364 ( 2000 ) .
j. kishine and k. yonemitsu , int .
b 16 , 711 ( 2002 ) .
h. kontani , unpublished .
h. kohno and k. yamada : prog .
. phys . * 85 * , 13 ( 1991 ) .
a. rosch , phys .
* 82 * , 4280 ( 1999 ) .
m. abdel - jawad , m.p .
kennett , l. balicas , a. carrington , a.p .
mackenzie , r.h .
mckenzie and n.e .
hussey : nature physics * 2 * , 821 ( 2006 ) ; m. abdel - jawad , j.g .
analytis , l. balicas , a. carrington , j.p.h .
charmant , m.m.j .
french , n.e .
hussey : phys .
lett . * 99 * , 107002 ( 2007 ) .
a. narduzzo , g. albert , m. m. j. french , n. mangkorntong , m. nohara , h. takagi , n.e .
hussey : arxiv:0707.4601 .
k. yamaji : j. phys .
* 58 * ( 1989 ) 1520 ; r. yagi , y. iye , t. osada and s. kagoshima , j. phys .
. jpn . * 59 * , 3069 ( 1990 ) .
h. nakano : int .
b * 7 * ( 1993 ) 2397 .
r. kubo : j. phys .
* 12 * ( 1957 ) 570 .
m. s. green : j. chem
. phys . * 20 * ( 1952 ) 1281 ; ibid * 22 * ( 1954 ) 398 .
a. rosch , ann .
( leipzig ) 15 , 526 ( 2006 ) .
y. fukuzumi , k. mizuhashi , k. takenaka , and s. uchida : phys .
* 76 * ( 1996 ) 684 .
s. d. obertelli , j.r .
cooper and j.l .
tallon : phys .
b * 46 * ( 1992 ) 14928 . c. p. popoviciu and j. l. cohn , phys . rev .
b * 55 * , 3155 ( 1997 ) .
a. yamamoto , w .- z .
hu , and s. tajima , phys .
b * 63 * , 024504 ( 2000 )
. a. yamamoto , k. minami , w .- z .
hu , a. miyakita , m. izumi , and s. tajima , phys .
b * 65 * , 104505 ( 2002 ) .
t. honma , p. h. hor , h. h. hsieh , and m. tanimoto , phys .
b * 70 * , 214517 ( 2004 ) .
t. takemura , t. kitajima , t. sugaya , i. terasaki , t. takemura , t. kitajima and t. sugaya , j. phys . : condens .
matter * 12 * ( 2000 ) 6199 .
g. v. williams , r. dupree , a. howes , s. kramer , h. j. trodahl , c. u. jung , min - seok park , and sung - ik lee , phys .
b * 65 * , 224520 ( 2002 ) .
g. hildebrand , t. j. hagenaars , w. hanke , s. grabowski , and j. schmalian , phys .
b * 56 * , r4317 ( 1997 ) .
j. m. luttinger : phys . rev . * 135 * ( 1964 ) a1505 . m. jonson and g. d. mahan : phys .
b * 42 * ( 1990 ) 9350 ; g. d. mahan : solid state physics , * 51 * ( 1998 ) 81 . h. kontani , k. kanki and k. ueda , j. phy . chem .
* 62 * ( 2001 ) 83 .
m. tsuji : j. phys .
* 13 * ( 1958 ) 979 .
ong : phys .
b * 43 * ( 1991 ) 913 .
i. tsukada and s. ono , phys .
b * 74 * , 134508 ( 2006 ) . w. kohn , phys . rev . * 123 * , 1242 ( 1961 ) .
y. takada and t. ando : j. phys .
* 44 * ( 1978 ) 905 .
k. kanki and k. yamada : j. phys .
* 66 * ( 1997 ) 1103 .
t. okabe : j. phys .
* 67 * ( 1998 ) 2792 .
t. jujo , j. phys .
* 70 * ( 2001 ) 1349 .
t.p . devereaux and a.p .
kampf , phys .
b * 59 * , 6411 ( 1999 ) .
z. konstantinovic , z. z. li , and h. raffy , phys .
b * 62 * , r11989 ( 2000 ) . e.w .
carlson , v.j .
emery , s.a .
kivelson , d. orgad , _ the physics of superconductors , vol . 2 _ , ed .
bennemann and j. b. ketterson , ( springer , berlin , 2002 ) .
s. tewari , h.y .
kee , c. nayak and s. chakravarty , phys .
b * 64 * , 224516 ( 2001 ) .
ong , y. wang , t , kakeshita and uchida : nature * 406 * ( 2000 ) 486 ; y. wang , z. a. xu , t. kakeshita , s. uchida , s. ono , y. ando , and n. p. ong : phys .
b * 64 * ( 2001 ) 224519 ; y. wang , l. li , and n. p. ong , phys . rev .
b * 73 * , 024510 ( 2006 ) . c.c .
capan et al .
* 88 * ( 2002 ) 056601 .
f. rullier - albenque , r. tourbot , h. alloul , p. lejay , d. colson , and a. forget , phys .
lett . * 96 * , 067002 ( 2006 ) .
i. ussishkin , s. l. sondhi , and d. a. huse phys .
lett . * 89 * , 287001 ( 2002 ) . s .- k
. yip , phys . rev .
b * 41 * , 2612 ( 1990 ) .
t. nakano , m. oda , c. manabe , n. momono , y. miura , and m. ido , phys .
b * 49 * , 16000 ( 1994 ) .
k. takenaka et al . :
phys . rev .
* b50 * ( 1994 ) 6534 .
t. noda , h. eisaki and s. uchida , science * 286 * ( 1999 ) 265 .
z. a. xu , j. q. shen , s. r. zhao , y. j. zhang , and c. k. ong , phys .
b * 72 * , 144527 ( 2005 ) .
l. b. rigal , d. c. schmadel , h. d. drew , b. maiorov , e. osquiguil , j. s. preston , r. hughes , and g. d. gu : phys .
* 93 * ( 2004 ) 137002 .
m. grayson , l. b. rigal , d. c. schmadel , h. d. drew , and p .- j .
kung : phys .
* 89 * ( 2002 ) 037003 .
j. cerne , m. grayson , d. c. schmadel , g. s. jenkins , h. d. drew , r. hughes , a. dabkowski , j. s. preston , and p .- j .
kung : phys .
* 84 * ( 2000 ) 3418 .
kaplan , s.wu , h .- t.s .
lihn , h.d . drew , q. li , d.b .
fenner , j.m .
phillips and s.y .
hou : phys .
* 76 * ( 1996 ) 696 .
schmadel , g.s .
jenkins , j.j .
gu , h. kontani , and h.d
. drew , phys .
b * 75 * , 140506(r ) ( 2007 ) .
l. shi , d. schmadel , h. d. drew , i. tsukada , and y. ando , cond - mat/0510794 .
a. zimmers , l. shi , d. c. schmadel , w. m. fisher , r. l. greene , h. d. drew , m. houseknecht , g. acbas , m .- h .
kim , m .- h .
yang , j. cerne , j. lin , and a. millis , hys .
b * 76 * , 064515 ( 2007 ) .
h. kontani , j. phys .
jpn * 75 * ( 2006 ) 013703 .
h. kontani , j. phys .
jpn ( 2007 ) 074707 .
h. d. drew and p. coleman phys .
lett . * 78 * , 1572 ( 1997 ) .
k. mcelroy , jinho lee , j. a. slezak , d .- h .
lee , h. eisaki , s. uchida , and j. c. davis , science * 309 * ( 2005 ) 1048 .
andersen , a. melikyan , t.s .
nunner , and p.j .
hirschfeld , phys .
b * 74 * , 060501(r ) ( 2006 ) .
fang , l. capriotti , d.j .
scalapino , s.a .
kivelson , n. kaneko , m. greven , and a. kapitulnik , phys .
lett . * 96 * , 017007 ( 2006 ) .
h. eisaki , n. kaneko , d.l .
feng , a. damascelli , p.k .
mang , k.m .
shen , z .- x .
shen , and m. greven , phys .
b * 69 * , 064512 ( 2004 ) .
wang , p. j. hirschfeld , and h .-
cheng , phys .
b * 72 * ( 2005 ) 224516 .
a. v. mahajan , h. alloul , g. collin , and j. f. marucco : phys .
* 72 * ( 1994 ) 3100 .
w. a. macfarlane , j. bobroff , h. alloul , p. mendels , n. blanchard , g. collin , and j .- f .
marucco : phys .
* 85 * ( 2000 ) 1108 .
a. v. mahajan , h. alloul , g. collin , j. f.marucco , g. collin , and j .- f .
marucco : eur .
j. * b * 13 ( 2000 ) 457 .
j. bobroff , w. a. macfarlane , h. alloul , p. mendels , n. blanchard , g. collin , and j .- f .
marucco : phys .
* 83 * ( 1999 ) 4381 .
julien , t. feher , m. horvatic , c. berthier , o.n .
bakharev , p. segransan , g. collin , j .- f .
marucco : phys .
* 84 * ( 2000 ) 3422 .
j. riera and s. koval , d. poilblanc and f. pantigny , phys .
b * 54 * ( 1996 ) 4771 .
t. tohyama , m. takahashi and s. maekawa : physica c * 357 - 360 * ( 2001 ) 93 .
h. tsuchiura , y. tanaka , m. ogata , and s. kashiwaya : phys .
b * 64 * ( 2001 ) 140501(r ) .
p. mendels , j. bobroff , g. collin , h. alloul , m. gabay , j.f .
marucco , n. blanchard and b. grenier : europhys . lett .
* 46 * ( 1999 ) 678 .
k. ishida , y. kitaoka , k. yamazoe , k. asayama , and y. yamada : phys .
* 76 * ( 1996 ) 531 .
r. s. islam , j. r. cooper , j. w. loram , s. h. naqib , cond - mat/0607153 n. bulut , d. hone , d.j .
scalapino , and e.y .
loh : phys .
* 62 * ( 1989 ) 2192 .
sandvik , e. dagotto , and d.j .
scalapino : phys .
b * 56 * ( 1997 ) 11701 . g.b .
martins , m. laukamp , j. riera and e. dagotto : phys .
* 78 * ( 1997 ) 3563 .
m. laukamp et al .
b * 57 * ( 1998 ) 10755 . phys .
lett . * 96 * , 017204 ( 2006 ) .
fabrizio anfuso and sebastian eggert : phys .
lett . * 96 * , 017204 ( 2006 ) .
n. bulut : physica c * 363 * ( 2001 ) 260 .
n. bulut : phys .
b * 61 * ( 2000 ) 9051 .
y. ohashi : j. phys .
* 70 * ( 2001 ) 2054 .
y. ando , g.s .
boebinger , a. passner , t. kimura and k. kishio : phys .
* 75 * ( 1995 ) 4662 .
boebinger , y. ando , a. passner , t. kimura , m. okuya , j. shimoyama , k. kishio , k. tamasaku , n. ichikawa , and s. uchida : phys .
* 77 * ( 1996 ) 5417 .
t. sekitani , m. naito , and n. miura : phys .
b * 67 * ( 2003 ) 174503 .
d. jaccard and j. sierro , physica b * 206 - 207 * ( 1995 ) 625 .
h. wilhelm , s. raymond , d. jaccard , o. stockert , h.v .
lohneysen and a. rosch , j. phys . : condens .
matter * 13 * l329 .
k. miyake and o. narikiyo : j. phys .
* 71 * ( 2002 ) 867 .
holmes , d. jaccard and k. miyake , phys .
b * 69 * , 024508 ( 2004 ) .
h. kontani and m. ohno , unpublished .
s. ouazi , j. bobroff , h. alloul , m. le tacon , n. blanchard , g. collin , m. h. julien , m. horvati , and c. berthier , phys .
lett . * 96 * , 127005 ( 2006 ) .
harter , b.m .
andersen , j. bobroff , m. gabay and p.j .
hirshfeld , phys .
b * 75 * ( 2007 ) 054520 .
p. coleman , _ handbook of magnetism and advanced magnetic materials , vol . 1 _ ( j. wiley and sons , 2007 ) ; cond - mat/0612006 .
j. paglione , m. a. tanatar , d. g. hawthorn , e. boaknin , r. w. hill , f. ronning , m. sutherland , l. taillefer , c. petrovic , and p. c. canfield , phys .
lett . * 91 * , 246405 ( 2003 ) .
k. sakurazawa , h. kontani and t. saso : j. phys .
. jpn . * 74 * , 271 ( 2004 ) .
p. li , f.f .
balakirev and r.l .
greene , phys .
lett . * 99 * , 047003 ( 2007 ) .
s. onari , h. kontani and y. tanaka : to be published in j. phys .
* 77 * ( 2008 ) .
k. miyagawa and h. taniguchi , private communication .
h. taniguchi , t. okuhara , t. nagai , k. satoh , m. mori , y. himizu , m. hedo and y. uwatoko , j. phys .
* 76 * ( 2007 ) no.11 .
rosenbaum , a. husmann , s.a .
carter and j.m honig : phys .
b * 57 * ( 1998 ) r13997 .
s. yoshii , k. murata and m. sato , j. phys .
* 69 * ( 2000 ) 17 ; s. yoshii , k. murata and m. sato , j. phys . chem .
solids * 62 * ( 2001 ) 129 .
renner , b. revaz , j .- y .
genoud , k. kadowaki and .
fischer , phys .
* 80 * ( 1998 ) 149 .
n. miyakawa , p. guptasarma , j. f. zasadzinski , d. g. hinks and k. e. gray , phys .
* 80 * ( 1998 ) 157 .
r. m. dipasupil , m. oda , n. momono and m. ido , j. phys .
* 71 * ( 2002 ) 1535 .
t. sato , t. yokoya , y. naitoh , t. takahashi , k. yamada and y. endoh , phys .
* 83 * ( 1999 ) 2254 .
h. alloul , t. ohno and p. mendels , phys .
* 63 * ( 1989 ) 1700 .
k. ishida , y. kitaoka , g .- q .
zheng and k. asayama , j. phys .
jpn * 60 * ( 1991 ) 3516 .
d. c. johnston , phys .
* 62 * ( 1989 ) 957 .
m. oda , h. matsuki and m. ido , solid state commun .
* 74 * ( 1990 ) 1321 .
b. holm and f. aryasetiawan : phys .
b * 56 * ( 1997 ) 12825 .
b. holm and u. von barth : phys .
b * 57 * ( 1998 ) 2108 . p. garcia - gonzalez and r.w .
godby : phys .
b * 63 * ( 2001 ) 075112 .
j. schmalian , d. pines , and b. stojkovi , phys .
b * 60 * , 667 ( 1999 ) ; j. schmalian , d. pines , and b. stojkovi , phys . rev .
lett . * 80 * , 3839 ( 1998 ) .
s. fujimoto , j. phys .
* 71 * ( 2002 ) 1230 .
h. ikeda , s. shinkai and k. yamada , submitted to j. phys .
s. shinkai , h. ikeda and k. yamada , j. phys .
* 74 * ( 2005 ) 2592 ; _ ibid _ * 75 * ( 2006 ) 104712 . t. nomura and k. yamada : j. phys .
* 69 * ( 2000 ) 3678 ; _ ibid _ * 71 * ( 2002 ) 1993 .
m. vilk and a .- m.s .
tremblay , j. phys .
i france * 7 * , 1309 ( 1997 ) .
h. kusunose , j. phys .
* 75 * ( 2006 ) 054713 .
t. saso : j. phys .
* 68 * ( 1999 ) 3941 .
k. izawa , k. behnia , y. matsuda , h. shishido , r.settai , y. onuki , j. flouquet , phys .
lett . * 99 * , 147005 ( 2007 ) .
t. shibauchi , private communication .
tanatar , j. paglione , c. petrovic , and l. taillefer , science * 316 * , 1320 ( 2007 ) .
h. ding , m. r. norman , t. yokoya , t. takeuchi , m. randeria , j. c. campuzano , t. takahashi , t. mochiku , and k. kadowaki , phys .
lett . * 78 * , 2628 ( 1997 ) .
m. r. norman , m. randeria , h. ding , and j. c. campuzano , phys .
b * 57 * , r11093 ( 1998 ) .
u. chatterjee , m. shi , a. kaminski , a. kanigel , h. m. fretwell , k. terashima , t. takahashi , s. rosenkranz , z. z. li , h. raffy , a. santander - syro , k. kadowaki , m. r. norman , m. randeria , and j. c. campuzano : phys .
lett . * 96 * , 107006 ( 2006 ) .
t. yoshida , x. j. zhou , t. sasagawa , w. l. yang , p. v. bogdanov , a. lanzara , z. hussain , t. mizokawa , a. fujimori , h. eisaki , z .- x .
shen , t. kakeshita , and s. uchida : phys .
lett . * 91 * , 027001 ( 2003 ) .
u. chatterjee , m. shi , a. kaminski , a. kanigel , h. m. fretwell , k. terashima , t. takahashi , s. rosenkranz , z. z. li , h. raffy , a. santander - syro , k. kadowaki , m. r. norman , m. randeria , and j. c. campuzano , phys .
lett . * 95 * , 107006 ( 2006 ) .
h. matsui , k. terashima , t. sato , t. takahashi , s .- c .
wang , h .- b .
yang , h. ding , t. uefuji , and k. yamada , phys .
lett . * 94 * , 047005 ( 2005 ) .
s. ono , s. komiya , and y. ando , phys .
b * 75 * , 024515 ( 2007 ) .
r. karplus and j. m. luttinger : phys . rev .
* 95 * ( 1954 ) 1154
. j. m. luttinger : phys . rev .
* 112 * ( 1958 ) 739 .
j. smit : physica * 24 * ( 1958 ) 39 .
y. nuki , t. yamayoshi , i. ukon , t. komatsubara , a. umezawa , w. k. kwok , g. w. crabtree and d. g. hinks : j. phys .
* 58 * ( 1989 ) 2119 .
y. nuki , t. yamayoshi , t. omi , i. ukon , a. kobori and t. komatsubara : j. phys .
* 58 * ( 1989 ) 2126 .
k. hanzawa , k. yosida and k. yamada , prog .
* 81 * ( 1989 ) 960 .
h. kontani and k. yamada : j. phys .
* 63 * ( 1994 ) 2627 .
t. hiraoka , t. sada , t. takabatake and h. fujii , physica b * 186 - 188 * ( 1993 ) 7603 .
p. coleman , p. w. anderson and t. v. ramakrishman : phy .
* 55 * ( 1985 ) 414 .
a. fert and p. m. levy : phy .
b * 36 * ( 1987 ) 1907 . h. kontani and k. yamada : j. phys .
* 66 * ( 1997 ) 2252 .
t. miyasato , n. abe , t. fujii , a. asamitsu , s. onoda , y. onose , n. nagaosa , y. tokura : cond - mat/0610324 . m. miyazawa , h. kontani and k. yamada : j. phys .
* 68 * ( 1999 ) 1625 .
h. kontani , t. tanaka , k. yamada , phys .
b * 75 * , 184416 ( 2007 ) .
j. jin et al . , cond - mat/0112405 ; l.m .
galvin et al . , phys .
b * 63 * , 161102(r ) ( 2001 ) .
y. yao , l. kleinman , a.h .
macdonald , j. sinova , t. jungwirth , d. wang , e. wang and q. niu : phys .
* 92 * ( 2004 ) 037204 .
z. fang et al . :
science * 302 * ( 2003 ) 92 .
kontani , t. tanaka , d.s .
hirashima , k. yamada , and j. inoue , cond - mat/0702447 .
kontani , m. naito , d.s .
hirashima , k. yamada , and j. inoue , j. phys .
* 76 * ( 2007 ) 103702 .
t. kimura , y. otani , t. sato , s. takahashi , and s. maekawa , phys .
lett . * 98 * , 156601 ( 2007 ) .
t. tanaka , h. kontani , m. naito , t. naito , d.s .
hirashima , k. yamada and j. inoue , unpublished .
s. murakami , n. nagaosa , and s. c. zhang , science * 301 * , 1348 ( 2003 ) .
j. sinova et al . , phys .
lett . * 92 * , 126603 ( 2004 ) .
k. kadowaki and s.b .
woods : solid state commun .
* 58 * ( 1986 ) 507 .
h. kontani : j. phy .
* 73 * ( 2004 ) 515 .
h. kontani : physica b. * 359 - 361 * ( 2005 ) 202 .
k. yamada : prog .
* 53 * ( 1975 ) 970 ; k. yamada and k. yosida : prog .
* 53 * ( 1975 ) 1286 ; k. yamada and k. yosida : prog .
* 54 * ( 1975 ) 316 . in refs .
@xcite , we put @xmath1352 by mistake ; the correct relation is @xmath1353 .
for this reason , eq .
( [ eqn : g - kw ] ) is two times eq .
( 8) in ref . @xcite . since the author of ref .
@xcite assumed @xmath1354 cm , the grand - kw relation in ref .
@xcite is equal to eq .
( [ eqn : kw ] ) in the present paper .
n. tsujii , h. kontani and k. yoshimura : phys .
* 94 * ( 2005 ) 057201 .
k. behnia , d. jaccard and j. flouquet : j. phys . : condens .
matter * 16 * ( 2004 ) 5187 .
k. miyake and h. kohno : j. phy .
* 74 * ( 2005 ) 254 .
m. s. torikachvili , s. jia , e. d. mun , s. t. hannahs , r. c. black , w. k. neils , dinesh martien , s. l. budko , p. c. canfield , proc .
u. s. a. * 104 * , 9960 ( 2007 ) .
h. kontani and k. yamada : j. phys .
* 65 * ( 1996 ) 172 .
h. kontani and k. yamada : j. phys .
* 66 * ( 1997 ) 2232 .
h. okamura et al.:j .
* 76 * ( 2007 ) 023703 .
h. kontani : jpsj online - news and comments [ feb . 13 , 2007 ] | in this paper , we present a recent developments in the theory of transport phenomena based on the fermi liquid theory . in conventional metals , various transport coefficients are scaled according to the quasiparticles relaxation time , @xmath0 , which implies that the relaxation time approximation ( rta ) holds well .
however , such a simple scaling does not hold in many strongly correlated electron systems .
the most famous example would be high-@xmath1 superconductors ( htscs ) , where almost all the transport coefficients exhibit a significant deviation from the rta results .
this issue has been one of the most significant unresolved problems in htscs for a long time .
similar anomalous transport phenomena have been observed in metals near their antiferromagnetic ( af ) quantum critical point ( qcp ) .
the main goal of this study is to demonstrate whether the anomalous transport phenomena in htsc is the evidence of non - fermi liquid ground state , or it is just the rta violation in strongly correlated fermi liquids .
another goal is to establish a unified theory of anomalous transport phenomena in metals with strong magnetic fluctuations . for these purposes ,
we develop a method for calculating various transport coefficients beyond the rta by employing field theoretical techniques . in a fermi liquid
, an excited quasiparticle induces other excited quasiparticles by collision , and current due to these excitations is called a current vertex correction ( cvc ) .
landau had noticed the existence of cvc first , which is indispensable for calculating transport coefficient in accord with the conservation laws . here , we develop a transport theory involving resistivity and hall coefficient on the basis of the microscopic fermi liquid theory , by considering the cvc . in nearly af fermi liquids , we find that the strong backward scattering due to af fluctuations induces the cvc with prominent momentum dependence .
this feature of the cvc can account for the significant enhancements in the hall coefficient , magnetoresistance , thermoelectric power , and nernst coefficient in nearly af metals . according to the present numerical study , aspects of anomalous transport phenomena in htsc
are explained _ in a unified way _ by considering the cvc , without introducing any fitting parameters ; this strongly supports the idea that htscs are fermi liquids with strong af fluctuations .
further , the present theory also explains very similar anomalous transport phenomena occurring in ce@xmath2in@xmath3 ( @xmath2=co or rh ) , which is a heavy - fermion system near the af qcp , and in the organic superconductor @xmath4-(bedt - ttf ) .
in addition , the striking @xmath5-dependence of the ac hall coefficient and the remarkable effects of impurities on the transport coefficients in htscs appear to fit naturally into the present theory .
many aspects of the present theory are in accord with the anomalous transport phenomena in htscs , organic superconductors , and heavy - fermion systems near their af qcps .
we discuss some of the open questions for future work . |
cet article tudie deux quantits associes un fibr adlique hermitien : son minimum absolu et sa pente maximale .
nous donnons plusieurs ingalits faisant intervenir ces deux nombres , en particulier pour estimer leur comportement relativement au produit tensoriel des fibrs et , subsquemment , aux puissances symtriques et extrieures
. la notion de fibr adlique hermitien gnralise un corps de nombres celle de rseau euclidien avec laquelle elle concide sur @xmath3 .
alors quun rseau euclidien est constitu dun @xmath4-module libre de rang fini @xmath5 et dune norme euclidienne sur lespace vectoriel rel @xmath6 , nous voyons plutt un fibr adlique hermitien sur un corps de nombres @xmath7 comme un couple form dun espace vectoriel @xmath8 sur @xmath7 et dune collection de normes @xmath9 sur @xmath8 pour toute place @xmath10 de @xmath7 ( la dfinition prcise est rappele dans la partie suivante ) : dans le cas dun rseau euclidien @xmath11 , la norme @xmath12 est celle donne sur @xmath6 tandis que les normes aux places finies dfinissent @xmath5 dans @xmath8 . on associe classiquement un rseau euclidien son ( premier ) minimum : la plus petite norme dun lment non nul de @xmath5 .
traduite dans le langage des fibrs adliques hermitiens , cette dfinition devient la plus petite hauteur dun lment non nul de @xmath8 ( la hauteur sobtient en faisant le produit des normes , voir [ defh ] ) mais nous souhaitons considrer ici le minimum _ absolu _ de @xmath8 cest - - dire linfimum des hauteurs des lments non nuls de @xmath13 : nous le notons @xmath14 .
nous utilisons ensuite la pente dun fibr adlique hermitien .
dans le cas dun rseau euclidien , celle - ci est un avatar du covolume : si @xmath15 est dduit de @xmath16 alors @xmath17 o le volume dsigne la mesure de haar sur @xmath6 associe la norme @xmath18 ( cest - - dire la mesure de lebesgue si lon identifie @xmath6 @xmath19 _ via _ une base orthonorme pour @xmath18 ) .
cette dfinition stend un corps de nombres et nous nous concentrons sur la pente maximale @xmath20 de @xmath21 , cest - - dire le maximum des pentes des sous - fibrs non nuls de @xmath21 .
tandis quune comparaison entre le minimum et le covolume dun rseau euclidien fait lobjet du premier thorme de minkowski ( appel aussi lemme de siegel dans certains contextes ) , une comparaison analogue pour le minimum absolu rsulte dun thorme profond de zhang , qui joue un rle crucial dans ce texte .
nous citons ici une forme lgrement simplifie ( voir la partie [ pzh ] pour une discussion plus complte ) .
[ tzhi ] pour tout fibr adlique hermitien @xmath21 de dimension @xmath22 sur un corps de nombres nous avons @xmath23 ici la minoration rsulte facilement des dfinitions , le contenu du thorme rside dans la majoration qui mrite le nom de thorme de minkowski absolu ou de lemme de siegel absolu ( suivant la terminologie de roy - thunder @xcite ) .
nous montrons dans un premier temps que lordre de grandeur de cette majoration ne peut pas tre amlior .
un rsultat analogue sur @xmath3 portant classiquement le nom de thorme de minkowski - hlawka , notre premier rsultat correspond donc un thorme de minkowski - hlawka absolu .
[ thlawka ] pour tout @xmath22 il existe un fibr adlique hermitien @xmath21 de dimension @xmath24 sur un corps de nombres avec @xmath25 il sagit ici encore dune forme simplifie ( voir thorme [ mha ] ) et nous donnerons un procd explicite pour trouver un tel @xmath21 . tournons - nous prsent vers le comportement vis - - vis du produit tensoriel . en utilisant le thorme [ tzhi ] ( de manire indirecte ) nous dmontrons lingalit centrale suivante . [ prin ] pour deux fibrs adliques hermitiens @xmath21 et @xmath26 sur un corps de nombres , on a @xmath27 cet nonc stend immdiatement par rcurrence @xmath28 fibrs puis en le combinant ( de manire directe cette fois ) avec le thorme [ tzhi ] nous aboutissons aux estimations suivantes . [ cor ] soient @xmath29 un entier naturel et @xmath30 des fibrs adliques hermitiens sur un corps de nombres .
nous avons@xmath31 @xmath32 lingalit a dj fait lobjet de plusieurs travaux .
dans le seul publi ce jour , chen @xcite obtient comme majorant @xmath33 .
avec le coefficient @xmath34 , notre majoration tait connue dans le cas @xmath35 , tablie indpendamment par bost et andr @xcite .
en outre , dans un travail en cours , bost et chen @xcite donnent une version plus forte de o le majorant devient @xmath36 et prouvent donc lexact analogue de .
dans le corollaire , les minorations suivent facilement des dfinitions mais lon peut se demander si elles peuvent tre strictes .
dans le cas une conjecture de bost prdit quil y a toujours galit : @xmath37 pour tous fibrs adliques hermitiens @xmath21 et @xmath26 .
ceci reste un problme ouvert : voir @xcite pour des rsultats partiels ; plus bas nous donnons seulement un exemple de situation o il y a galit ( en prsence dune action de groupe ) . en revanche dans le cadre de nous montrons que la conjecture analogue se trouve tre
_ fausse_. [ nml ] pour tous entiers naturels @xmath38 il existe deux fibrs adliques hermitiens sur un corps de nombres @xmath21 et @xmath39 de dimensions respectives @xmath24 et @xmath40 avec @xmath41 dans le langage de larticle dandr @xcite , ceci signifie que la proprit dtre numriquement effectif ne se conserve pas par produit tensoriel .
dans la fin du texte , nous utilisons le thorme [ prin ] pour estimer minimum et pente maximale des puissances symtriques ou extrieures dun fibr adlique hermitien .
par exemple nous amliorons la borne @xmath42 due bost o @xmath43 ( voir ( * ? ? ?
* thorme 7.1 ) ) en remplaant @xmath44 par @xmath45 .
nous montrons aussi par un exemple que ceci est le bon ordre de grandeur et ne peut pas tre remplac par @xmath46 si @xmath47 . *
remerciements*. nous remercions jean - benot bost et huayi chen ainsi que yves andr de nous avoir communiqu leurs textes @xcite et @xcite avant publication .
soit @xmath7 un corps de nombres .
notons @xmath48 lensemble des places de @xmath7 et @xmath49 les adles de @xmath7 .
pour @xmath50 , le corps @xmath51 dsigne la compltion dune clture algbrique du complt @xmath52 de @xmath7 .
nous le munissons de la valeur absolue @xmath53 qui tend la valeur absolue usuelle de @xmath54 , adhrence de @xmath3 dans @xmath52 .
la _ formule du produit _ scrit @xmath55}}=1.\ ] ] soit @xmath56 .
soit @xmath57 le couple dit _ fibr hermitien standard _
constitu de lespace vectoriel @xmath58 et dune collection de normes @xmath59 dfinies de la manire suivante : pour tout vecteur @xmath60 de @xmath61 on a @xmath62 [ definitionfva ] un _ fibr adlique hermitien _
@xmath63 sur @xmath7 , de dimension @xmath24 , est la donne dun @xmath7-espace vectoriel @xmath8 de dimension @xmath24 et dune famille de normes @xmath64 sur @xmath65 , indexes par les places @xmath10 de @xmath7 , qui satisfont la condition suivante : il existe une @xmath7-base @xmath66 de @xmath8 et une matrice adlique @xmath67 telles que , pour toute place @xmath68 , la norme sur @xmath65 est donne par @xmath69 cette dfinition est quivalente celle de _ fibr vectoriel hermitien sur @xmath70_. si @xmath71 , cette notion correspond celle de _ rseau euclidien _
au sens de la gomtrie des nombres classique .
nous renvoyons le lecteur @xcite pour les proprits de ces objets mentionnes ci - aprs ( rappelons quune base @xmath72 de @xmath73 est dite orthonorme si @xmath74 pour tout @xmath75 ) .
soit @xmath21 un fibr adlique hermitien sur un corps de nombres @xmath7 .
soient @xmath76 une extension algbrique de @xmath7 et @xmath77 un lment de @xmath78 .
soit @xmath79 une extension finie de @xmath7 telle que @xmath80 .
le @xmath81-espace vectoriel @xmath82 a une structure de fibr adlique hermitien induite par celle de @xmath21 ( pour @xmath83 au - dessus de @xmath50 on identifie @xmath84 @xmath85 ) .
nous noterons @xmath86 plutt que @xmath87 la norme de @xmath82 relative @xmath83 .
avec ces conventions , la _ hauteur normalise _ de @xmath77 est le nombre rel : @xmath88/[k':\mathbf{q}]}}.\ ] ] cette dfinition ne dpend pas de lextension finie @xmath81 choisie .
le ( premier ) _ minimum _ de @xmath8 sur @xmath76 , not @xmath89 , est le nombre rel : @xmath90 on remarque que @xmath91 o @xmath81 parcourt les extensions finies de @xmath7 incluses dans @xmath76 .
dans la suite nous voyons toujours @xmath7 comme un sous - corps de @xmath92 et nous intressons principalement @xmath93 ( not @xmath94 dans @xcite ) .
par ailleurs , si @xmath95 est une application linaire entre deux espaces munis de structures adliques hermitiennes alors , en chaque place @xmath10 de @xmath7 , lon dispose de la norme doprateur @xmath96 de lapplication @xmath97 induite par @xmath98 .
la dfinition de cette norme implique immdiatement que si @xmath98 est _ injective _ alors @xmath99/[k:\mathbf{q}]}}.\ ] ] tant donn un fibr hermitien @xmath21 sur @xmath7 donn par une matrice adlique @xmath100 , la _ pente ( normalise ) _ @xmath101 de @xmath21 est le nombre rel @xmath102 } { [ k:\mathbf{q}]}\log\vert\det a_{v}\vert_{v}}.\ ] ] la _ pente maximale _
@xmath20 de @xmath21 est le maximum des pentes @xmath103 lorsque @xmath104 parcourt les sous - espaces vectoriels non nuls de @xmath8 ( le fibr @xmath26 est @xmath104 muni des mtriques induites par celles de @xmath21 ) . le fibr @xmath21 est dit _ semi - stable _ si @xmath105 .
la pente maximale est atteinte en un unique sous - espace de dimension maximale et elle est invariante par extension de corps ( nous rappelons largument au paragraphe [ act ] ) .
soit @xmath21 un fibr adlique hermitien sur @xmath7 .
le _ dual _
de @xmath21 , not @xmath106 , est le fibr adlique hermitien despace sous - jacent @xmath107 muni des normes ( doprateurs ) duales de celles sur @xmath8 . on a @xmath108 ( voir ( * ? ? ?
* proposition @xmath109 ) ) .
soit @xmath26 un autre fibr adlique hermitien sur @xmath7 .
somme directe ( hermitienne ) _ @xmath110 de
@xmath21 et @xmath26 est le fibr adlique hermitien despace sous - jacent @xmath111 de normes @xmath112 pour tous @xmath113 et @xmath114 . on a @xmath115 .
soient @xmath21 et @xmath26 deux fibrs adliques hermitiens sur @xmath7 .
produit tensoriel _ @xmath116 est
le fibr adlique hermitien despace sous - jacent @xmath117 et do nt la norme sur @xmath118 ( @xmath50 ) est telle que des bases orthonormes @xmath66 de @xmath119 et @xmath120 de @xmath121 donnent une base orthonorme @xmath122 de @xmath123 . on a la formule
@xmath124 ( voir ( * ? ? ?
* proposition @xmath125 ) ) .
soient @xmath21 un fibr adlique hermitien sur @xmath7 et @xmath126 un entier .
la puissance symtrique @xmath127 de @xmath8 , note @xmath128 , est un quotient de la puissance tensorielle @xmath129 .
cet espace vectoriel est de dimension @xmath130 o @xmath43 .
il possde une structure adlique hermitienne @xmath131 dfinie de la manire suivante : soient @xmath10 une place de @xmath7 et @xmath66 une base orthonorme de @xmath132 ; la norme @xmath133 est lunique norme pour laquelle la base @xmath134 , pour @xmath135 de somme @xmath136 , est orthonorme si @xmath10 est ultramtrique et orthogonale avec @xmath137 si @xmath10 est archimdienne .
il revient au mme de dire que @xmath128 est muni de la structure quotient de celle de @xmath138 ( voir @xcite ) .
soient @xmath21 un fibr adlique hermitien sur @xmath7 de dimension @xmath139 et @xmath140 .
la puissance extrieure @xmath141 peut tre munie de normes hermitiennes aux places @xmath10 de @xmath7 en dcrtant quune base orthonorme @xmath66 de @xmath65 fournit une base orthonorme @xmath142 de @xmath143 .
ceci confre une structure de fibr hermitien @xmath144 la puissance extrieure , qui , contrairement au cas de la puissance symtrique , nest pas exactement la structure quotient que lon obtient par la projection canonique @xmath145 ( un facteur @xmath146 diffrencie les normes aux places archimdiennes ) .
la pente de ce fibr vrifie lgalit @xmath147(une dmonstration de cette formule repose sur lgalit @xmath148 o @xmath149 est un endomorphisme despace vectoriel ) .
lorsque @xmath150 , cette formule peut tre utilise comme dfinition de la pente de @xmath21 en termes de hauteurs .
en effet , dans ce cas , @xmath151 est une droite et lon a @xmath152 , pour un lment @xmath77 quelconque de @xmath153 .
[ convexe ] soient @xmath28 un entier @xmath154 et @xmath30 des fibrs adliques hermitiens sur @xmath7 .
pour tout @xmath155 , on a @xmath156 par convexit de lapplication @xmath157 , on a , pour tout ensemble @xmath158 de cardinal @xmath159 et toute famille de nombres rels positifs @xmath160 , @xmath161 par rcurrence sur @xmath28 , on en dduit que , pour tout @xmath162 , on a @xmath163 on applique alors cette ingalit lensemble @xmath158 des plongements complexes de @xmath7 et @xmath164 puis lon multiplie par les contributions ultramtriques pour conclure .
avec les notations du lemme , nous dduisons en particulier que pour toute extension algbrique @xmath76 de @xmath7 on a @xmath165
soient @xmath21 un fibr adlique hermitien de dimension @xmath22 sur un corps de nombres et @xmath166 un entier avec @xmath167 . on dfinit
le @xmath166-me minimum de zhang @xmath168 comme la borne infrieure des nombres rels @xmath169 tels que ladhrence de zariski de @xmath170 soit de dimension au moins @xmath166 .
nous avons @xmath171 et , si lon note @xmath172 le nombre harmonique , le thorme des minima successifs de zhang dans le cas des fibrs adliques hermitiens snonce comme suit . [ tzh ]
pour tout fibr adlique hermitien @xmath21 de dimension @xmath22 , on a @xmath173 ce rsultat dcoule du thorme 5.2 de @xcite avec @xmath174 et il faut valuer la hauteur de @xmath175 comme par exemple dans @xcite .
mentionnons quen toute gnralit zhang traite le cas dune varit arithmtique @xmath175 avec des mtriques adliques non ncessairement hermitiennes .
dans ce cadre largi , les ingalits du thorme sont optimales car il est possible davoir @xmath176 comme on le voit en prenant @xmath177 muni des mtriques du supremum en toutes les places ( on a @xmath178 pour tout @xmath166 car les points coordonnes dans les racines de lunit forment un ensemble dense de points de hauteur minimale 1 ) .
lorsque lon se limite au cadre hermitien comme ici ,
lon peut tout de mme voir que lordre de grandeur de lencadrement du thorme ne peut pas tre amlior .
en effet si @xmath21 est le fibr standard @xmath179 alors on a @xmath180 pour tout @xmath166 .
pour le voir on utilise le lemme [ convexe ] de convexit des hauteurs pour montrer que @xmath181 ds que @xmath77 a @xmath166 coordonnes non nulles , avec galit lorsque toutes ces coordonnes sont des racines de lunit .
dans ce cas , lassertion du thorme scrit @xmath182 o @xmath183 .
le thorme [ tzh ] implique le lemme suivant qui lui - mme entrane le thorme [ tzhi ] .
pour tout fibr adlique hermitien de dimension @xmath22 , on a @xmath184 lingalit @xmath185 fournit @xmath186 et nous appliquons cette estimation un sous - espace @xmath104 tel que @xmath187 : @xmath188 la premire minoration @xmath189 traduit simplement lingalit tautologique @xmath190 lorsque lon sait que la pente maximale est invariante par extension des scalaires .
au lieu du thorme [ tzhi ] nous aurions pu donner un nonc intermdiaire avec des minima successifs plus classiques cest - - dire dfinis comme les @xmath191 en remplaant ladhrence de zariski par lespace vectoriel engendr . on obtient alors
un lemme de siegel absolu exactement de la forme de celui de roy - thunder ( avec une meilleure constante ) .
ceci ne prsente pas dintrt pour ce qui suit car lon sait que la constante optimale est la mme dans les deux formulations ( voir @xcite et @xcite ) .
tant donn un fibr adlique hermitien @xmath21 sur @xmath7 , notons @xmath192 le nombre rel @xmath193 ( analogue de linvariant dhermite dun rseau euclidien ) .
le rsultat que nous prsentons dans ce paragraphe est une version plus prcise du thorme [ thlawka ] . [ mha ] pour tout entier @xmath139 , pour tout @xmath194 , il existe un fibr adlique hermitien @xmath21 de dimension @xmath24 ( sur un corps de nombres qui dpend de @xmath24 ) tel que @xmath195 . en particulier il existe @xmath21 de dimension @xmath24 tel que @xmath196 .
le thorme de minkowski - hlawka classique affirme que , pour tout @xmath197 , il existe un rseau euclidien @xmath198 sur @xmath199 de dimension @xmath24 avec @xmath200 o @xmath201 est le volume de la boule unit de @xmath202 . plus gnralement thunder @xcite a dmontr une minoration analogue avec un fibr @xmath198 sur un corps de nombres @xmath7 fix : @xmath203 pour un rel @xmath204 .
toutefois , contrairement ces rsultats o lexistence de lexemple sobtient par un argument de moyenne , nous donnons ici un exemple _ explicite _ de @xmath21 satisfaisant la minoration . si @xmath205 alors @xmath206 pour tout @xmath21 et le thorme est dmontr .
supposons maintenant @xmath197 .
soit @xmath207 une matrice do nt aucun mineur ( dordre quelconque ) nest nul .
par exemple la matrice de coefficient @xmath208 gal @xmath209 convient .
choisissons ensuite un nombre premier @xmath210 tel que la proprit des mineurs de @xmath211 reste vraie modulo @xmath210 ( par exemple si @xmath210 est strictement plus grand que les valeurs absolues des mineurs de @xmath211 ) .
par ailleurs , sans restriction , on peut supposer que @xmath212 .
prenons alors @xmath213 et @xmath214 tels que , pour tout @xmath215 , on a @xmath216 .
la condition sur @xmath217 entrane @xmath218 .
soit @xmath219 et , pour @xmath220 , posons @xmath221 la matrice diagonale @xmath222 de @xmath223 .
considrons le @xmath7-fibr adlique hermitien @xmath21 gal
@xmath57 sauf en les places @xmath10 de @xmath7 au - dessus de @xmath210 pour lesquelles on pose @xmath224 la pente de @xmath21 vaut @xmath225/[k:\mathbf{q}]}}=\log ( a_1\cdots a_n)^{1/n}$ ] car @xmath226 pour tout @xmath227 .
montrons maintenant que @xmath228 .
soient @xmath220 et @xmath229 ayant exactement @xmath230 coordonnes non nulles , disons @xmath231 .
soient @xmath232 et @xmath40 la matrice extraite @xmath233 de @xmath211 de lignes @xmath234 et de colonnes @xmath235 .
le vecteur @xmath236 est le vecteur des @xmath230 premires coordonnes de @xmath237 et donc , pour toute place @xmath10 de @xmath7 au - dessus de @xmath210 , on a @xmath238 .
cette dernire quantit vaut @xmath239 o les @xmath240 sont les coordonnes de @xmath241 .
on obtient alors @xmath242 car @xmath40 est une isomtrie en la place @xmath10 ( choix de @xmath211 ) . on en dduit que @xmath243 en utilisant le lemme [ convexe ] . comme de plus @xmath244 , ceci montre que @xmath228 puis @xmath245 .
la minoration @xmath246 dcoule alors du choix des @xmath247 et lingalit @xmath248 entrane @xmath249 ( avec un choix convenable de @xmath217 ) .
ltude du comportement du premier minimum dun fibr hermitien par produit tensoriel remonte aux travaux de kitaoka dans les annes soixante - dix ( * ? ? ?
* chapitre @xmath250 ) .
il a montr que @xmath251 lorsque @xmath252 sont des fibrs adliques hermitiens sur @xmath253 avec @xmath254 . _ a contrario _
steinberg a tabli lexistence , pour tout entier @xmath255 , dun fibr @xmath21 sur @xmath253 , de dimension @xmath24 , tel que @xmath256 ( voir @xcite ) .
par la suite , coulangeon a obtenu des rsultats analogues pour des minima de fibrs sur un corps imaginaire quadratique @xcite .
dans cette partie nous examinons le comportement du minimum sur @xmath257 vis - - vis du produit tensoriel .
la situation est plus tranche : nous montrons que le couple @xmath258 sur @xmath253 peut tre remplac par @xmath259 .
nous donnons un bref exemple pour lequel on a multiplicativit par rapport au produit tensoriel .
ce serait dj le cas pour le fibr standard mais nous proposons une variante qui nous servira galement illustrer dautres rsultats .
soient @xmath24 un entier @xmath154 et @xmath261 lhyperplan @xmath262 de @xmath263 .
le fibr adlique hermitien @xmath264 possde la proprit suivante .
pour toute extension algbrique @xmath265 , pour tout fibr adlique hermitien @xmath21 sur @xmath253 , on a @xmath266 on peut supposer que @xmath76 est un corps de nombres .
pour @xmath267 , lapplication @xmath268 est une injection ( isomtrique ) de @xmath269 dans @xmath270 et donc @xmath271 par le lemme [ convexe ] .
dans le minorant au moins deux termes ne sont pas nuls si @xmath272 et , dans ce cas , @xmath273 puis @xmath274 .
lgalit sobtient en considrant le vecteur @xmath275 avec @xmath276 .
de cette dmonstration lon dduit galement que les vecteurs @xmath77 de @xmath277 tels que @xmath278 sont _ scinds _ , de la forme @xmath279 avec @xmath280 et @xmath281 .
dans ce paragraphe nous dmontrons que le premier minimum sur @xmath257 nest en gnral pas multiplicatif relativement au produit tensoriel ( thorme [ nml ] ) .
soit @xmath282 .
pour @xmath283 , soit @xmath284 une puissance @xmath285 de @xmath286 .
soit @xmath7 une extension finie de @xmath287 contenant @xmath288 et @xmath289 .
les nombres @xmath290 et @xmath291 engendrent les deux idaux premiers de @xmath292 $ ] au - dessus de @xmath293 et la valeur absolue @xmath10-adique de @xmath286 vaut @xmath294 ou @xmath295 selon que la place @xmath10 de @xmath7 divise ou non la place @xmath296 .
considrons alors le fibr adlique hermitien @xmath297 sur @xmath7 despace sous - jacent @xmath298 , de normes @xmath299 en toutes les places @xmath10 de @xmath7 qui ne divisent pas @xmath293 et , pour @xmath283 , @xmath300en dautres termes , @xmath301 est le fibr hermitien donn par la matrice @xmath302 en une place @xmath303 ( et la matrice identit en les autres places ) .
en particulier , on a @xmath304 .
[ ce ] il existe un nombre rel @xmath305 tel que , pour tout nombre rationnel @xmath306 vrifiant @xmath307 , on a @xmath308 . en particulier , pour un tel @xmath306 , on a @xmath309 .
la dmonstration utilise le lemme suivant .
soient @xmath310 et @xmath311 avec @xmath312 et @xmath313 .
alors @xmath314 .
si @xmath315 nous avons @xmath316 donc @xmath317 o la seconde ingalit rsulte du lemme [ convexe ] .
supposons quil y ait galit @xmath318 .
alors @xmath319 avec @xmath320 racine de lunit vrifiant @xmath321 pour toute place @xmath322 et tout @xmath283 .
ainsi @xmath323 , do nt toutes les valeurs absolues ultramtriques sont plus petites que @xmath295 , est un entier algbrique .
lgalit @xmath324 montre que @xmath325 est aussi un entier algbrique .
il en est alors de mme pour @xmath326 ce qui est absurde puisque sa valeur absolue en une place au - dessus de @xmath290 est @xmath327 .
nous appliquons le thorme [ tzh ] @xmath328 : on a @xmath329 .
par dfinition de @xmath330 , les couples @xmath311 tels que @xmath331 appartiennent un nombre fini de droites .
daprs le lemme ,
il existe donc @xmath305 tel que , pour tout @xmath332 avec @xmath333 , on a @xmath334 .
choisissons maintenant @xmath310 tel que @xmath307 , ce qui implique @xmath335 .
pour tous nombres algbriques @xmath336 et toute place @xmath10 , on a @xmath337 et donc @xmath338 .
si @xmath333 ces hauteurs sont suprieures @xmath339 .
si @xmath340 et @xmath341 alors @xmath342 .
ceci montre que @xmath308 .
quant au produit tensoriel @xmath343 , il sidentifie @xmath344 muni de sa structure hermitienne usuelle sauf en les places @xmath10 au - dessus de @xmath286 , @xmath283 , pour lesquelles la norme est donne par le produit de kronecker @xmath345 : @xmath346de la sorte on a @xmath347 et la seconde assertion du thorme sensuit .
le thorme [ nml ] sobtient en choisissant pour @xmath21 et @xmath26 des sommes directes hermitiennes dun @xmath297 fix et dun fibr quelconque de minimum absolu @xmath348 . en calculant le minimum dune somme directe comme indiqu aprs le lemme [ convexe ] , nous trouvons @xmath349 et @xmath350 .
nous donnons dans ce paragraphe un calcul de pente maximale en prsence dune action de groupe .
nous disons quun _ groupe @xmath351 agit sur un fibr vectoriel adlique @xmath198 _
lorsque @xmath351 agit sur lespace vectoriel @xmath8 et que , pour tout @xmath352 et toute place @xmath10 de @xmath7 , lautomorphisme de @xmath353 donn par @xmath354 est une isomtrie pour la norme @xmath355 . de plus , laction est dite : * _ irrductible _ si les seuls sous - espaces @xmath351-stables de @xmath8 sont @xmath356 et @xmath8 ; * _ gomtriquement irrductible _
si laction de @xmath351 sur @xmath357 est irrductible .
avec cette terminologie , nous avons lnonc suivant .
[ actgr ] soient @xmath358 et @xmath359 deux fibrs vectoriels adliques hermitiens .
soit @xmath351 un groupe agissant sur @xmath358 de faon gomtriquement irrductible .
alors @xmath198 est semi - stable et @xmath360 .
nous utilisons dans la dmonstration le lemme ci - dessous de thorie des reprsentations .
soient @xmath8 et @xmath104 deux espaces vectoriels de dimension finie sur un corps @xmath7 .
soient @xmath351 un groupe agissant sur @xmath8 de manire gomtriquement irrductible et @xmath361 un sous - espace @xmath351-stable de @xmath362 .
alors il existe un sous - espace @xmath363 de @xmath104 tel que @xmath364 .
nous raisonnons par labsurde et considrons un espace @xmath104 de dimension minimale pour lequel le lemme est faux ( @xmath8 fix ) .
nous choisissons ensuite un sous - espace @xmath361 de dimension minimale qui viole lnonc .
ceci entrane dune part que le seul sous - espace @xmath365 de @xmath104 tel que @xmath366 est @xmath104 lui - mme et dautre part que @xmath361 est irrductible pour laction de @xmath351 : sil contenait un sous - espace strict non nul et @xmath351-stable celui - ci serait de la forme @xmath367 et @xmath368 fournirait un contre - exemple de dimension infrieure .
par consquent si @xmath369 alors @xmath370 induit un @xmath351-morphisme @xmath371 qui ne peut tre que nul ou bijectif .
sil est nul @xmath372 donne @xmath373 donc @xmath374 .
par suite , pour tous @xmath375 , nous obtenons un @xmath351-automorphisme @xmath376 de @xmath8 . par le lemme de schur , @xmath98 scrit @xmath377 pour @xmath378 ( cest ici quintervient lhypothse dirrductibilit sur @xmath379 : il existe @xmath380 tel que @xmath381 est non inversible donc nul donc @xmath382 puis @xmath378 ) .
nous en dduisons @xmath383
puis @xmath384 et , comme ci - dessus , cela entrane @xmath385 .
on conclut @xmath386 mais comme @xmath387 linclusion @xmath388 ne peut
pas tre stricte et cest la contradiction cherche .
lautre ingrdient de notre dmonstration est lexistence , pour un fibr adlique hermitien @xmath358 , dun unique
sous - espace @xmath389 de pente et dimension maximales ( voir ( * ? ? ?
* lemme 5.12 ) ) parfois appel sous - espace _
dstabilisant_. en particulier , si un groupe @xmath351 agit sur @xmath358 , ceci montre immdiatement que @xmath389 est @xmath351-stable ( si @xmath352 le sous - espace @xmath390 a mme pente et mme dimension que @xmath389 ) . on en dduit donc ( voir ( * ? ? ?
* proposition 5.17 ) ) que si laction est irrductible alors @xmath358 est semi - stable ( car @xmath391 ) .
ceci permet galement de voir que la pente maximale est invariante par extension de corps _ via _ laction dun groupe de galois .
soit @xmath361 lunique sous - espace de @xmath362 avec @xmath392 et de dimension maximale parmi les sous - espaces de mme pente .
par unicit @xmath361 est @xmath351-stable donc par le lemme il est de la forme @xmath393 .
par suite nous avons @xmath394 et le rsultat en dcoule car lingalit contraire est toujours vraie .
comme application directe , nous voyons que @xmath260 est semi - stable et que , pour tout fibr adlique hermitien @xmath21 sur un corps de nombres , on a @xmath395 en effet le groupe symtrique @xmath396 sur @xmath397 lments agit sur @xmath260 de manire gomtriquement irrductible ( en permutant les coordonnes ) . [ [ section ] ] le but de ce paragraphe est de dmontrer le thorme [ prin ] et ses consquences directes .
soit @xmath76 une extension finie de @xmath7 . on choisit
@xmath398 tel que i ) @xmath399 et ii ) lentier @xmath400 est minimal pour cette proprit .
la minimalit de @xmath400 implique que les espaces vectoriels @xmath401 et @xmath402 sont de dimension @xmath400 .
montrons que du livre de kitaoka @xcite . ]
@xmath403 on le fait place par place . on choisit des bases orthonormes de @xmath404 et @xmath405 en une place @xmath10 et lon note @xmath406 ( _ resp_. @xmath407 ) la matrice de @xmath408 ( _ resp_. @xmath409 ) dans les bases orthonormes choisies . alors @xmath410 vaut @xmath411 , _
i.e. _ la racine carre de la somme des carrs des valeurs absolues des coefficients de @xmath412 ( norme de hilbert - schmidt ) si @xmath10 est archimdienne et le maximum des valeurs absolues des coefficients de @xmath412 si @xmath10 est ultramtrique . or , comme consquence de lingalit arithmtico - gomtrique dans le cas archimdien et de la simple ingalit ultramtrique dans le cas ultramtrique , on a @xmath413 on obtient alors par produit car la matrice adlique @xmath414 dfinit la structure hermitienne de @xmath404 ( _ idem _ pour les @xmath407 et @xmath405 ) .
ceci tant tabli , on applique le thorme [ tzhi ] @xmath404 : @xmath415 .
en majorant @xmath416 par @xmath417 et @xmath418 par @xmath419 , on obtient @xmath420 pour toute extension @xmath421 .
ceci permet de conclure .
nous venons de dmontrer le thorme [ prin ] laide du thorme [ tzhi ] .
inversement , ce dernier est contenu dans le thorme [ prin ] comme le montre largument suivant ( inspir de celui du thorme 4.6 de @xcite ) : soit @xmath389 le sous - espace dstabilisant de @xmath8 ; on a @xmath422 en considrant lidentit ; le thorme [ prin ] donne @xmath423 car @xmath424 par semi - stabilit de @xmath425 et de son dual . on obtient ainsi @xmath426 .
nous dduisons de cet nonc une version lgrement plus prcise du corollaire [ cor ] .
[ theoprecis]soient @xmath427 un entier naturel et @xmath428 des fibrs adliques hermitiens , de dimensions respectives @xmath429 .
alors on a @xmath430 et @xmath431 nous pouvons galement citer la consquence suivante du thorme [ prin ] .
[ corominima ] pour tous fibrs adliques hermitiens @xmath30 , on a @xmath432
dans cette partie , nous tablissons le rsultat suivant .
[ pentesym ] soient @xmath433 un entier et @xmath21 un fibr adlique hermitien sur @xmath7 , de dimension @xmath139 .
soit @xmath434 le ppcm des coefficients multinomiaux @xmath435 ( voir lappendice ) . on a @xmath436et @xmath437
la dmonstration du thorme [ pentesym ] repose sur le rsultat suivant .
[ lemmaquetredeux ] avec les donnes du thorme [ pentesym ] , on a @xmath438 la projection canonique @xmath439 fournit une injection isomtrique @xmath440 soit @xmath441 lapplication linaire dfinie de la manire suivante : soient @xmath442 ( @xmath443 ) et @xmath444 ( @xmath445 )
. posons @xmath446
( @xmath447 est le groupe symtrique de @xmath448 ) .
cette formule a un sens car , par symtrie , elle ne dpend pas du choix de lordre dans lequel sont les @xmath449 .
la dfinition de @xmath450 se prolonge naturellement des lments quelconques de @xmath128 et @xmath451 par bilinarit .
lapplication @xmath450 est un isomorphisme despaces vectoriels et lingalit applique @xmath452 donne @xmath453 montrons que @xmath454 .
soit @xmath455 le prolongement de @xmath450 @xmath456 . on a une criture alternative de @xmath455 , valable pour toute base @xmath66 de @xmath65 , de base duale @xmath457 , de la forme ( les notations sont celles du [ puissancesym ] ) : @xmath458(@xmath459 ) .
pour utiliser cette formule , nous choisissons une base orthonorme @xmath66 de @xmath73 et la base duale est automatiquement orthonorme .
lapplication @xmath455 est une isomtrie si @xmath10 est archimdienne .
en effet , en une telle place @xmath10 , en crivant @xmath460 et en utilisant lingalit de cauchy - schwarz , on a@xmath461 le choix @xmath462 donne lgalit et @xmath455 est une isomtrie en @xmath10 .
si @xmath10 est ultramtrique , les bases @xmath463 et @xmath464 sont orthonormes et par ingalit ultramtrique , on a @xmath465 ( lgalit est satisfaite comme on le voit en considrant @xmath466 , @xmath166 ralisant le maximum ) . il ne reste plus qu observer que le dernier maximum qui apparat vaut @xmath467 et que le produit de ces nombres ( renormaliss avec les degrs locaux ) sur les places ultramtriques @xmath10 de @xmath7 vaut @xmath434 .
la positivit de la diffrence des pentes maximales est bien connue ( voir ( * ? ? ?
* proposition @xmath468 ) ) .
pour la majoration de cette diffrence , le corollaire [ corominima ] minore @xmath469 par @xmath470 et le thorme de zhang [ tzhi ] donne @xmath471 on conclut avec le lemme [ lemmaquetredeux ] . en ce qui concerne le minimum de @xmath131 , la majoration par @xmath472 sobtient simplement en considrant @xmath473 pour @xmath474 . on a @xmath475et lon fait ensuite tendre @xmath476 vers @xmath418 .
pour la minoration , on utilise encore le lemme [ lemmaquetredeux ] et lon minore @xmath469 avec le corollaire [ theoprecis ] .
afin de juger de la finesse de cette estimation , voici un calcul de la pente maximale dune puissance symtrique lorsque @xmath477 .
[ propquatredeux]soient @xmath478 deux entiers .
posons @xmath479 $ ] .
pour toute extension algbrique @xmath76 de @xmath253 , on a @xmath480 soient @xmath72 les vecteurs de la base canonique de @xmath481 .
alors @xmath482 est la somme directe hermitienne des droites @xmath483 pour @xmath135 de longueur @xmath484 .
ainsi on a @xmath485(voir @xcite ) .
par ailleurs @xmath486 est le minimum des hauteurs des @xmath487 , qui vaut @xmath488 ( voir [ ich ] ) .
ceci tablit la premire galit de la proposition [ propquatredeux ] .
quant la seconde , considrons le multiplet @xmath489 qui ralise le maximum dans .
pour tous entiers @xmath490 , on a @xmath491 comme on le voit en posant @xmath492 ( lorsque @xmath493 ) et en utilisant @xmath494 .
les valeurs prises par les coordonnes de @xmath495 sont donc soit @xmath496 soit @xmath497 pour un certain entier naturel @xmath496 .
posons @xmath498 . on a @xmath499 car @xmath500 .
si @xmath501 alors @xmath502 et la proposition est dmontre .
si @xmath503 alors lgalit @xmath504 est la division euclidienne de @xmath400 par @xmath24 et donc @xmath505 .
il suffit alors dobserver que @xmath506 pour conclure .
le thorme [ pentesym ] et la proposition [ propquatredeux ] donnent un encadrement de la quantit @xmath507 o @xmath21 parcourt tous les fibrs adliques hermitiens de dimension @xmath24 . plus
prcisment , on a@xmath508 lorsque @xmath509 .
lgalit de gauche provient de la formule de stirling et celle de droite est dmontre au [ subsecconse ] o nous verrons aussi @xmath510 .
le dbut de la dmonstration de la proposition [ propquatredeux ] montre galement que la puissance symtrique @xmath482 est semi - stable si et seulement si @xmath400 ou @xmath24 gale @xmath295 . en effet ,
la pente @xmath511 est la moyenne des pentes des @xmath483 ( la pente dune somme directe hermitienne est une moyenne pondre des pentes , voir [ dsd ] ) tandis que montre que la pente maximale est le maximum de ces pentes .
par consquent , lgalit @xmath512 nest possible que si @xmath513 reste constant sur @xmath514 .
on voit alors que les seules possibilits sont @xmath515 ou @xmath205 en considrant les multiplets @xmath516 et @xmath517 ( lorsque @xmath197 ) .
soient @xmath21 un fibr adlique hermitien sur @xmath7 de dimension @xmath139 et @xmath140 . on a @xmath518et @xmath519 la minoration
de la pente maximale dcoule de la formule donnant la pente de @xmath144 en fonction de celle de @xmath21 ( [ secpuissanceexterieure ] ) .
pour la majoration , considrons lapplication @xmath520 induite par la forme multilinaire alterne @xmath521 ( @xmath522 est la signature de la permutation @xmath523 ) .
lapplication @xmath524 est injective et ses normes doprateur sont @xmath525 si @xmath526 et @xmath527 si @xmath528 .
par consquent , on a @xmath529 daprs . en utilisant le corollaire [ corominima ] et la majoration @xmath530 du thorme [ tzhi ] , nous obtenons la majoration souhaite de la pente maximale de @xmath144 .
si nous appliquons plutt le corollaire [ theoprecis ] nous obtenons la majoration de @xmath531 du thorme .
dans la minoration de la pente maximale de @xmath144 donne par ce thorme , il nest pas possible en gnral de remplacer @xmath101 par @xmath20 car si @xmath150 alors @xmath532 . revenons au rseau de racines @xmath260 introduit au [ an ] .
soient @xmath139 et @xmath140 des entiers .
pour tout fibr adlique hermitien @xmath21 sur @xmath199 , on a @xmath533 en utilisant la base canonique @xmath534 de @xmath263 , nous avons une description de la puissance extrieure @xmath535 de la manire suivante .
soit @xmath536 un vecteur de @xmath537 .
si @xmath538 sont des entiers quelconques de @xmath539 , posons @xmath540 si deux indices @xmath541 sont gaux et @xmath542 si @xmath523 est la permutation de @xmath543 , de signature @xmath522 , telle que @xmath544 .
avec cette convention , les quations qui dcrivent @xmath545 sont @xmath546 pour tout @xmath547 tel que @xmath548 .
si @xmath549 , considrons @xmath550 un coefficient non nul de @xmath77 avec @xmath551 .
pour toute partie @xmath158 de @xmath543 @xmath552 lments , les quations @xmath553 montrent quil existe @xmath554 tel que @xmath555 .
comme il y a @xmath400 possibilits pour @xmath158 et que les ensembles @xmath556 sont tous distincts , il y a au moins @xmath557 vecteurs @xmath550 qui ne sont pas nuls .
en une place @xmath10 dun corps de dfinition de @xmath77 , la norme de @xmath77 vaut @xmath558 si @xmath528 et @xmath559 sinon
. on obtient alors la minoration @xmath560 par le lemme de convexit [ convexe ] , ce qui donne @xmath561 .
lgalit sobtient en choisissant @xmath77 de la forme @xmath562 avec @xmath563 qui se rapproche de @xmath418 .
pour les pentes maximales , le groupe symtrique @xmath564 agit encore sur @xmath565 de manire gomtriquement irrductible ( voir ( * ? ? ?
* proposition @xmath566 ) ) . on applique alors la proposition [ actgr ] qui donne @xmath567 .
la pente maximale de @xmath568 se calcule au moyen du dterminant de ce fibr @xmath569 et la proposition est dmontre .
[ [ section-1 ] ] le cas @xmath573 figure dj dans un article de williams @xcite .
nous navons pas trouv de rfrence dans la littrature pour le cas gnral .
aussi donnons - nous ci - dessous une dmonstration ( lmentaire ) du thorme [ thmppcm ] .
soient @xmath582 avec @xmath583 pour @xmath584 .
dfinissons des entiers @xmath585 par le procd suivant : si @xmath586 sont construits on choisit un multiple @xmath585 de @xmath587 avec @xmath588 et @xmath589 ( ceci est possible car @xmath583 montre quau moins @xmath590 multiples de @xmath587 appartiennent @xmath591 $ ] ) .
alors les @xmath585 sont deux deux distincts et @xmath592 .
soit @xmath603 tel que @xmath604 .
par dfinition de @xmath605 et pour tout @xmath606 , on a @xmath607 ce qui , en sommant sur @xmath495 , conduit @xmath608 ceci montre la premire assertion .
pour la seconde , nous transitons par des intgrales , comme nair@xcite dans le cas @xmath573 .
considrons le simplexe @xmath609^n;\ x_1+\cdots+x_n\le1\}.\ ] ] si @xmath610 et @xmath611 nous avons @xmath612 pour le voir , effectuons le changement de variables @xmath613 ( @xmath614 ) et @xmath615 qui donne @xmath616 par intgrations par parties successives lintgrale sur @xmath230 vaut @xmath617 et lon conclut par rcurrence .
dautre part , nous avons aussi @xmath618 avec @xmath619
. fixons une famille @xmath620 et notons @xmath621 la suite ordonne des nombres @xmath622 .
nous avons @xmath623 donc @xmath624 .
ainsi @xmath625 divise @xmath626 .
maintenant le nombre @xmath627 appartient @xmath628 .
en combinant tout ceci nous avons @xmath629 qui donne le lemme .
il est bien connu que le thorme des nombres premiers entrane une estimation asymptotique de la _ fonction de tchebychev de seconde espce _
@xmath630)$ ] : @xmath631 lorsque @xmath632 .
ce rsultat conduit lestimation asymptotique @xmath633o , comme plus haut , @xmath634 dsigne le nombre harmonique .
il peut tre intressant davoir un encadrement effectif de @xmath434 .
par exemple , en minorant @xmath434 par le maximum des coefficients multinomiaux , on a une minoration simple et asymptotiquement prcise : @xmath635 ( on rappelle que la somme des multinomiaux vaut @xmath636 ) .
en sens inverse , un thorme de rosser & schoenfeld ( * ? ? ?
* thorme @xmath637 ) implique @xmath638 pour tout nombre rel @xmath639 , qui , son tour , fournit la majoration suivante .
lnonc tant trivial si @xmath24 ou @xmath400 gale @xmath295 , on peut supposer @xmath641 .
si @xmath642 on utilise la borne @xmath643 qui implique @xmath644 .
lorsque @xmath645 on utilise le thorme [ thmppcm ] et la majoration de @xmath524 ci - dessus qui permettent de majorer @xmath646 par @xmath647 .
par une tude de fonction , on a @xmath648 ds que @xmath649 .
la majoration @xmath650 reste vraie pour @xmath651 par vrification directe .
en utilisant cette borne on trouve @xmath652car @xmath653 et lon conclut en majorant @xmath654 par le produit des @xmath655 , @xmath656 . | nous tudions les relations entre la pente maximale dun fibr adlique hermitien , issue de la thorie des pentes de bost , et le minimum absolu de ce fibr .
en particulier , nous tablissons un thorme de minkowski - hlawka absolu et nous montrons que le minimum absolu nest pas multiplicatif par produit tensoriel . de plus
nous montrons comment un lemme de siegel d zhang permet de majorer la pente maximale dun produit tensoriel de fibrs adliques hermitiens et damliorer ainsi un thorme obtenu rcemment par chen .
nous en dduisons des consquences pour les puissances symtriques et extrieures de tels fibrs , do nt lune fait intervenir le ppcm des coefficients multinomiaux , que nous calculons . + abstract .
slopes of an adelic vector bundle exhibit a behaviour akin to successive minima .
comparisons between the two amount to a siegel lemma . here
we use zhang s version for absolute minima over the algebraic numbers .
we prove a minkowski - hlawka theorem in this context .
we also study the tensor product of two hermitian bundles bounding both its absolute minimum and maximal slope , thus improving an estimate of chen .
we further include similar inequalities for exterior and symmetric powers , in terms of some lcm of multinomial coefficients . * coordonnes des auteurs :* ric gaudron + universit grenoble i , institut fourier . +
umr @xmath0 , bp @xmath1 + @xmath2 saint - martin - dhres cedex , france . + courriel : ` [email protected] ` + ( 33 ) 04 76 51 45 72 + page internet : ` http://www-fourier.ujf-grenoble.fr/\simgaudron ` gal rmond + universit grenoble i , institut fourier .
+ umr @xmath0 , bp @xmath1 + @xmath2 saint - martin - dhres cedex , france . + courriel : ` [email protected] ` + ( 33 ) 04 76 51 49 86 |
undirected formation control has been one of the most studied subjects in multi - agent systems .
the formation control model is described by two characteristics : one is an undirected graph describing the pattern of interaction , and the other is a set of scalar functions each of which describes the interaction law between a pair of adjacent agents .
a detailed description is given below let @xmath0 be an undirected connected graph of @xmath1 vertices @xmath2 .
let @xmath3 be the set of vertices adjacent to vertex @xmath4 .
we then consider the motion of @xmath1 agents @xmath5 in @xmath6 by @xmath7 each @xmath8 with @xmath9 is a scalar function describing how adjacent agents @xmath10 and @xmath11 interact with each other .
we require that @xmath8 be identical with @xmath12 for all @xmath9 , in other words , interactions among agents are reciprocal . in its most general form ,
each @xmath8 could be a function of @xmath5 , and possibly the time variable @xmath13 as well . in this case , we have proved in @xcite that if @xmath14 is connected , then system , treated as a centralized control system , is approximately path - controllable over an open dense subset of the configuration space . yet ,
if we regard system as a decentralized control system , i.e , each agent @xmath10 only accesses part of the information , then there is a restriction on what variables each @xmath8 can depend on .
for example , it is often assumed that each agent @xmath10 knows agent @xmath11 if and only if @xmath11 is its neighbor , i.e , @xmath15 . then in this case
each @xmath8 is at most a function of @xmath10 and @xmath11 , and possibly the time @xmath13 . over the last decade ,
there have been many solid works about using system to achieve decentralized formation control .
questions about the level of interaction laws that are necessary for organizing such systems , questions about system convergence , questions about counting and locating stable equilibria , and questions about the issues of robustness and etc .
have all been treated to some degree ( see , for example , @xcite ) .
for example , a popular decentralized algorithm , known as the krick s law @xcite , is that we assume each agent @xmath10 measures the mutual distance between himself and its neighbors @xmath11 .
the control law is then given by @xmath16 for all @xmath9 where @xmath17 is the prescribed distance between @xmath10 and @xmath11 in the target formation . by following this decentralized algorithm , system
will then be a gradient system with respect to the potential function @xmath18 .
in fact , it has been shown that if each @xmath8 is a continuous function depending only on the distance @xmath19 , then the resulting system is always a gradient system @xcite .
however , the associated potential function @xmath20 often has multiple local minima , and in some cases , the number of local minima has an exponential growth with respect to the number of agents ( see , for example , @xcite ) .
also , we note that if each @xmath8 depends only on the mutual distance @xmath19 as is the case if we adopt the krick s law for @xmath8 , then how a configuration is embedded into the euclidean space is not relevant , only the shape of the formation matters . in other words , if a configuration is an equilibrium associated with system , then any rotation or translation of the configuration will also be an equilibrium . in any of such case
, the group of rigid motions is introduced to describe this phenomena .
two configurations will be recognized as the same target formation if they are in the same orbit with respect to the group action . in this paper , as in many earlier work on this problem , we will investigate system , treated as a decentralized formation control system , by equipping it with a set of new control laws .
what distinguishes this paper from others is that in addition to the shape of the target configuration , we also emphasize its euclidean embedding . to be more precise ,
we let @xmath21 and @xmath22 be two configurations with the same centroid , i.e , @xmath23 , and we distinguish @xmath24 and @xmath25 in the sense that these two configurations are recognized as the same target formation if and only if @xmath26 .
we impose the condition that @xmath24 and @xmath25 have the same centroid because of the fact that the centroid of a configuration in an undirected formation control system is invariant along the evolution regardless of the choice of the control laws .
the decentralized control law is then designed for each agent so that the solution of the control system may converge to the target configuration . in particular , we show that there is a quadratic lyapunov function associated with system whose unique local ( global ) minimum point is the target configuration .
but we also note ( and we will see later in the paper ) that there may exist a continuum of equilibria for system , thus a solution of system may fail to converge to the global minimum point . to fix this problem , we then modify the formation control laws by adding noise terms .
this is an application of simulated annealing to formation control systems .
simulation results then show that sample paths of the modified stochastic system approach the global minimum point .
the rest of this paper is organized as follows . in section 2
, we will first specify what information each agent knows , or in other words what variables @xmath8 can depend on .
then we will introduce the decentralized formation control law , and establish the convergence of system . in section 3
, we will explore one of the limitations of this formation control model by showing that there may exist a continuum of equilibria of system .
for simplicity , we will only focus on trees as a special type of network topologies . in this special case ,
we show that there is a simple condition for determining whether a configuration @xmath24 is an equilibrium or not , and thus there is a geometric characterization of the set of equilibria of system .
the existence of continuum of equilibria poses a problem about the convergence of system to the target configuration . in section 4
, we will focus on fixing this problem by applying the technique of simulated annealing to the algorithm .
simulation results then show that a typical sample path will converge to the target configuration .
in this section , we will introduce the decentralized formation control law and show that system , when equipped with this control law , converges to the set of equilibria .
but before that , we need to be clear about what we mean by a decentralized formation control system .
so in the first part of this section , we will specify what information each agent knows , i.e , what variables each @xmath8 can depend on .
also we will specify how the information of the target configuration is distributed among agents .
we first introduce the underlying space of system . as interactions among agents are reciprocal , the centroid of a configuration is always invariant along the evolution in an undirected formation control system , so we may as well assume that the centroid of a configuration is located at the origin .
the configuration space @xmath27 , as the underlying space of system , is then defined by @xmath28 it is clear that @xmath27 is a euclidean space of dimension @xmath29 . in this paper
, we assume that @xmath30 is the target configuration , i.e , each @xmath31 is the target position for agent @xmath10 .
we will now specify what information each agent @xmath10 can access . in this paper ,
if @xmath32 is an edge of @xmath14 , we then assume that * agent @xmath10 knows @xmath33 ; * agent @xmath10 is able to measure @xmath34 at any time @xmath13 ; consequently , if @xmath32 is an edge of @xmath14 , then we require that each scalar function @xmath8 depend only on @xmath34 , @xmath33 and possibly the time variable @xmath13 .
suppose @xmath32 is an edge of @xmath14 , we then let the control law @xmath8 be defined as @xmath35 where @xmath36 is the standard inner - product of two vectors , and @xmath37 is the standard euclidean norm of a vector .
we note that in this definition , if we exchange roles of vertex @xmath4 and vertex @xmath38 , then @xmath12 will be identical with @xmath8 .
this is consistent with our assumption that the formation control system is undirected .
the main result of this section we will prove is about the convergence of system as stated below .
let @xmath8 be the decentralized control law defined by expression .
then there is a quadratic lyapunov function associated with system defined as @xmath39 let @xmath40 be a solution of system , then @xmath41 the derivative is zero if and only if @xmath42 is an equilibrium . this proof is done by explicit computation .
we check that @xmath43 it is then clear that the time derivative is zero if and only if each @xmath44 is zero which implies that @xmath24 is an equilibrium . .
there may exist multiple equilibria of system .
in fact , as we will see in the next section that in the case @xmath14 is a tree graph there exists a continuum of equilibria . nevertheless , there is only one local ( and also global ) minimum point of the potential function @xmath20 which is @xmath45 .
it thus suggests that we apply simulated annealing to this formation control law as we will discuss in the last section of the paper .
+ _ remark ii_. notice that the potential function @xmath46 approaches infinity as @xmath47 goes to infinity . on the other hand ,
we have @xmath48 so each solution @xmath42 of system has to remain in a bounded set , and thus converges to the set of equilibria . in other words
, no agent escapes to infinity along the evolution .
in this section , we will explore the set of equilibria of system .
the main purpose of doing this is to illustrate one of the limitations of this formation control law .
it is well - known that if @xmath20 is a lyapunov function for a dynamical system @xmath49 and @xmath50 is a unique equilibrium , then @xmath50 is stable and all solutions of the system will converge to @xmath50 .
however , this is not the case here , i.e , the target configuration @xmath45 will not be the unique equilibrium of system . as we will see in this section there may exist a continuum of equilibria of system . for simplicity
, we will only focus on the case where the interaction pattern @xmath14 is a tree graph .
we focus on this special class of interaction patterns because in this case there is a simple condition telling us whether a configuration @xmath24 is an equilibrium or not .
in particular , we will use this condition to characterize the set of equilibria of system in a geometric way .
a path in a graph @xmath14 is a finite sequence of edges which connects a sequence of vertices .
a simple path then refers to a path which does not have repeated vertices , and a circle refers to a path without repeated vertices or edges , other than the repetition of the starting and ending vertices .
an undirected graph @xmath14 is a * tree * if any two vertices of @xmath14 are connected by a unique simple path , i.e , there is no circle in @xmath14 .
each tree graph can be inductively built up starting with one vertex , and then at each step , we join a new vertex via one new edge to an existing vertex .
this , in particular , implies that each tree graph has a leaf , i.e , a vertex of degree one .
an example of a tree graph is given in figure [ tree ] .
vertices and @xmath51 edges .
vertices are labeled with respect to an inductive construction .
for any two vertices in the graph , there is a unique path connecting them .
the four vertices @xmath52 , @xmath53 , @xmath51 and @xmath54 are leaves of the tree graph above . ] in this part , we show that if the graph @xmath14 is a tree graph , then the set of equilibria associated with system can be characterized by a simple condition stated below .
[ equic ] let @xmath55 be a tree graph , then @xmath24 is an equilibrium associated with system if and only if @xmath56 for all @xmath9 .
it is clear that if each @xmath8 is zero , then @xmath24 is an equilibrium .
we now prove the other direction , and the proof is done by induction on the number of agents .
+ _ base case_. suppose @xmath57 , then @xmath58 .
so if @xmath59 is an equilibrium , then either @xmath60 or @xmath61 , but they both imply @xmath62 .
+ _ inductive step_. assume the lemma holds for @xmath63 with @xmath64 , and we prove for the case @xmath65 .
since each tree graph has at least one leaf , we may assume that vertex @xmath66 is a leaf of @xmath14 and it joins the graph via edge @xmath67 .
suppose @xmath68 is an equilibrium , then we must have @xmath69 then by the same arguments we used for proving @xmath70 in the base case , we conclude that @xmath71 .
now let @xmath72 and let @xmath73 be the subgraph induced by @xmath74 , i.e , for any two vertices @xmath4 and @xmath38 in @xmath74 , the pair @xmath32 is an edge of @xmath75 if and only if it is an edge of @xmath14 .
it is clear that @xmath75 is also a tree graph .
let @xmath25 be a sub - configuration of @xmath24 consisting of agents @xmath76 , then @xmath25 is also an equilibrium under @xmath8 with @xmath77 .
this holds because @xmath24 is an equilibrium and meanwhile @xmath78 , so the agent @xmath79 does nt attract or repel any agent in @xmath25 . by induction
, we have @xmath80 for all @xmath77 .
this then , combined with the condition @xmath81 , establishes the proof . in this part
, we will use the equilibrium condition to characterize the set of equilibria of system .
[ conte ] let @xmath14 be a tree graph .
let @xmath82 be the set of equilibria of system with @xmath8 defined by expression .
then there is a diffeomorphism of @xmath82 given by @xmath83 as a product of @xmath84 copies of the unit sphere @xmath85 in @xmath86 .
the proof of the theorem will again carried out by induction on the number of agents .
so we first prove for the case @xmath57 , and the inductive step will be given after that .
+ _ base case_. we show that theorem [ conte ] holds in the case @xmath57 .
suppose @xmath59 is an equilibrium , then by lemma [ equic ] , we have @xmath62 , this then implies @xmath87 the set of equilibria associated with system is characterized by equation , together with the condition that @xmath88 .
let @xmath89 be a subset of @xmath6 defined by @xmath90 then it is clear that @xmath82 is diffeomorphic to @xmath89 . to see this
, we define a map @xmath91 by @xmath92 it is clear that the map @xmath93 is a diffeomorphism .
we will now show that the set @xmath89 is itself a sphere in @xmath6 .
let @xmath94 and let @xmath95 be the sphere of radius @xmath96 centered at @xmath97 in @xmath86 , i.e , @xmath98 it is clear by computation that @xmath99 in fact , if @xmath100 lies inside @xmath101 , then @xmath102 and if @xmath103 lies outside @xmath101 , then @xmath104 this then completes the proof of the base case .
+ _ inductive step_. we will now use induction to prove theorem [ conte ] .
we assume that the theorem holds for @xmath63 with @xmath64 , and we prove for the case @xmath65 .
we again assume that vertex @xmath66 is a leaf of @xmath14 and it joins the graph via edge @xmath105 .
let @xmath73 be the subgraph of @xmath14 induced by vertices @xmath106 , then @xmath75 is a tree graph .
let @xmath25 be a sub - configuration of @xmath24 consisting of agents @xmath76 .
let @xmath107 be a subset of @xmath108 defined by @xmath109 the equilibria set @xmath82 is then characterized by the condition that @xmath25 is an equilibrium , together with the condition that @xmath81 .
since these two conditions are independent of each other , there is a diffeomorphism of @xmath82 given by @xmath110 we may translate each @xmath111 in @xmath6 so that the centroid of @xmath25 is zero after translation . since @xmath75 is a tree graph , by induction the set @xmath107 is diffeomorphic to @xmath112 , and hence @xmath82 is diffeomorphic to @xmath113 .
one may ponder at this point whether the existence of continuum of equilibria is a consequence of the fact that a tree graph is not a rigid graph .
however , it is not the case . for example , if we consider three agents evolving on a plane with @xmath14 being the complete graph , one can then show that the set of equilibria is diffeomorphic to a disjoint union of two circles .
though at this moment we do not have a statement about the set of equilibria in the most general case , the tree - graph cases , as well as the three - agents example suggest that it may be inevitable for system to possess a continuum of equilibria .
in the previous section , we have showed that there may exist a continuum of equilibria of system under the proposed formation control law .
this certainly affects the efficacy of the algorithm because there may exist a solution of system which converges to an equilibrium other than the target configuration . in this section
, we will focus on fixing the problem . in view of the fact that there is only one local ( global ) minimum of the quadratic lyapunov function @xmath20 which is the target configuration , we attempt to apply simulated annealing , as a heuristic method , to the decentralized formation control algorithm .
in particular , if we add an appropriate noise term to each @xmath8 , the resulting stochastic system is then described by @xmath114 where the @xmath115 are independent standard wiener processes , and @xmath116 is a scalar function of time @xmath13 and @xmath117 defined by @xmath118 with @xmath119 and @xmath120 positive constants .
as @xmath121 decays along time , the impact of the noise tends to zero as @xmath13 goes to infinity . with these noise terms
, we expect that the centroid of the stochastic formation control system is still invariant along the evolution because otherwise the entire configuration may drift to some place which is neither predictable nor controllable .
fortunately this is the case here as stated in the next theorem .
the centroid of the configuration is invariant along the evolution of the stochastic formation control system described by expression .
let @xmath122 be a function defined by @xmath123 let @xmath124 be the @xmath66-th coordinate of agent @xmath10 .
we need to show that @xmath125 for all @xmath126 .
let @xmath127 be a vector collecting the @xmath66-th coordinates of all the agents , i.e , @xmath128 by defining vector @xmath129 , we can rewrite the system equation in a matrix form as @xmath130 where @xmath131 is a symmetric matrix of zero - column / row - sum with the @xmath132-th , @xmath133 , entry defined by @xmath134 and each @xmath135 is also a symmetric matrix of zero - column / row - sum defined by @xmath136 where @xmath137 is the standard basis of @xmath138 .
we now apply the it rule , and get the stochastic differential equation for @xmath139 as follows @xmath140 notice that for any @xmath141 , we have @xmath142 where @xmath143 is a vector of all ones , and @xmath144 so then all inner - products in equation vanish .
this completes the proof .
we now give some examples of this stochastic formation control system , and illustrate how sample paths of this stochastic system evolve over time @xmath13 .
+ * examples*. consider five agents @xmath145 , @xmath146 , @xmath147 , @xmath148 and @xmath149 evolving in @xmath150 .
let @xmath151 be the target configuration given by @xmath152 we will work with two network topologies , one is a star graph which is a special type of tree graph and the other is a circle .
details are described below + _ 1 .
star as the network topology_. we assume that @xmath55 is a star graph with @xmath153 defined by @xmath154 we then pick an initial condition @xmath155 given by @xmath156 in figure [ spstar ] , we show how the value of the quadratic lyapunov function @xmath157 evolves over time .
the smooth curve ( the green one ) refers to the solution of system where there is no noise term added into the control law . as this solution converges to an equilibrium ,
so we see from the figure that the green curve converges to a constant line along the evolution .
also it is clear that the solution , with the initial condition given by equation , does not converge to the target configuration . on the other hand
, the ragged curve refers to the solution of system where we have added noise terms into it . in the simulation ,
we have chosen @xmath158 and @xmath159 .
we see from the figure that @xmath160 approaches zero , in a stochastic way , along time @xmath13 which implies that @xmath42 approaches @xmath45
. evolves over time @xmath13 with / without noise term under the condition that @xmath14 is a star graph , with the initial condition given by expression . ]
+ _ 2 . circle as the network topology_. we assume that @xmath161 is now a circle with @xmath153 defined by @xmath162 we adopt the same initial condition given by expression . figure [ spcircle ] shows how @xmath160 evolves over time @xmath13 .
similarly , we see that if there is no noise term , then the solution of system does not converge to the target configuration . on the other hand , the sample path @xmath42 approaches @xmath45 along the evolution .
the two parameters @xmath119 and @xmath120 are again chosen to be @xmath163 and @xmath164 respectively .
evolves over time @xmath13 with / without noise term under the condition that @xmath14 is a circle , with the initial condition given by expression . ]
these two examples have demonstrated that simulated annealing can be used to modify the formation control law in order to achieve global convergence to the target configuration .
more provable facts are needed at this moment for this heuristic algorithm .
in this paper , we have proposed a decentralized formation control law for agents to converge to a target configuration in the physical space . in particular
, there is a quadratic lyapunov function associated with the formation control system whose unique local ( also global ) minimum point is the target configuration .
we then focussed on one of the limitations of this formation control model , i.e , there may exist a continuum of equilibria of the system , and thus there are solutions of system which do not converge to the target configuration . to fix this problem
, we then applied the technique of simulated annealing to the formation control law , and showed that the modified stochastic system preserves one of the basic properties of the undirected formation control system , i.e , the centroid of the configuration is invariant along the evolution over time .
we then worked on two simple examples of the stochastic system .
simulations results showed that sample paths approach to the target configuration .
the author here thanks prof .
ali belabbas , prof .
tamer baar , as well as the reviewers of the earlier draft of this work for their comments on this paper .
helmke , u. , & anderson , b. d. ( 2013 , october ) .
equivariant morse theory and formation control . " in communication , control , and computing ( allerton ) , 2013 51st annual allerton conference on ( pp .
1576 - 1583 ) .
ieee .
sun , z. , mou , s. , anderson , b. d. , & morse , a. s. ( 2014 ) . formation movements in minimally rigid formation control with mismatched mutual distances . " in the 53rd ieee conference on decision and control ( cdc 2014 ) . | in this paper , we investigate a decentralized formation control algorithm for an undirected formation control model . unlike other formation control problems where only the shape of a configuration counts ,
we emphasize here also its euclidean embedding . by following this decentralized formation control law
, the agents will converge to certain equilibrium of the control system .
in particular , we show that there is a quadratic lyapunov function associated with the formation control system whose unique local ( global ) minimum point is the target configuration . in view of the fact that there exist multiple equilibria ( in fact , a continuum of equilibria ) of the formation control system , and hence there are solutions of the system which converge to some equilibria other than the target configuration , we apply simulated annealing , as a heuristic method , to the formation control law to fix this problem .
simulation results show that sample paths of the modified stochastic system approach the target configuration . |
the gauge hierarchy problem is one of the most important issues in the standard model ( sm ) , which has been motivating us to seek new physics beyond the sm for decades .
the problem lies in the fact that quantum corrections to the self - energy of the sm higgs doublet quadratically diverges , and this divergence ( cut off by some new physics scale ) should be canceled by a fine - tuning of the higgs mass parameter when the cutoff scale is much higher than the electroweak scale , such as the planck mass .
because of the chiral nature of the sm , the sm lagrangian possesses the conformal ( scale ) invariance at the classical level , except for the higgs mass term .
it has been argued in @xcite that once the classical conformal invariance and its minimal violation by quantum anomalies are imposed on the sm , it can be free from the quadratic divergences and hence the gauge hierarchy problem .
this picture fits a setup first investigated by coleman and weinberg @xcite , namely , a u(1 ) gauge theory with a massless higgs field , where the classical conformal invariance is broken by quantum corrections in the coleman - weinberg effective potential , and the u(1 ) gauge symmetry is radiatively broken ( coleman - weinberg mechanism ) .
although it is tempting to apply this coleman - weinberg mechanism to the sm higgs sector , this can not work with the observed values of top quark and weak boson masses , since the coleman - weinberg potential for the sm higgs field is found to be unbounded from below @xcite .
therefore , in order to pursue this scheme , it is necessary to extend the sm . among several new physics model proposals ( see , for example , @xcite ) ,
classically conformal b - l extended sm proposed in @xcite is a very simple and well - motivated model .
the b - l ( baryon number minus lepton number ) is a unique anomaly - free global symmetry in the sm , and it can be easily gauged .
associated with gauging the b - l symmetry , three generation of right - handed neutrinos and a b - l higgs field are introduced to make the model free from all gauge and gravitational anomalies , and to break the b - l gauge symmetry . once the b - l gauge symmetry is broken , the b - l gauge field ( @xmath1 boson ) and the right - handed ( majorana ) neutrinos obtain their masses . with the majorana heavy neutrinos , the seesaw mechanism @xcite is automatically implemented . in @xcite , under a requirement of the classically conformal invariance
, the radiative b - l symmetry breaking by the coleman - weinberg mechanism has been investigated .
the b - l gauge symmetry breaking also triggers the electroweak symmetry breaking by generating a negative mass squared for the sm higgs doublet .
naturalness of the model requires the b - l symmetry breaking scale at or below the tev scale @xcite , so that the lhc run-2 can test the model . in this work ,
we first consider a generalization of the classically conformal b - l model . with the same new particles ( three generations of right - handed neutrinos and a new higgs field ) ,
the most general gauged u(1 ) extension of the sm is to introduce a u(1)@xmath0 gauge group , which is defined as a linear combination of the sm @xmath8 and the u(1 ) b - l gauge groups .
thus , we investigate the classically conformal u(1)@xmath0 extended sm .
u(1)@xmath0 charge assignments for particles which are different from those in the b - l model yield a drastic change in phenomenological consequences . in this context
, we consider the sm higgs vacuum stability .
the sm higgs boson has been discovered at the lhc , and this marks the beginning of the experimental confirmation of the sm higgs sector
. the observed higgs boson mass of around 125 gev @xcite ( see also the recent update from a combined analysis by the atlas and the cms @xcite ) indicates that the electroweak vacuum is unstable @xcite , since the sm higgs quartic coupling becomes negative far below the planck mass , for the top quark pole mass @xmath9 from the combined measurements by the tevatron and the lhc experiments @xcite .
this is a serious problem in our framework , since the instability of the higgs potential induces a large tree - level mass for the u(1)@xmath0 higgs field through its interaction term with the sm higgs doublet field , and spoils the coleman - weinberg mechanism for the u(1)@xmath0 sector .
in addition to the proposal of the classically conformal u(1)@xmath0 extend sm , the main purpose of this work is to resolve the higgs vacuum instability in this context .
particle contents . in addition to the sm particle contents
, the right - handed neutrino @xmath11 ( @xmath12 denotes the generation index ) and a complex scalar @xmath13 are introduced . [ cols="^,^,^,^,>,^,>",options="header " , ] the model we will investigate is the anomaly - free u(1)@xmath0 extension of the sm with the classically conformal invariance , which is based on the gauge group su(3)@xmath14su(2)@xmath15u(1)@xmath16u(1)@xmath0 .
the particle contents of the model are listed in table [ tab : particle_contents ] .
the covariant derivative relevant to u(1)@xmath16 u(1)@xmath0 is given by @xmath17 where @xmath18 ( @xmath19 ) are u(1)@xmath20 ( u(1)@xmath0 ) charge of a particle , and the gauge coupling @xmath21 is introduced associated with a kinetic mixing between the two u(1 ) gauge bosons .
the particle contents include three generations of right - hand neutrinos @xmath11 and a u(1)@xmath0 higgs field @xmath13 , in addition to the sm particle contents . for generation - independent charge assignments ,
the u(1)@xmath0 charges of the fermions are defined to satisfy the gauge and gravitational anomaly - free conditions : @xmath22 ^ 2&\ : & 2x_q - x_u - x_d & \ = \ 0 , \nonumber \\ { \rm u}(1)^\prime \times \left [ { \rm su}(2)_l \right]^2&\ : & 3x_q + x_\ell & \ = \ 0 , \nonumber \\ { \rm u}(1)^\prime \times \left [ { \rm u}(1)_y \right]^2&\ : & x_q - 8x_u - 2x_d + 3x_\ell - 6x_e & \ = \ 0 , \nonumber \\ \left [ { \rm u}(1)^\prime \right]^2 \times { \rm u}(1)_y&\ : & x_q^2 - 2x_u^2 + x_d^2 - x_\ell^2 + x_e^2 & \ = \ 0 , \nonumber \\ \left [ { \rm u}(1)^\prime \right]^3&\ : & 6x_q^3 - 3x_u^3 - 3x_d^3 + 2x_\ell^3
- x_\nu^3 - x_e^3 & \ = \ 0 , \nonumber \\ { \rm u}(1)^\prime \times \left [ { \rm grav . }
\right]^2&\ : & 6x_q - 3x_u - 3x_d + 2x_\ell
- x_\nu - x_e & \ = \ 0 .
\label{eq : anomaly_free_cond}\end{aligned}\ ] ] in order to reproduce observed fermion masses and flavor mixings , we introduce the following yukawa interactions : @xmath23 where @xmath24 and the third and fifth terms in the right - handed side are for the seesaw mechanism to generate neutrino masses .
these yukawa interaction terms impose @xmath25 solutions to these conditions are listed in table [ tab : particle_contents ] , which are controlled by only two parameters , @xmath26 and @xmath27 .
the two parameters reflect the fact that the u(1)@xmath0 gauge group can be defined as a linear combination of the sm u(1)@xmath20 and the u(1 ) b - l gauge groups . since the u(1)@xmath0 gauge coupling @xmath28 is a free parameter of the model and it always appears as a product @xmath29 , we fix @xmath30 without loss of generality throughout this paper .
this convention excludes the case that u(1)@xmath0 gauge group is identical with the sm u(1)@xmath20 .
the choice of @xmath31 corresponds to the u(1)@xmath32 model .
another example is @xmath33 , which corresponds to the sm with the so - called u(1)@xmath34 symmetry .
when we choose @xmath35 , the beta function of @xmath21 at the 1-loop level becomes proportional to @xmath21 ( see appendix a ) .
this is the orthogonal condition for the u(1)@xmath20 and u(1)@xmath0 , under which @xmath21 dose not evolve once we have set @xmath36 at an energy scale .
imposing the classically conformal invariance , the scalar potential is given by @xmath37 where the mass terms are forbidden by the conformal invariance . clearly ,
if @xmath38 is negligibly small , we can analyze the higgs potential separately for @xmath13 and @xmath39 .
this will be justified in the following sections . when the majorana yukawa couplings @xmath40 is negligible compared to the u(1)@xmath0 gauge coupling , the @xmath13 sector is identical with the original coleman - weinberg model @xcite , so that the radiative u(1)@xmath0 symmetry breaking will be achieved .
once @xmath13 develops a vev through the coleman - weinberg mechanism , the tree - level mass term for the sm higgs is effectively generated through @xmath38 in eq .
( [ eq : classical_potential ] ) . taking @xmath38 negative , the induced mass squared for the higgs doublet is negative and , as a result , the electroweak symmetry breaking is driven in the same way as in the sm .
assuming @xmath38 is negligibly small , we first analyze the u(1)@xmath0 higgs sector . without mass terms , the coleman - weinbeg potential @xcite at the 1-loop level is found to be @xmath41 - \frac{25}{6 } \right ) , \label{eq : cw_potential } \end{aligned}\ ] ] where @xmath42 $ ] , and we have chosen the renormalization scale to be the vev of @xmath13 ( @xmath43 ) . here , the coefficient of the 1-loop quantum corrections is given by @xmath44 \\ & \simeq & \frac{1}{16 \pi^2 } \left [ 6 \left(x_\phi g^\prime \right)^4 - 16\sum_i(y_m^i)^4 \right ] , \end{aligned}\ ] ] where in the last expression , we have used @xmath45 as usual in the coleman - weinberg mechanism and set @xmath46 at @xmath47 , for simplicity .
the stationary condition @xmath48 leads to @xmath49 and this @xmath50 is nothing but a renormalized self coupling at @xmath51 defined as @xmath52 for more detailed discussion , see @xcite .
associated with this radiative u(1)@xmath0 symmetry breaking ( as well as the electroweak symmetry breaking ) , the u(1)@xmath0 gauge boson ( @xmath1 boson ) and the right - handed majorana neutrinos acquire their masses as @xmath53 where @xmath54 gev is the sm higgs vev , and we have used @xmath55 , which will be verified below . in this paper , we assume degenerate masses for the three majorana neutrinos , @xmath56 ( equivalently , @xmath57 ) for all @xmath12 , for simplicity .
the u(1)@xmath0 higgs boson mass is given by @xmath58 when the yukawa coupling is negligibly small , this reduces to the well - known relation derived in the radiative symmetry breaking by the coleman - weinberg mechanism @xcite . for a sizable majorana mass
, this formula indicates that the potential minimum disappears for @xmath59 , so that there is an upper bound on the right - handed neutrino mass for the u(1)@xmath0 symmetry to be broken radiatively .
this is in fact the same reason as why the coleman - weinberg mechanism in the sm higgs sector fails to break the electroweak symmetry when the top yukawa coupling is large as observed . in order to avoid the destabilization of the u(1)@xmath0 higgs potential
, we simply set @xmath60 in the following analysis .
note that this condition does not mean that the majorana neutrinos must be very light , even though a factor difference between @xmath61 and @xmath62 is enough to satisfy the condition .
let us now consider the sm higgs sector . in our model ,
the electroweak symmetry breaking is achieved in a very simple way .
once the u(1)@xmath0 symmetry is radiatively broken , the sm higgs doublet mass is generated through the mixing term between @xmath39 and @xmath13 in the scalar potential ( see eq .
( [ eq : classical_potential ] ) ) , @xmath63 where @xmath64 . choosing @xmath65 , the electroweak symmetry is broken in the same way as in the sm @xcite
however , the crucial difference from the sm is that in our model the electroweak symmetry breaking originates from the radiative breaking of the u(1)@xmath0 gauge symmetry . at the tree level , the stationary condition @xmath66 leads to the relation @xmath67 , and the higgs boson mass @xmath68 is given by @xmath69 in the following rge analysis , this is used as the boundary condition for @xmath38 at the normalization scale @xmath70 .
note that since @xmath71 and @xmath72 tev by the lep constraint @xcite , @xmath73 , which is very small . in our discussion about the u(1)@xmath0 symmetry breaking
, we neglected @xmath38 by assuming it to be negligibly small . here
we justify this treatment . in the presence of @xmath38 and the higgs vev , eq .
( [ eq : stationary ] ) is modified as @xmath74 considering the current lhc bound form search for @xmath1 boson resonances @xcite , @xmath75 tev , we find that the first term in the parenthesis in the last equality is 5 orders of magnitude greater than the second term , and therefore we can analyze the two higgs sectors separately .
( solid line ) for the inputs @xmath76 gev and @xmath77 gev , along with the sm case ( dashed line ) .
( b ) the rg evolutions of @xmath50 ( solid line ) and @xmath38 ( dashed line ) . here , we have taken @xmath78 , @xmath79 tev and @xmath80 .
, title="fig : " ] ( a ) ( solid line ) for the inputs @xmath76 gev and @xmath77 gev , along with the sm case ( dashed line ) .
( b ) the rg evolutions of @xmath50 ( solid line ) and @xmath38 ( dashed line ) . here , we have taken @xmath78 , @xmath79 tev and @xmath80 .
, title="fig : " ] ( b ) in our classically conformal u(1)@xmath0 extended sm , the u(1)@xmath0 gauge symmetry is radiatively broken by the coleman - weinberg mechanism .
associated with this symmetry breaking , the negative higgs mass squared is generated to break the electroweak symmetry as in the sm . in the sm with the observed higgs boson mass around 125 gev , the rge evolution of the sm higgs quartic
coupling shows that the coupling becomes negative at the intermediate scale @xmath81 gev @xcite ( dependently of input masses for the higgs boson and top quark ) , and hence the electroweak vacuum is unstable .
this vacuum instability might not be a serious problem in the sm , since the lifetime of the electroweak vacuum is much longer than the age of the universe @xcite .
however , in our model , this sm higgs vacuum instability generates a large negative mass squared of @xmath13 through the @xmath38 term , and hence the coleman - weinberg mechanism in the u(1)@xmath0 higgs sector is spoiled . in this section ,
we investigate rg evolution of the higgs quartic coupling and a possibility to solve the higgs vacuum stability problem in our u(1)@xmath0 extended sm . without the classical conformal invariance , ref .
@xcite has considered the same problem and identified parameter regions which can remove the higgs vacuum instability .
a crucial difference in our model is that because of the classical conformal invariance and the symmetry breaking by the coleman - weinberg mechanism , the initial values of @xmath50 and @xmath38 at @xmath51 are not free parameters .
therefore , it is nontrivial to resolve the higgs vacuum instability in the present morel .
the higgs vacuum stability has been investigated in @xcite for classically conformal extension of the sm with an extend gauge groups and particle contents ( including a dark matter candidate ) . for our rge analysis
, we employ the sm rges at 2-loop level @xcite from the top pole mass to the u(1)@xmath0 higgs vev , and connect the rges to those of the u(1)@xmath0 extended sm at the 1-loop level .
all formulas used in our analysis are listed in appendices . for inputs for the higgs boson mass and top quark pole mass
, we employ a central value of the cms measurement @xmath77 gev @xcite , which is slightly smaller than the atlas measurement @xmath82 gev @xcite , while @xmath83 which is the central value of combined results of the tevatron and the lhc measurements of top quark mass @xcite .
there are only 3 free parameters in our model , by which inputs at @xmath51 are determined : @xmath26 , @xmath51 , and @xmath84 . in fig .
[ fig : higgs_quartics ] ( a ) , we show the rg evolution of the sm higgs quartic coupling in our model ( solid line ) , along with the sm case ( dashed line ) . here , we have taken @xmath85 , @xmath79 tev , and @xmath80 as an example . recall that we have fixed @xmath30 without loss of generality .
the higgs quartic coupling remains positive all the way up to the planck mass , so that the higgs vacuum instability problem is solved .
there are complex , synergetic effects in the coupled rges to resolve the higgs vacuum instability ( see appendices for rges ) .
for example , the u(1)@xmath20 gauge coupling grows faster than the sm case in the presence of the mixing gauge coupling @xmath86 , which makes the evolution of top yukawa coupling decreasing faster than in the sm case .
the evolution of the mixing gauge coupling is controlled by the u(1)@xmath0 gauge coupling . both of them are asymptotic non - free .
the gauge couplings positively contribute to the beta function of the sm higgs quartic coupling , while the top yukawa coupling gives a negative contribution . as a result
, the rg evolutions of the gauge and top yukawa couplings work to change the sign of the the beta function of the sm higgs quartic coupling at @xmath87 gev in fig .
[ fig : higgs_quartics ] ( a ) . fig .
[ fig : higgs_quartics ] ( b ) shows the rg evolutions of the other higgs quartic couplings .
note that the input of @xmath50 and @xmath38 are very small because of the radiative gauge symmetry breaking , and the two couplings remain very small even after reaching the planck scale .
thus , the positive contribution of @xmath38 to the beta function of the sm higgs quartic coupling is negligible .
this is in sharp contrast to u(1 ) extended modes without the conformal invariance , where @xmath38 is a free parameter and we can take its input to give a large , positive contribution to the beta function , so that the higgs vacuum instability problem is relatively easier to solve . and @xmath84 with a fixed @xmath88 tev , shown in ( @xmath89)-plane with @xmath90 . as a reference ,
horizontal lines are depicted for @xmath91 , @xmath92 ( u(1)@xmath32 case ) , @xmath93 ( orthogonal case ) , and @xmath94 ( u(1)@xmath34 case ) .
( b ) same as ( a ) , but parameter scan for @xmath26 and @xmath51 with a fixed @xmath95 .
, title="fig : " ] ( a ) and @xmath84 with a fixed @xmath88 tev , shown in ( @xmath89)-plane with @xmath90 . as a reference
, horizontal lines are depicted for @xmath91 , @xmath92 ( u(1)@xmath32 case ) , @xmath93 ( orthogonal case ) , and @xmath94 ( u(1)@xmath34 case ) .
( b ) same as ( a ) , but parameter scan for @xmath26 and @xmath51 with a fixed @xmath95 .
, title="fig : " ] ( b ) in order to identify parameter regions to resolve the higgs vacuum instability , we also perform parameter scans for the free parameters @xmath26 , @xmath51 and @xmath84 . in this analysis , we impose several conditions on the running couplings at @xmath96 ( @xmath97 gev is the reduced planck mass ) : stability conditions of the higgs potential ( @xmath98 ) , and the conditions that all the running couplings remain in the perturbative regime , namely , @xmath99 , @xmath100 , @xmath101 , @xmath102 and @xmath103 . in fig .
[ fig : scan1 ] we show the results of parameter scans for @xmath26 and @xmath84 with a fixed @xmath88 tev ( a ) , and for @xmath26 and @xmath51 with a fixed @xmath95 ( b ) , in ( @xmath89)-plane with @xmath104 . as a reference
, we also show horizontal lines corresponding to @xmath91 , @xmath92 ( u(1)@xmath32 case ) , @xmath93 ( orthogonal case ) , and @xmath94 ( u(1)@xmath34 case ) .
the resultant parameter space is very restricted .
for example , the higgs vacuum instability can not be resolved in the classically conformal b - l extended sm , which is also observed in @xcite . and @xmath84 with a fixed @xmath91 , shown in ( @xmath105)-plane .
( b ) the allowed region at the tev scale in ( a ) is magnified , along with the lep bound ( dashed line ) and the lhc bound ( solid line ) from direct search for @xmath1 boson resonance .
the region on the left side of the lines are excluded . here , the naturalness bounds for 10% ( dotted line ) and 30% ( dashed - dotted line ) fine - tuning levels are also depicted . ,
title="fig : " ] ( a ) and @xmath84 with a fixed @xmath91 , shown in ( @xmath105)-plane .
( b ) the allowed region at the tev scale in ( a ) is magnified , along with the lep bound ( dashed line ) and the lhc bound ( solid line ) from direct search for @xmath1 boson resonance .
the region on the left side of the lines are excluded . here ,
the naturalness bounds for 10% ( dotted line ) and 30% ( dashed - dotted line ) fine - tuning levels are also depicted .
, title="fig : " ] ( b ) , but for @xmath106 .
, title="fig : " ] ( a ) , but for @xmath106 .
, title="fig : " ] ( b ) the result of parameter scan for @xmath51 and @xmath107 with a fixed @xmath91 is depicted in fig .
[ fig : scan2_xh=2 ] ( a ) .
the allowed region at the tev scale is magnified in fig .
[ fig : scan2_xh=2 ] ( b ) . here
we also show the current collider bounds from search for @xmath1 boson mediated processes ( details of this analysis will be presented in @xcite ) .
the dashed line is obtained by interpreting the lep results @xcite for effective 4-fermi interactions mediated by a heavy @xmath1 boson , while the solid line corresponds to the bound obtained by interpreting the cms results of @xmath1 boson search @xcite ( the bound obtained from the atlas results @xcite is similar , but slightly weaker ) .
the region on the left side of the lines are excluded .
naturalness bound , which is obtained in the next section , is also shown .
the region , in which the classically conformal u(1)@xmath0 model is natural , will be tested by the lhc run-2 .
same plot but for @xmath106 is shown in fig .
[ fig : scan2_xh=-2.5 ] .
once the classical conformal symmetry is broken and a mass scale is generated , it contributes to the sm higgs boson self - energy in general . hence ,
if the u(1)@xmath0 gauge symmetry breaking scale is very large , we may need a fine - tuning to cancel the radiative corrections by some heavy states associated with the u(1)@xmath0 gauge symmetry breaking .
see @xcite for related discussions .
we consider two heavy states , the right - handed neutrino and @xmath1 boson , whose masses are generated by the u(1)@xmath0 gauge symmetry breaking . once the right - handed neutrinos obtain their majorana masses by the u(1)@xmath0 gauge symmetry breaking , the sm higgs self - energy is induced through the dirac yukawa coupling at the 1-loop level , which is roughly estimated as @xmath108 where we have used the seesaw formula , @xmath109 @xcite , and the quadratic divergence has been dropped and the logarithmic divergence has been ignored . for the stability of the electroweak vacuum ,
we impose @xmath110 as the naturalness .
for example , when the light neutrino mass scale is around @xmath111 ev , we have an upper bound for the majorana mass as @xmath112 gev .
this bound is much larger than the scale that we are interested in , @xmath113 tev . a more important contribution to the higgs self - energy
is generated through the 1-loop diagram with the @xmath114 gauge boson .
this contribution is typical in the u(1)@xmath0 extended sm , where the sm higgs doublet has a non - zero u(1)@xmath0 charge .
it is in sharp contrast to the b - l extension where the sm higgs doublet has no b - l charge , and the @xmath1 boson contribution arises at the 2-loop level @xcite . since the sm higgs self - energy at the 1-loop level with the massive @xmath1 boson includes quadratic and logarithmic divergences , this simple calculation does nt seem consistent with our scheme for the coleman - weinberg mechanism
. it may be more reasonable to calculate a series of 1-loop corrections with the @xmath1 boson running in the loop , with external lines of @xmath115 ( @xmath116 ) , and extract a coefficient of @xmath117 with replacing the other lines by their vevs . since @xmath118
, we expect that the dominant contributions come from 1-loop corrections with external lines of @xmath119 .
we may simply calculate the corrections from the coleman - weinberg potential in eq .
( [ eq : cw_potential ] ) , @xmath120 - \frac{25}{6 } \right ) , \label{eq : cw_potential_loop } \end{aligned}\ ] ] which is from 1-loop corrections with @xmath1 boson running in the loop , with external lines of @xmath121 . here
we ignore @xmath40 in @xmath122.s , although we have concluded from the rough estimate of eq .
( [ eq : contribution_ym ] ) that the contributions are not significant . ] with the 1-loop corrections , we replace one combination of external line @xmath123 and a corresponding vertex @xmath124 to @xmath117 with its vertex @xmath125 .
thus , we evaluate the sm higgs self - energy by @xmath126 if @xmath127 is much larger than the electroweak scale , we need a fine - tuning of the tree - level higgs mass ( @xmath128 ) to reproduce the correct value for the sm higgs vev , @xmath129 .
we simply evaluate a fine - tuning level as @xmath130 here , @xmath131 , for example , indicates that we need to fine - tune the tree - level higgs mass squared at the accuracy of 10% level .
some of fine - tuning levels are shown in figs .
[ fig : scan2_xh=2 ] and [ fig : scan2_xh=-2.5 ] , along with the results of parameter scans .
the classical conformal symmetry with its violation through quantum anomalies could be a solution to the gauge hierarchy problem in the sm . because of the absence of the mass term in the higgs potential in this system , the gauge symmetry breaking should be radiatively induced by the coleman - weinberg mechanism .
unfortunately , we can not simply apply this mechanism to the sm , since the large top yukawa coupling destabilizes the effective higgs potential .
we have extended the sm by introducing an anomaly - free u(1)@xmath0 symmetry , along with three right - handed neutrinos and a u(1)@xmath0 higgs field .
the u(1)@xmath0 symmetry is radiatively broken by the coleman - weinberg mechanism , by which the @xmath1 boson as well as the right - handed neutrinos acquire their masses . through a mixing terms between the u(1)@xmath132 higgs and the sm higgs doublet fields , a negative mass squared for the sm higgs doublet
is generated and , as a result , the electroweak symmetry breaking is triggered .
therefore , all mass generations occur through the dimensional transmutation . in the context of the classically conformal u(1)@xmath0 model , we have investigated a possibility to resolve the sm higgs vacuum instability .
since the gauge symmetry is broken by the coleman - weinberg mechanism , all quartic couplings in the higgs potential except the sm higgs one are very small , and hence their positive contributions to u(1)@xmath0 model , are not effective in resolving the sm higgs vacuum instability . on the other hand , in the u(1)@xmath0 model , the sm higgs doublet has a non - zero u(1)@xmath0 charge , and this gauge interaction positively contributes to the beta function .
in addition , the u(1)@xmath0 gauge interaction negatively contributes to the beta function of the top yukawa coupling , so that the running top yukawa coupling is decreasing faster than in the sm case , and its negative contribution to the beta function of the sm higgs quartic coupling becomes milder . for three free parameters of the model
, we have performed parameter scan , and found a parameter region to solve the sm higgs vacuum instability problem .
we have also considered naturalness of our model .
after the u(1)@xmath0 gauge symmetry breaking , the heavy states , @xmath1 boson and the right - handed neutrinos , contribute to the sm higgs self - energy .
therefore , the self - energy exceeds the electroweak scale , if the states are too heavy . since the sm higgs doublet has non - zero u(1)@xmath0 charge
, this self - energy corrections from @xmath1 boson occur at the one loop level .
this is in sharp contrast with the classically conformal b - l model @xcite , where the higgs doublet has no b - l charge , and the self - energy corrections from @xmath1 boson occur at the two loop level .
the naturalness constraint leads to the upper bound on the @xmath1 boson mass as @xmath2 tev .
the @xmath1 boson with this mass range will be tested at the lhc run-2 in the near future .
we would like to thank h. sugawara , s. hikami , h. shimada , h. ueda , n. haba , and s. okada for discussions and useful comments .
s.o . and d.t .
acknowledge department of physics and astronomy , university of alabama , for their hospitality during their visit .
the work of n.o . is supported in part by the united states department of energy .
in this appendix we present the one - loop rges for the u(1)@xmath0 extension of the sm , which are used in our analysis .
the definitions of the covariant derivative , the yukawa interactions and the scalar potential are given by eqs .
( [ eq : covariant_derivative ] ) , ( [ eq : l_yukawa ] ) and ( [ eq : classical_potential ] ) , respectively .
we only include the top quark yukawa coupling @xmath133 and the right - handed neutrino majorana yukawa coupling @xmath40 , since the other yukawa couplings are negligibly small . the u(1)@xmath0 charges @xmath134 are defined in table [ tab : particle_contents ] .
the rges for the gauge couplings at the one - loop level are given by @xmath135 , \nonumber \\
% \label{eq : u(1)'_rge_g_3 } % % % % % u(1 ) ' rge for g_2 % % % % % \mu \frac{d g_2}{d\mu } & = & \frac{g_2 ^ 3}{(4\pi)^2 } \left [ -\frac{19}{6 } \right ] , \nonumber \\ % % % % % u(1 ) ' rge for g_1 % % % % % \mu \frac{d g_1}{d\mu } & = & \frac{g_1}{(4\pi)^2 } \left [ 12\left(\frac{1}{6}g_1+x_q\tilde{g } \right)^{\!\!2 } + 6\left(\frac{2}{3}g_1+x_u\tilde{g } \right)^{\!\!2 } + 6\left(-\frac{1}{3}g_1+x_d\tilde{g } \right)^{\!\!2 } \right .
\nonumber \\ & + & \left .
4\left(-\frac{1}{2}g_1+x_\ell\tilde{g } \right)^{\!\!2 } + 2\left(x_\nu\tilde{g } \right)^2 + 2\left(-g_1+x_e\tilde{g } \right)^2 + \frac{2}{3}\left(\frac{1}{2}g_1+x_h\tilde{g } \right)^{\!\!2 } + \frac{1}{3}\left(x_\phi\tilde{g } \right)^2 \right ] , \nonumber \\ % % % % % u(1 ) ' rge for g ' % % % % % \mu \frac{d g^\prime}{d\mu } & = & \frac{g^{\prime 3}}{(4\pi)^2 } \left [ 12x_q^2 + 6x_u^2 + 6x_d^2 + 4x_\ell^2 +
2x_\nu^2 + 2x_e^2 + \frac{2}{3}x_h^2 + \frac{1}{3}x_\phi^2 \right ] , \nonumber \\ % % % % % u(1 ) ' rge for \tilde{g } % % % % % \mu \frac{d \tilde{g}}{d\mu } & = & \frac{1}{(4\pi)^2 } \left [ \tilde{g } \left\ { 12\left(\frac{1}{6}g_1+x_q\tilde{g } \right)^{\!\!2 } + 6\left(\frac{2}{3}g_1+x_u\tilde{g } \right)^{\!\!2 } + 6\left(-\frac{1}{3}g_1+x_d\tilde{g } \right)^{\!\!2 } \right
\nonumber \\ & + & 4\left(-\frac{1}{2}g_1+x_\ell\tilde{g } \right)^{\!\!2 } + 2\big(x_\nu\tilde{g } \big)^2 + 2\big ( - g_1 + x_e\tilde{g } \big)^2 \left .
+ \frac{2}{3}\left(\frac{1}{2}g_1+x_h\tilde{g } \right)^{\!\!2 } + \frac{1}{3}\left(x_\phi\tilde{g } \right)^2 \right\ } \nonumber \\ & + & 2g^{\prime 2 } \ !
\left\ { 12x_q \ ! \left(\frac{1}{6}g_1+x_q\tilde{g } \right ) + 6x_u \!\left(\frac{2}{3}g_1+x_u\tilde{g } \right ) + 6x_d \!\left(-\frac{1}{3}g_1+x_d\tilde{g } \right ) \right .
\nonumber \\ & + & 4x_\ell\left(-\frac{1}{2}g_1+x_\ell\tilde{g } \right ) + 2x_\nu \big(x_\nu \tilde{g } \big ) + 2x_e\big(-g_1+x_e\tilde{g } \big ) \nonumber \\ & + & \left .
\left .
\frac{2}{3}x_h \left(\frac{1}{2}g_1+x_h\tilde{g } \right ) + \frac{1}{3}x_\phi \left(x_\phi\tilde{g } \right ) \right\ } \right ] .
\label{eq : rge_gauge } \end{aligned}\ ] ] for the rges for the yukawa couplings at the one - loop level we have @xmath136 , \nonumber \\ % % % % % u(1 ) ' rge for y_m % % % % % \mu \frac{d y_m^i}{d\mu } & = & \frac{y_m^i}{(4\pi)^2 } \bigg [ 4(y_m^i)^2 + 2\sum_j(y_m^j)^2 + \left ( 6x_\nu^2 - 3x_\phi^2 \right ) \ !
\left ( \tilde{g}^2 + g^{\prime 2 } \right ) \bigg ] .
\label{eq : rge_yukawa } \end{aligned}\ ] ] finally , the rges for the scalar quartic couplings are given by @xmath137 , \nonumber \\ % % % % % u(1 ) ' rge for \lambda_\phi % % % % % \mu \frac{d \lambda_\phi}{d\mu } & = & \frac{1}{(4\pi)^2 } \bigg [ \lambda_\phi \big\ { 20\lambda_\phi + 8\sum_i(y_m^i)^2 - 12\big ( x_\phi \tilde{g } \big)^2 - 12\big ( x_\phi g^\prime \big)^2 \big\ } \bigg .
\nonumber \\ & + & \bigg .
2\lambda_{mix}^2 - 16\sum_i(y_m^i)^4 + 6 \big\ { \big ( x_\phi \tilde{g }
\big)^2 + \big ( x_\phi g^\prime \big)^2 \big\}^2 \bigg ] , \nonumber \\ % % % % % u(1 ) ' rge for \lambda_{mix } % % % % % \mu \frac{d \lambda_{mix}}{d\mu } & = & \frac{1}{(4\pi)^2 } \left [ \lambda_{mix } \bigg\ { 12\lambda_h + 8\lambda_\phi + 4\lambda_{mix } + 6y_t^2 + 4\sum_i(y_m^i)^2 \bigg
. \right .
\nonumber \\ & - & \left .
\frac{9}{2}g_2 ^ 2 - 6\left ( \frac{1}{2}g_1+x_h\tilde{g } \right)^{\!\!2 } - 6\big(x_\phi\tilde{g } \big)^2 - 6\big ( x_h g^\prime \big)^2 - 6\big ( x_\phi g^\prime \big)^2 \right\ } \nonumber \\ & + & \left . 12 \left\ { \left ( \frac{1}{2}g_1 + x_h\tilde{g } \right ) \!\ ! \big ( x_\phi \tilde{g } \big ) + \big ( x_h g^\prime \big ) \
! \big ( x_\phi g^\prime \big ) \right\}^{2 } \right ] .
\label{eq : rge_lambda}\end{aligned}\ ] ]
the rges for coupling constants of the sm up to two - loop level @xcite are give by @xmath138 + \frac{g_3 ^ 3}{(4\pi)^4 } \left [ - 26g_3 ^ 2 + \frac{9}{2}g_2 ^ 2 + \frac{11}{6}g_1 ^ 2 - 2y_t^2 \right ] , \nonumber \\
% \label{eq : sm_rge_g_3}\\ % % % % % 2-loop sm rge for g_2 % % % % % \mu \frac{d g_2}{d\mu } & = & \frac{g_2 ^ 3}{(4\pi)^2 } \left [ -\frac{19}{6 } \right ] + \frac{g_2 ^ 3}{(4\pi)^4 } \left[12g_3 ^ 2 + \frac{35}{6}g_2 ^ 2 + \frac{3}{2}g_1 ^ 2 - \frac{3}{2}y_t^2 \right ] , \nonumber \\ % % % % % 2-loop sm rge for g_1 % % % % % \mu \frac{d g_1}{d\mu } & = & \frac{g_1 ^ 3}{(4\pi)^2 } \left [ \frac{41}{6 } \right ] + \frac{g_1 ^ 3}{(4\pi)^4 } \left [ \frac{44}{3}g_3 ^ 2 + \frac{9}{2}g_2 ^ 2 + \frac{199}{18}g_1 ^ 2 - \frac{17}{6}y_t^2 \right ] , \nonumber\\ % % % % % 2-loop sm rge for y_t % % % % % \mu \frac{d y_t}{d\mu } & = & \frac{y_t}{(4\pi)^2 } \left [ \frac{9}{2}y_t^2 - 8g_3 ^ 2 - \frac{9}{4}g_2 ^ 2 - \frac{17}{12}g_1 ^ 2 \right ] \nonumber\\ & + & \frac{y_t}{(4\pi)^4 } \left [ y_t^2 \ ! \left
( - 12y_t^2 - 12\lambda_h + 36g_3 ^ 2 + \frac{225}{16}g_2 ^ 2 + \frac{131}{16}g_1 ^ 2 \right ) \right . \nonumber\\ & + & \left . 6\lambda_h^2 - 108 g_3 ^ 4 - \frac{23}{4}g_2 ^ 4 + \frac{1187}{216}g_1 ^ 4 + 9g_3 ^ 2 g_2 ^ 2 + \frac{19}{9}g_3 ^ 2 g_1 ^ 2- \frac{3}{4}g_2 ^ 2 g_1 ^ 2 \right ] , \nonumber\\ % % % % % 2-loop sm rge for \lambda_h % % % % % \mu \frac{d \lambda_h}{d\mu } & = & \frac{1}{(4\pi)^2 } \left [ \lambda_h \big ( 24\lambda_h + 12y_t^2 - 9g_2 ^ 2 - 3g_1 ^ 2 \big ) - 6y_t^4 + \frac{9}{8}g_2
^ 4 + \frac{3}{8}g_1 ^ 4 + \frac{3}{4}g_2 ^ 2 g_1 ^ 2 \right ] \nonumber\\ & + & \frac{1}{(4\pi)^4 } \left [ \lambda_h^2 \big ( - 312\lambda_h - 144y_t^2 + 108g_2 ^ 2 + 36g_1 ^ 2 \big ) \right . \nonumber\\ & + & \lambda_h y_t^2 \ ! \left ( - 3y_t^2 + 80g_3 ^ 2 + \frac{45}{2}g_2 ^ 2 + \frac{85}{6}g_1 ^ 2 \right ) + \lambda_h \ !
\left ( - \frac{73}{8}g_2 ^ 4 + \frac{629}{24}g_1 ^ 4 + \frac{39}{4}g_1 ^ 2 g_2 ^ 2 \right ) \nonumber\\ & + & y_t^4 \ !
\left ( 30y_t^2 - 32g_3 ^ 2 - \frac{8}{3}g_1 ^ 2 \right ) + y_t^2 \ ! \left ( - \frac{9}{4}g_2 ^ 4 - \frac{19}{4}g_1 ^ 4 + \frac{21}{2}g_2 ^ 2 g_1 ^ 2 \right ) \nonumber\\ & + & \left . \frac{305}{16}g_2 ^ 6 - \frac{379}{48}g_1
^ 6 - \frac{289}{48}g_2 ^ 4 g_1 ^ 2 - \frac{559}{48}g_2 ^ 2 g_1 ^ 4 \right ] .
\label{eq : sm_rge_lambda_h}\end{aligned}\ ] ] in our analysis , we numerically solve these sm rges with the following boundary conditions at @xmath139 @xcite @xmath140 using the inputs , @xmath141 , @xmath76 gev , @xmath77 gev , and @xmath142 gev .
r. hempfling , `` the next - to - minimal coleman - weinberg model , '' phys .
b * 379 * , 153 ( 1996 ) [ arxiv : hep - ph/9604278 ] ; j. r. espinosa and m. quiros , `` novel effects in electroweak breaking from a hidden sector , '' phys .
d * 76 * , 076004 ( 2007 ) [ arxiv : hep - ph/0701145 ] ; w. f. chang , j. n. ng and j. m. s. wu , " shadow higgs from a scale - invariant hidden u(1)s model , phys .
d * 75 * , 115016 ( 2007 ) [ arxiv : hep - ph/0701254 ] ; r. foot , a. kobakhidze and r. r. volkas , `` electroweak higgs as a pseudo - goldstone boson of broken scale invariance , '' phys .
b * 655 * , 156 ( 2007 ) [ arxiv:0704.1165 [ hep - ph ] ] ; r. foot , a. kobakhidze , k. l. mcdonald and r. r. volkas , `` neutrino mass in radiatively - broken scale - invariant models , '' phys .
d * 76 * , 075014 ( 2007 ) [ arxiv:0706.1829 [ hep - ph ] ] ; `` a solution to the hierarchy problem from an almost decoupled hidden sector within a classically scale invariant theory , '' phys .
d * 77 * , 035006 ( 2008 ) [ arxiv:0709.2750 [ hep - ph ] ] ; k. a. meissner and h. nicolai , `` conformal symmetry and the standard model , '' phys .
b * 648 * , 312 ( 2007 ) [ arxiv : hep - th/0612165 ] ; `` effective action , conformal anomaly and the issue of quadratic divergences , '' phys .
b * 660 * , 260 ( 2008 ) [ arxiv:0710.2840 [ hep - th ] ] ; `` neutrinos , axions and conformal symmetry , '' eur .
j. c * 57 * , 493 ( 2008 ) [ arxiv:0803.2814 [ hep - th ] ] ; m. holthausen , m. lindner and m. a. schmidt , `` radiative symmetry breaking of the minimal left - right symmetric model , '' phys .
d * 82 * , 055002 ( 2010 ) [ arxiv:0911.0710 [ hep - ph ] ] ; m. lindner , s. schmidt and j. smirnov , `` neutrino masses and conformal electro - weak symmetry breaking , '' jhep * 1410 * , 177 ( 2014 ) [ arxiv:1405.6204 [ hep - ph ] ] .
s. iso , n. okada and y. orikasa , `` classically conformal @xmath143 extended standard model , '' phys .
b * 676 * , 81 ( 2009 ) [ arxiv:0902.4050 [ hep - ph ] ] ; `` the minimal b - l model naturally realized at tev scale , '' phys .
d * 80 * , 115007 ( 2009 ) [ arxiv:0909.0128 [ hep - ph ] ] .
p. minkowski , phys .
b * 67 * , 421 ( 1977 ) ; t. yanagida , in _ proceedings of the workshop on the unified theory and the baryon number in the universe _ ( o. sawada and a. sugamoto , eds . ) , kek , tsukuba , japan , 1979 , p. 95 ; m. gell - mann , p. ramond , and r. slansky , _ supergravity _ ( p. van nieuwenhuizen et al
. eds . ) , north holland , amsterdam , 1979 , p. 315 ; s. l. glashow , _ the future of elementary particle physics _ , in _ proceedings of the 1979 cargse summer institute on quarks and leptons _ ( m. lvy et al . eds . ) , plenum press , new york , 1980 , p. 687 ; r. n. mohapatra and g. senjanovi , phys . rev
. lett . * 44 * , 912 ( 1980 ) .
g. aad _ et al . _
[ atlas collaboration ] , `` measurement of the higgs boson mass from the @xmath144 and @xmath145 channels with the atlas detector using 25 fb@xmath146 of @xmath147 collision data , '' phys .
d * 90 * , no .
5 , 052004 ( 2014 ) [ arxiv:1406.3827 [ hep - ex ] ] .
g. aad _ et al . _ [ atlas and cms collaborations ] , `` combined measurement of the higgs boson mass in @xmath147 collisions at @xmath148 and 8 tev with the atlas and cms experiments , '' arxiv:1503.07589 [ hep - ex ] .
d. buttazzo , g. degrassi , p. p. giardino , g. f. giudice , f. sala , a. salvio and a. strumia , `` investigating the near - criticality of the higgs boson , '' jhep * 1312 * , 089 ( 2013 ) [ arxiv:1307.3536 [ hep - ph ] ] .
t. s. electroweak [ lep and aleph and delphi and l3 and opal and lep electroweak working group and sld electroweak group and sld heavy flavor group collaborations ] , `` a combination of preliminary electroweak measurements and constraints on the standard model , '' hep - ex/0312023 .
g. aad _ et al . _ [ atlas collaboration ] , `` search for high - mass dilepton resonances in pp collisions at @xmath149 tev with the atlas detector , '' phys .
d * 90 * , no . 5 , 052005 ( 2014 ) [ arxiv:1405.4123 [ hep - ex ] ] .
cms collaboration [ cms collaboration ] , `` search for resonances in the dilepton mass distribution in pp collisions at sqrt(s ) = 8 tev , '' cms - pas - exo-12 - 061 ; v. khachatryan _ et al . _
[ cms collaboration ] , `` search for physics beyond the standard model in dilepton mass spectra in proton - proton collisions at @xmath150 tev , '' jhep * 1504 * , 025 ( 2015 ) [ arxiv:1412.6302 [ hep - ex ] ]
. j. elias - miro , j. r. espinosa , g. f. giudice , g. isidori , a. riotto and a. strumia , `` higgs mass implications on the stability of the electroweak vacuum , '' phys .
b * 709 * , 222 ( 2012 ) [ arxiv:1112.3022 [ hep - ph ] ] . c. coriano , l. delle rose and c. marzo , `` vacuum stability in u(1)-prime extensions of the standard model with tev scale right handed neutrinos , '' phys .
b * 738 * , 13 ( 2014 ) [ arxiv:1407.8539 [ hep - ph ] ] . | we consider the minimal u(1)@xmath0 extension of the standard model ( sm ) with conformal invariance at the classical level , where in addition to the sm particle contents , three generations of right - handed neutrinos and a u(1)@xmath0 higgs field are introduced . in the presence of the three right - handed neutrinos , which are responsible for the seesaw mechanism ,
this model is free from all the gauge and gravitational anomalies .
the u(1)@xmath0 gauge symmetry is radiatively broken via the coleman - weinberg mechanism , by which the u(1)@xmath0 gauge boson ( @xmath1 boson ) mass as well as the majorana mass for the right - handed neutrinos are generated .
the radiative u(1)@xmath0 symmetry breaking also induces a negative mass squared for the sm higgs doublet to trigger the electroweak symmetry breaking . in this context
, we investigate a possibility to solve the sm higgs vacuum instability problem .
the model includes only three free parameters ( u(1)@xmath0 charge of the sm higgs doublet , u(1)@xmath0 gauge coupling and @xmath1 boson mass ) , for which we perform parameter scan , and identify a parameter region resolving the sm higgs vacuum instability .
we also examine naturalness of the model .
the heavy states associated with the u(1)@xmath0 symmetry breaking contribute to the sm higgs self - energy .
we find an upper bound on @xmath1 boson mass , @xmath2 tev , in order to avoid a fine - tuning severer than 10% level .
the @xmath1 boson in this mass range can be discovered at the lhc run-2 in the near future .
satsuki oda@xmath3 , nobuchika okada@xmath4 , and dai - suke takahashi@xmath3 _ @xmath5okinawa institute of science and technology graduate school ( oist ) , + 1919 - 1 tancha , onna - son , kunigami - gun , okinawa , 904 - 0495 , japan _ @xmath6research institute , meio university , + 1220 - 1 biimata , nago city , okinawa , 905 - 8585 , japan + @xmath7department of physics and astronomy , university of alabama , + tuscaloosa , alabama 35487 , usa |
we have employed a box - type potential with hard walls for a quantum dot , which is given by where the position vector @xmath4 , @xmath107 , @xmath108 and @xmath109 are the widths of the potential in the @xmath24 , @xmath110 and @xmath20 directions , respectively . the schrdinger equation for a single electron or hole in a quantum dot
is written as @xmath120 ^ 2 + 4|{\cal m}^{\rm eh}_{\alpha,\alpha}(t)|^2}\,\right ) & \\
\mbox{if $ \hbar\omega\leq { \cal e}_{\rm g}(t)+\varepsilon_\alpha^{\rm e}+\varepsilon_\alpha^{\rm h}$}\\ \\
\frac{1}{2}\left(\hbar\omega-\sqrt{[{\cal e}_{\rm g}(t)+\varepsilon_\alpha^{\rm e}+\varepsilon_\alpha^{\rm h}-\hbar\omega]^2 + 4|{\cal m}^{\rm eh}_{\alpha,\alpha}(t)|^2}\,\right ) & \\
\mbox{if $ \hbar\omega\geq { \cal e}_{\rm g}(t)+\varepsilon_\alpha^{\rm e}+\varepsilon_\alpha^{\rm h}$ } \end{array}\right.\ , \label{a5}\ ] ] where the composite index @xmath121 .
moreover , we get the energy levels of dressed electrons in the lab frame , i.e. , @xmath122 and @xmath123 for @xmath124 .
similarly , we obtain the energy levels of dressed holes @xmath125 and @xmath126 for @xmath127 . | the resonant scattering of surface plasmon - polariton waves by embedded semiconductor quantum dots above the dielectric / metal interface is explored in the strong - coupling regime .
in contrast to non - resonant scattering by a localized dielectric surface defect , a strong resonant peak in the scattering field spectrum is predicted and accompanied by two side valleys .
the peak height depends nonlinearly on the amplitude of surface plasmon - polariton waves , reflecting the feedback dynamics from a photon - dressed electron - hole plasma inside the quantum dots .
this unique behavior in the scattering field peak strength is correlated with the occurrence of a resonant dip in the absorption spectrum of surface plasmon - polariton waves due to interband photon - dressing effect .
our result on the scattering of surface plasmon - polariton waves may be experimentally observable and applied spatially selective illumination and imaging of individual molecules .
most of the previously reported research carried out on the optical properties of condensed matter , including well - known optical absorption and inelastic light scattering , made use of a weak probe field as a perturbation to the system under investigation@xcite . in this weak - coupling limit
, the optical response of electrons depends only on the material characteristics .
however , with increased field intensity , the optical properties of materials are found to depend nonlinearly on the strength of the external perturbation@xcite .
strong photon - electron interaction in semiconductors is known to produce dressed states@xcite with a rabi gap for electrons and create substantial nonlinearity in the semiconductor@xcite .
the presence of an induced polarization field , treated as a source term@xcite arising from photo - excited electrons in metals , allows for a _ resonant _ scattering of surface plasmon - polariton waves@xcite .
this is quite different from the _ non - resonant _ scattering of surface plasmon - polariton waves by a localized dielectric surface defect@xcite or by surface roughness@xcite . in this letter , based on the obtained analytical solution@xcite for the green s function of the coupled quantum dot and semi - infinite metallic material system
, we report for the first time our semi - analytic solutions of the self - consistent equations for strongly coupled electromagnetic field dynamics and quantum kinetics of electrons in a quantum dot above the surface of a thick metallic film . in our formalism
, strong light - electron interaction is manifested in the photon - dressed electronic states and in the feedback from induced optical polarization of dressed electrons to the incident light as well .
our calculated results predict a strong resonant peak in the scattering field spectrum , which is accompanied by two side valleys at the same time .
furthermore , we have discovered that the peak height varies nonlinearly with the amplitude of the surface plasmon - polariton waves .
this clearly demonstrates the effect due to feedback dynamics from photon - dressed electron - hole ( e - h ) plasma inside quantum dots .
in addition , this unique observation in the scattering field spectrum is proven to be correlated with a resonant dip observed in the absorption spectrum@xcite of surface plasmon - polariton waves which can be directly attributed to an effect of the inter - band photon - dressing of electronic states .
our model system consists of a semi - infinite metallic material and a semiconductor quantum dot above its surface .
a surface plasmon - polariton ( spp ) wave is locally excited through a surface grating by a normally - incident light .
this propagating spp wave further excites an inter - band e - h plasma within the quantum dot .
the induced local optical polarization field of the photo - excited e - h plasma is resonantly coupled to the spp wave to produce a splitting in degenerate e - h plasma and spp modes with an anti - crossing gap . by using the green s function @xmath0
, we may express maxwell s equation for the electric field component @xmath1 of an electromagnetic field in a semi - infinite non - magnetic medium in position - frequency space as a three - dimensional integral equation , i.e. , @xmath2 where @xmath3 is a solution in the absence of semiconductor quantum dots , @xmath4 is a three - dimensional position vector , @xmath5 is the angular frequency of the incident light , @xmath6 and @xmath7 are the permittivity and speed of light in vacuum , @xmath8 is an off - surface local polarization field that is generated by optical transitions of electrons in a quantum dot , which generally depends on the electric field in a nonlinear way and should be determined by the optical bloch equations .
additionally , the position - dependent dielectric constant @xmath9 is equal to @xmath10 for the semi - infinite dielectric material in the region @xmath11 , but is given by @xmath12 for the semi - infinite metallic material in the region @xmath13 . by assuming a surface plasmon - polariton wave propagating within the @xmath14-plane ,
we can write @xmath15 \,e^{i{\bf k}_0(\omega_{\rm sp})\cdot{\bf x}_\|}\,e^{-\beta_3(k_0,\,\omega_{\rm sp})x_3}\ , \label{e58}\ ] ] where @xmath16 , @xmath17 and @xmath18 are the unit vectors in the @xmath19 and @xmath20 directions , @xmath21 is the field amplitude , @xmath22 is the field frequency , @xmath23 is the angle of the incident surface plasmon - polariton wave with respect to the @xmath24 direction , @xmath25 is the position vector of the center of a surface grating , and the two wave numbers in eq.([e58 ] ) are given by @xmath26 and @xmath27 with @xmath28\geq 0 $ ] and @xmath29\geq 0 $ ] . here , the in - plane wave vector @xmath30 is produced by the surface - grating diffraction of the @xmath31-polarized normally - incident light , which in turn determines the resonant frequency @xmath22 of the surface plasmon - polariton mode . in order for us to explicitly determine the electric field dependence for @xmath8
, we now turn to the quantum kinetics of electrons in a quantum dot . here , the optical polarization field @xmath8 plays a unique role in bridging the gap between the classical maxwell s equations for electromagnetic fields and the quantum - mechanical schrdinger equation for electrons .
the quantum kinetics of electrons in photo - excited quantum dots should be adequately described by the so - called semiconductor bloch equations@xcite which are a generalization of the well - known optical bloch equations in two ways , namely the incorporation of electron scattering by impurities , phonons and other electrons , as well as the many - body effects on dephasing in the photo - induced optical coherence . for photo - excited spin - degenerate electrons ( holes ) in the conduction ( valence ) band ,
their semiconductor bloch equations with @xmath32 are given , within the rotating - wave approximation , by @xmath33 + \left.\frac{\partial n^{\rm e(h)}_{\ell(j)}}{\partial t}\right|_{\rm rel } -\delta_{\ell(j),1}\,{\cal r}_{\rm sp}\,n_1^{\rm e}\,n_1^{\rm h}\ , \label{e19}\ ] ] where @xmath34 is the spontaneous emission rate , which should be calculated by using the kubo - martin - schwinger relation@xcite and including band gap energy and interband dipole moment renormalizations , and @xmath35 represents the electron ( hole ) level population .
in eq.([e19 ] ) , the boltzmann - type scattering term for non - radiative energy relaxation of electrons ( holes ) is @xmath36 , where @xmath37 and @xmath38 are the scattering - in and scattering - out rates for electrons ( holes ) , respectively , and should be calculated by including carrier - carrier and carrier-(optical ) phonon interactions .
moreover , we know from eq.([e19 ] ) that the total number @xmath39 of photo - excited electrons ( holes ) is conserved .
the induced optical polarization in the semiconductor bloch equations with @xmath32 satisfies the following equations for a spin - averaged e - h plasma , i.e. , @xmath40y_\ell^j+\left(1-n^{\rm e}_\ell - n^{\rm h}_j\right)\left({\cal m}^{\rm eh}_{\ell , j}-y_\ell^j\,v^{\rm eh}_{\ell , j;j,\ell}\right)\ ] ] @xmath41\ ] ] @xmath42\ , \label{e22}\ ] ] where @xmath43 represents the induced interband optical coherence , @xmath44 is the energy level broadening ( due to finite carrier lifetime plus the radiation loss of an external evanescent field ) , @xmath45 is the kinetic energy of dressed single electrons ( holes ) ( see supplementary materials ) . in eq.([e22 ] ) , the diagonal dephasing of @xmath43 , the renormalization of interband rabi coupling , the renormalization of electron and hole energies , as well as the exciton binding energy , are all taken into account . since the e - h plasma is not spin - dependent , they may be excited by both left and right circularly polarized light .
the off - diagonal dephasing of @xmath43 has been neglected due to low carrier density in quantum dots . in eqs.([e19 ] ) and ( [ e22 ] ) , we have introduced the coulomb matrix elements @xmath46 , @xmath47 and @xmath48 , for electron - electron , hole - hole and e - h interactions , respectively .
the steady - state solution of eq.([e22 ] ) , i.e. , subject to the condition that @xmath49 , has been obtained as @xmath50{\cal m}^{\rm eh}_{\ell , j}(t)\ , \label{e23}\ ] ] where the photon and coulomb renormalized interband energy level separation @xmath51 is given by @xmath52 @xmath53 and the matrix elements employed in eqs.([e19 ] ) and ( [ e22 ] ) for the rabi coupling between photo - excited carriers and an evanescent pump field @xmath54 are given by @xmath55\ . \label{e31}\ ] ] in this notation , @xmath56 is the heaviside unit step function , the static interband dipole moment denoted by @xmath57 is given by@xcite ( see supplementary materials ) @xmath58^\ast\,{\bf r}\,u_{\rm v}({\bf r})={\bf d}_{\rm c , v}^\ast\ , \label{e32}\ ] ] where @xmath59 and @xmath60 are the bloch functions associated with conduction and valence bands at the @xmath61-point in the first brillouin zone of the host semiconductor , and the effective electric field coupled to the quantum dot is evaluated by @xmath62^\ast\mbox{\boldmath$e$}({\bf r};\,\omega)\left[\psi^{\rm h}_j({\bf r})\right]^\ast\ , \label{e33}\ ] ] where @xmath63 is the envelope function of electrons ( holes ) in the quantum dot .
next , the photo - induced interband optical polarization @xmath8 by dressed electrons in the quantum dot is given by@xcite @xmath64 @xmath65{\cal m}^{\rm eh}_{1,1}(t)\ , \label{e45}\ ] ] where the profile function @xmath66 comes from the quantum confinement inside a quantum dot . in our numerical calculations
, we chose the quantum dot dimensions as @xmath67 and @xmath68 for along the @xmath69 and @xmath70 directions , respectively , @xmath71 and @xmath72 for the electron and hole effective masses , in terms of the free electron mass @xmath73 , @xmath74 , @xmath75 , @xmath76 for the quantum dot , @xmath77 for the cladding layer , @xmath78 and @xmath79 for the static and optical dielectric constants , @xmath80mev for the energy of optical phonons , @xmath81mev ( @xmath82 ) for the phonon broadening , @xmath83 , and @xmath84k for the lattice temperature .
the silver plasma frequency is @xmath85hz and the silver plasma dephasing parameter is @xmath86hz .
the energy gap @xmath87 for the active quantum - dot material is @xmath88ev at @xmath84k .
figure [ f1 ] presents the absorption coefficient @xmath89 for an spp wave by a quantum dot@xcite , the scattering field @xmath90 of the spp wave , and the energy - level occupations for electrons @xmath91 and holes @xmath92 with @xmath93 as functions of the frequency detuning @xmath94 with bare energies @xmath95 , @xmath96 for ground - state electrons and holes .
a dip is observed at resonance @xmath97 in the upper - left panel , which becomes deeper with decreasing amplitude @xmath21 of the spp wave in the strong - coupling regime due to a reduction of saturated absorption .
however , this dip disappears when @xmath21 drops to @xmath98kv / cm in the weak - coupling limit due to the suppression of the photon dressing effect , which is accompanied by a one - order of magnitude increase in the absorption - peak strength .
the dip in the upper - left panel corresponds to a peak in the scattering field , as may be seen from the upper - right panel of this figure .
the scattering field increases with the frequency detuning away from resonance , corresponding to decreasing absorption .
consequently , two local minima appear on both sides of resonance for the scattering field in the strong - coupling regime .
maxwell - bloch equations couple the field dynamics outside a quantum dot with the electron dynamics inside the dot . at @xmath99kv
/ cm in the lower - right - hand panel , we obtain peaks in energy - level occupations at resonance , which are broadened by the finite carrier lifetime as well as the optical power of the spp wave .
moreover , jumps in the energy level occupation may be seen at resonance due to rabi splitting of the energy levels in the dressed electron states .
the effect of resonant phonon absorption also plays a significant role in the finite value of @xmath100 with energy - level separations @xmath101 .
however , as @xmath21 decreases to @xmath98kv / cm in the lower - left panel , peaks in the energy - level occupations are greatly sharpened and negatively shifted due to the suppression of the broadening from the optical power and the excitonic effect , respectively . additionally , jumps in the energy level occupations become invisible because the rabi - split energy gap in this case is much smaller than the energy - level broadening from the finite lifetime of electrons ( i.e. , substantially dampened rabi oscillations between the first electron and hole levels ) .
although the interband dipole moment of a quantum dot is isotropic in space , the scattering field ( see fig.[f2 ] with @xmath102kv / cm ) in the @xmath69 direction ( upper panel ) and in the @xmath103 direction ( lower panel ) are still different due to the presence of a metallic surface perpendicular to the @xmath103 direction in our system .
however , this isotropic intensity distribution is mostly recovered at @xmath104 .
specifically , the scattering field in the @xmath103 direction is one order of magnitude larger than that in the @xmath69 direction . the field pattern in the lower panel tends to spread in the @xmath103 direction , while the pattern in the upper panel distributes in the @xmath69 direction . from this figure
, we also find that the intensities in both panels follow the pattern of strong - weak - strong - weak - strong as the frequency detuning is swept across @xmath104 , which agrees with the observation of the scattering field at the quantum dot in the upper - right panel of fig.[f1 ] .
color maps for the scattering field around a quantum dot displayed in fig.[f2 ] are for strong coupling between the dot and an spp wave .
we present in fig.[f3 ] the scattering field maps in the weak - coupling regime , where the strong - weak - strong - weak - strong pattern in the strong - coupling regimes has been changed to a weak - strong - weak pattern .
moreover , the spp - wave frequency for the resonant scattering field has been shifted from @xmath104 to @xmath105mev , demonstrating a positive depolarization shift of the optical excitation energy , as may be verified from the upper left - hand panel of fig.[f1 ] .
however , this depolarization effect is completely masked by the occurrence of a local minimum at @xmath102kv / cm . in conclusion , for a strong spp wave , we have demonstrated the unique effect of its resonant scattering by a dynamical semiconductor quantum dot very close to the metal / dielectric interface .
we have also predicted correlation between a resonant peak in the scattering field spectrum and a resonant local minimum in the absorption spectrum of the spp wave . in this letter , we only investigated the coupling between an spp wave and a single quantum dot for the simplest case .
our formalism may be generalized in a straightforward way to include multiple quantum dots close to the surface of a metallic film .
the open surface of the metallic film provides an easy solution to perform biochemical and biomedical tests under a microscope in a laboratory setting .
if the quantum dots are further coated with specially - selected chemically - reactive molecules , they should be expected to be able to adhere to special target tissue if the chemical properties are matched with each other .
therefore , the spontaneous emission by electrons in these quantum dots may be employed non - invasively for near - field imaging of target tissue with very high brightness and spatial resolution . additionally , instead of coupling to the lowest pair of e - h energy levels
, we may choose the surface plasmon frequency for a resonant coupling to the next pair of e - h levels .
in this case , optical pumping from the localized surface plasmon field may transfer a population inversion from the excited pair to the ground pair of e - h levels by exchanging thermal energy with lattice phonons , leading to a possible lasing action .
such a surface plasmon based quantum - dot laser would have a beam size as small as a few nanometers ( not limited by its wavelength ) , which is expected to be very useful to spatially selective illumination of individual molecules ( neuron cells ) in low - temperature photo - excited chemical reactions ( optogenetics and neuroscience ) .
dh would like to acknowledge the support by the air force office of scientific research ( afosr ) .
this research was also supported by contract # fa 9453 - 13 - 1 - 0291 of afrl .
g. gumbs and d. h. huang , _ properties of interacting low - dimensional systems _
( wiley - vch verlag gmbh & co. kgaa , weinheim , germany , 2011 ) , chaps . 4 and 5 .
s. schmitt - rink , d. s. chemla and h. haug , `` _ nonequilibrium theory of the optical stark effect and spectral hole burning in semiconductors _ '' , physical review b * 37 * , 941 ( 1988 ) .
d. dini , r. khler , a. tredicucci , g. biasiol and l. sorba , `` _ microcavity polariton splitting of intersubband transitions _ '' , physical review letters * 90 * , 116401 ( 2003 ) .
y. todorov , a. m. andrews , i. sagnes , r. colombelli , p. klang , g. strasser and c. sirtori , `` _ strong - light - matter coupling in subwavelength metal - dielectric microcavities at terahertz frequencies _ '' , physical review letters * 102 * , 186402 ( 2009 ) .
y. todorov , a. m. andrews , r. colombelli , s. de liberato , c. ciuti , p. klang , g. strasser and c. sirtori , `` _ ultrastrong light - matter coupling regime with polariton dots _ '' , physical review letters * 105 * , 196402 ( 2010 ) .
f. rossi and t. kuhn , `` _ theory of ultrafast phenomena in photoexcited semiconductors _ '' , review of modern physics * 74 * , 895 ( 2002 ) .
v. m. axt and t. kuhn , `` _ femtosecond spectroscopy in semiconductors : a key to coherences , correlations and quantum kinetics _ '' , reports on progress in physics * 67 * , 433 ( 2004 ) .
m. kira and s. w. koch , `` _ many - body correlations and excitonic effects in semiconductor spectroscopy _ '' , progress in quantum electronics * 30 * , 155 ( 2006 ) .
f. jahnke , m. kira and s. w. koch , `` _ linear and nonlinear optical properties of excitons in semiconductor quantum wells and microcavities _ '' , zeitschrift fr physik b * 104 * , 559 ( 1997 ) .
j. a. snchez - gil and a. a. maradudin , `` _ dynamic near - field calculations of surface - plasmon polariton pulses resonantly scattered at sub - micron metal defects _ '' , optics express * 12 * , 883 ( 2004 ) .
r. e. arias and a. a. maradudin , `` _ scattering of a surface plasmon polariton by a localized dielectric surface defect _ '' , optics express * 21 * , 9734 ( 2013 ) .
f. pincemin , a. sentenac and j .- j .
greffet , `` _ near - field scattered by a dielectric rod below a metallic surface _ '' , journal of optical society of american a * 11 * , 1117 ( 1994 ) .
a. a. maradudin and d. l. mills , `` _ the attenuation of rayleigh surface waves by surface roughness _ '' , annals of physics * 100 * , 262 ) 1976 ) .
a. a. maradudin and d. l. mills , `` _ scattering and absorption of electromagnetic radiation by a semi - infinite medium in the presence of surface roughness _ '' , physical review b * 11 * , 1392 ( 1975 ) .
m. g. cottam and a. a. maradudin , `` surface linear response functions '' , in _ surface excitations _ , eds . v. m. agranovich and r. loudon ( north - holland , amsterdam , 1984 ) , pp.1 - 194 .
d. h. huang and p. m. alsing , `` _ many - body effects on optical carrier cooling in intrinsic semiconductors at low lattice temperatures _ '' , physical review b * 78 * , 035206 ( 2008 ) .
e. o. kane , `` _ band structure of indium antimonide _ '' , journal of physics and chemistry of solids * 1 * , 249 ( 1957 ) .
u. bockelmann and g. bastard , `` _ interband absorption in quantum wires
. i. zero - magnetic - field case _ '' , physical review b * 45 * , 1688 ( 1992 ) . |
the recent explosion of interest in @xmath1 supersymmetric yang - mills ( sym ) and @xmath0 supergravity amplitudes is due to the discovery of a host of symmetries and structures that are completely hidden in the traditional feynman diagram approach to computing amplitudes .
perhaps even more surprising are the relations that have begun to emerge between perturbative gauge theory and gravity amplitudes .
tree - level relations between gauge and graviton amplitudes were originally discovered in the field - theory limit of the string - theoretic klt relations @xcite .
one - loop relations between @xmath1 sym and @xmath0 supergravity amplitudes were also first obtained using string theory @xcite , while unitarity methods have been used to derive relations at higher loops @xcite .
while much of the recent spectacular progress has involved planar ( large-@xmath6 ) sym amplitudes , the connection between sym and supergravity amplitudes intimately involves the non - planar ( subleading - in-@xmath7 ) contributions to sym amplitudes .
the connection between gauge and gravity amplitudes has recently been significantly tightened through the discovery of a new color - kinematic duality of gauge theory amplitudes by bern , carrasco , and johansson @xcite .
it was conjectured @xcite and proven @xcite that tree - level @xmath8-gluon amplitudes may be written as a sum over diagrams built from cubic vertices = _
i n_i c_i d_i [ bcj ] in which the kinematic numerators @xmath9 obey precisely the same set of algebraic relations observed by the color factors @xmath10 .
the denominators @xmath11 are products of the inverse propagators corresponding to each cubic diagram . given a tree - level @xmath8-gluon amplitude in a form that respects color - kinematic duality , it was further conjectured @xcite and proven @xcite that @xmath8-graviton amplitudes can be expressed as a sum over the same diagrams = _ i _ i n_i d_i with the color factors replaced by a second copy @xmath12 of the numerator factors . at the loop level
, color - kinematic duality was conjectured @xcite to hold for the numerator factors appearing in the integrands of gauge - theory loop diagrams built from cubic vertices , and this was verified through three loops for the @xmath1 sym four - point amplitude @xcite .
loop - level gravity amplitudes can then be obtained from the same diagrams using a double copy of the numerator factors .
related work on color - kinematic duality appears in refs .
@xcite , and some recent reviews are refs .
@xcite .
recently , carrasco and johansson @xcite demonstrated that the five - point amplitude can be written in a form in which the numerators of the integrands respect color - kinematic duality at one , two@xcite , and three @xcite loops .
the one - loop @xmath1 sym five - point amplitude is @xcite = i g^5 _ s_5 ( _ 12345 i^(p ) ( 12345 ) c^(p)_12345 + _ 12 i^(b ) ( 12;345 ) c^(b)_12;345 ) [ cjintro ] where @xmath13 and @xmath14 are the scalar pentagon and box integrals defined below in , @xmath15 and @xmath16 are the corresponding color factors defined in , and and @xmath17 was presented in ref .
@xcite . ] _
12345 = ^(8)(q ) [ 12][23][34][45][51 ] 4 ( 1,2,3,4 ) , _ 12 = ^(8)(q ) [ 12]^2 [ 34][45][35 ] 4 ( 1,2,3,4 ) [ betaandgamma ] where @xmath18 $ ] are the usual helicity spinor products , @xmath19 , and @xmath20 is the manifestly - supersymmetric delta function of the grassmann - valued supermomentum .
the bcj conjecture @xcite then suggests that the corresponding one - loop @xmath0 supergravity five - point amplitude is @xcite = - ( 2)^5 _ s_5 ( _ 12345 _ 12345 i^(p ) ( 12345 ) + _ 12 _ 12 i^(b ) ( 12;345 ) ) [ sgvintro ] where @xmath21 and @xmath22 are numerator factors of a second copy of @xmath1 sym , and it was verified @xcite that agrees with the known expression in ref .
@xcite .
loop amplitudes of @xmath1 sym theory possess ir divergences which can be regulated through dimensional regularization in @xmath23 dimensions with @xmath24 .
the leading ir divergence of one - loop @xmath8-point amplitudes , such as , goes as @xmath25 @xcite . in a @xmath7 expansion of the sym amplitude
, however , the coefficients of the subleading - color amplitudes have ir divergences that are less severe than the leading - color ( planar ) amplitudes@xcite . at one loop , for example , the ir divergence of the double - trace amplitude ( which is subleading in @xmath7 ) goes as @xmath26 , as ascertained for @xmath8-point functions in ref .
this matches the leading ir divergence of one - loop @xmath0 supergravity amplitudes , which also go as @xmath26 @xcite , suggesting the possibility of linear relations between supergravity amplitudes and subleading - color sym amplitudes .
such relations were found at one and two loops for four - point amplitudes in ref .
@xcite . in this paper
, we propose and prove the relation = -120 i g^5 ( 2)^5 _ s_5 _ 12345 a_5;3 ( 12;345 ) between the one - loop @xmath0 supergravity five - point amplitude @xmath27 and the subleading - color sym amplitude @xmath28 , defined as the coefficient of the double - trace term in the trace expansion of the one - loop @xmath1 sym five - point amplitude @xmath29 ( cf . ) .
this relation is distinct from one proposed in ref .
@xcite between the one - loop supergravity amplitude and the leading - color sym amplitudes @xmath30 , as we discuss at the end of sec .
[ secsgvsym ] .
the relation between loop - level gauge theory and supergravity amplitudes proposed in ref .
@xcite is a diagram - by - diagram map between the _ integrands_. in contrast , the various relations between subleading - color sym and supergravity amplitudes found for one- and two - loop four - point amplitudes in ref .
@xcite , for one - loop four- and five - point amplitudes in ref .
@xcite , and for one - loop five - point amplitudes in the present paper , are between the _ integrated _ amplitudes .
we observe that in all of these cases , the kinematic numerators are independent of the loop momenta and therefore can be factored out , leaving the same set of scalar integrals contributing to both the sym and supergravity amplitudes .
finding relations between the integrated sym and supergravity amplitudes when the numerator factors are dependent on the loop momenta will be more challenging .
this paper is organized as follows . in section 2 ,
we review the color and trace bases for representing sym amplitudes . in section 3
, we decompose the carrasco - johannson representation of the one - loop five - point amplitude into the trace basis . in section 4 ,
we present and prove a linear relation between the one - loop five - point supergravity amplitude and the one - loop subleading - color five - point sym amplitudes .
section 5 contains our conclusions , and in an appendix we review the leading ir divergence of the one - loop subleading - color five - point sym amplitude .
in this section , we review how the one - loop five - point amplitude of @xmath1 sym theory may be expressed in both the color basis and the trace basis , and how the coefficients in each of these bases are related to one another .
a complete color basis for one - loop five - point amplitudes is spanned by the pentagon color diagram in fig .
[ diagrams](a ) c_12345^(p ) = ^g a_1 b ^b a_2 c ^c a_3 d ^d a_4 e ^e a_5 g [ pentagon ] and its permutations ( , see ref .
@xcite ) , where @xmath31 are the structure constants for the su@xmath32 gauge group .
the one - loop five - point amplitude can then be expressed as = _
s_5 1 10 a(12345 ) c_12345^(p ) [ fivepointamp ] where @xmath33 depends on the momenta and helicities of the external states , and the sum is over all permutations of the momentum , helicity , and color index @xmath34 of the external states .
the color factor @xmath35 is manifestly symmetric under cyclic permutations of @xmath36 and antisymmetric under @xmath37 .
bose symmetry implies that @xmath33 is therefore also symmetric under cyclic permutations and antisymmetric under reversal of indices , so we can rewrite the amplitude as a sum over twelve terms = _ s_5/_5 _ 2 a(12345 ) c_12345^(p ) = _
i=1 ^ 12 a_i c_i [ oneloopfivepointcolor ] where we choose an explicit basis c_i = \ { c^(p)_12345 , c^(p)_14325 , c^(p)_13425 , c^(p)_12435 , c^(p)_14235 , c^(p)_13245 , c^(p)_12543 , c^(p)_14523 , c^(p)_13524 , c^(p)_12534 , c^(p)_14532 , c^(p)_13542 } [ colorbasis ] for later convenience .
alternatively , the one - loop five - point amplitude may be expressed in terms of a trace basis @xcite & = & _ s_5/_5 _ 2 a_5;1 ( 12345 ) n + & & + _ s_5/_2 s_3 a_5;3 ( 12;345 ) ( 12 ) [ oneloopfivepointtrace ] where @xmath38 and the matrices @xmath39 are the generators in the defining representation of su@xmath32 , normalized according to @xmath40 .
there are twelve independent single - trace coefficients @xmath30 , which are symmetric under cyclic permutations and antisymmetric under reversal of indices , and ten independent double - trace coefficients @xmath28 , which are symmetric under exchange of the first two indices , and completely antisymmetric under permutations of the last three indices .
we therefore write = _ = 1 ^ 22 a__where the basis @xmath41 consists of the following single - trace terms _ 1 & = & n_7 = n + _ 2 & = & n_8 = n + _ 3 & = & n_9 = n + _ 4 & = & n_10 = n + _ 5 & = & n_11 = n + _ 6 & = & n_12 = n [ singletrace ] and the following double - trace terms _ 13 & = & ( 12 ) _ 18 = ( 13 ) + _ 14 & = & ( 23 ) _ 19 = ( 24 ) + _ 15 & = & ( 34 ) _ 20 = ( 35 ) + _ 16 & = & ( 45 ) _ 21 = ( 41 ) +
_ 17 & = & ( 51 ) _ 22 = ( 52 ) . [ doubletrace ] the color ( [ colorbasis ] ) and trace bases ( [ singletrace ] ) and ( [ doubletrace ] ) are related by c_i = _ = 1 ^ 22 _ i _ .
[ trans ] using @xmath42 t^c ) $ ] to evaluate yields c_1 = _ 1 + _ = 13 ^ 22 _ which gives the first row @xmath43 of the transformation matrix .
the other rows may be obtained through permutations of the external state indices _ i = ( cccccccccccccccccccccc 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & & & & & & & & & + 0 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & & & -1 & -1 & -1 & & -1 & -1 & -1 + 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & & & & -1 & & -1 & & -1 & -1 & + 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & & & & -1 & & & & & -1 + 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & -1 & & & & & -1 & & -1 & & -1 + 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & 0 & & & & -1 & -1 & & & & -1 & + 0 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & -1 & -1 & & & -1 & -1 & -1 & & -1 & -1 + 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & & & -1 & & & & -1 & -1 & & -1 + 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & -1 & -1 & -1 & -1 & -1 & & & & & + 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & & -1 & & & & -1 & -1 & & -1 & + 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & & & -1 & -1 & -1 & & -1 & -1 & -1 & -1 + 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 & -1 & -1 & -1 & -1 & & -1 & & -1 & & ) [ transmatrix ] implies that the coefficients of the color and trace bases are related by a_= _
i=1 ^ 12 a_i _ i [ relatecoeff ] hence a_i = a_i , [ singletracecoeff ] i.e. , the coefficients @xmath33 in the color basis ( [ oneloopfivepointcolor ] ) are precisely equal to @xmath30 , the coefficients of the single - trace terms in . on the other hand ,
these planar amplitudes are well - known to be given by the sum of five 1 m scalar box integrals @xcite .
the leading ir divergence of these planar amplitudes therefore goes as @xmath25 .
equations ( [ transmatrix ] ) and ( [ relatecoeff ] ) also give the coefficients of the double - trace terms ; , a_13 = a_1 -a_2 + a_3 -a_4 -a_5 + a_6 -a_7 + a_8 -a_9 + a_10 + a_11 -a_12 .
[ doubletracecoeff ] comparing we see that each of the double - trace amplitudes can be written as a linear combination of the single - trace amplitudes , a result long known @xcite
. an equivalent way to derive this result is to observe that the rank - twelve matrix ( [ transmatrix ] ) possesses ten null eigenvectors @xmath44 : _ = 1 ^ 22 _ i r_j = 0 , j = 1 , 10 . given that @xmath45 has the form @xmath46 , one sees that @xmath47 .
the existence of these null eigenvectors , together with , implies ten relations among the color - ordered amplitudes _
= 1 ^ 22 a_r_j = 0 , j = 1 , 10 .
for example , the @xmath48 relation is a_13 = a_1 -a_2 + a_3 -a_4 -a_5 + a_6 -a_7 + a_8 -a_9 + a_10 + a_11 -a_12 [ athirteen ] that is , a_5;3 ( 12;345 ) & = & + a_5;1 ( 12345 ) -a_5;1 ( 14325 ) + a_5;1 ( 13425 ) -a_5;1 ( 12435 ) + & & -a_5;1 ( 14235 ) + a_5;1 ( 13245 ) -a_5;1 ( 12543 ) + a_5;1 ( 14523 ) + & & -a_5;1 ( 13524 ) + a_5;1 ( 12534 ) + a_5;1 ( 14532 ) -a_5;1 ( 13542 ) .
[ singledouble ] using the symmetries of the single - trace amplitudes , this can be written a_5;3 ( 12;345 ) & = & + a_5;1 ( 12345 ) + a_5;1 ( 23415 ) + a_5;1 ( 13425 ) + a_5;1 ( 34215 ) + & & + a_5;1 ( 32415 ) + a_5;1 ( 13245 ) + a_5;1 ( 21345 ) + a_5;1 ( 23145 ) + & & + a_5;1 ( 31425 ) + a_5;1 ( 34125 ) + a_5;1 ( 32145 ) + a_5;1 ( 31245 ) precisely the set of cyclically - ordered permutations in eq .
( 7.3 ) of ref .
@xcite .
although the planar amplitudes @xmath30 have a leading @xmath25 ir divergence , implies that the leading ir divergence of the subleading - color amplitude @xmath28 is only @xmath26 , as shown in ref .
we rederive this result in an appendix of this paper .
carrasco and johansson recently derived expressions for one- and two - loop @xmath1 sym five - point amplitudes that manifest color - kinematic duality @xcite . at one - loop
their ansatz takes the form @xcite = i g^5 _ i [ ansatz ] where the sum is over all cubic one - loop five - point diagrams , including relabelings of the external lines , with symmetry factors @xmath49 .
the @xmath50 are the momenta flowing through each of the internal legs of the diagram , which can depend on the external momenta @xmath51 and the loop momentum @xmath52 .
the numerator factors @xmath53 are as - yet - unspecified functions of momenta and helicity .
the color factors @xmath54 include the pentagon color diagram previously specified in , but to obtain a representation of the amplitude that satisfies color - kinematic duality , one must also include the box - plus - line color diagram in fig .
[ diagrams](b ) c^(b)_12;345 = ^a_1 a_2 b ^b c g ^c a_3 d ^d a_4 e ^e a_5 g [ boxplusline ] which is manifestly antisymmetric under @xmath55 and symmetric under @xmath56 .
the box color diagrams are not independent of the pentagon color diagrams ; the jacobi identity implies c^(b)_12;345 = c^(p)_12345 + c^(p)_12543 .
[ pentagonboxrelation ] other color diagrams containing triangles and bubbles need not be included since the corresponding numerator factors @xmath53 vanish for @xmath1 sym theory . in a beautiful analysis employing supersymmmetry , generalized unitarity , and color - kinematic duality of the numerator factors @xmath57 and @xmath58 , carrasco and johannson showed that the numerator factors at one loop are independent of the loop momentum , and so can be written as = i g^5 _ s_5 ( _ 12345 i^(p ) ( 12345 ) c^(p)_12345 + _ 12;345
i^(b ) ( 12;345 ) c^(b)_12;345 ) [ cjfirst ] where @xmath13 and @xmath14 are the scalar integrals shown in fig .
[ diagrams ] i^(p)(12345 ) & = & , [ ipentagon ] + i^(b)(12345 ) & = & , [ ibox ] and @xmath59 and @xmath60 are functions of external momenta and helicities given by , although we will not need their explicit forms in what follows .
one may see from that @xmath61 is invariant under cyclic permutations and reversals of @xmath36 , and @xmath62 is invariant under @xmath55 and @xmath56 . hence and @xmath60 possess further symmetries as well .
] @xmath59 is symmetric under cyclic permutations and antisymmetric under reversal of indices , and @xmath60 is antisymmetric under @xmath55 and symmetric under @xmath56 . on the basis of these symmetries ,
we rewrite as = i g^5 ( _ s_5/_5 _ 2 _ 12345 i^(p ) ( 12345 ) c^(p)_12345 + _ s_5/_2 _ 2 _ 12;345 i^(b ) ( 12;345 ) c^(b)_12;345 ) .
[ cjsecond ] we now recast the carrasco - johansson one - loop amplitude into the trace basis .
first we substitute into to obtain & = & i g^5 _ s_5 ( _ 12345 i^(p ) ( 12345 ) + _
12;345 i^(b ) ( 12;345 ) ) c^(p)_12345 + & = & i g^5 _ s_5 / _ 5 _ 2 ( _ 12345 i^(p ) ( 12345 ) + ) c^(p)_12345 . by comparing with
, one sees that the one - loop planar amplitude is given by a_5;1 ( 12345 ) & = & i g^5 ( _ 12345 i^(p ) ( 12345 ) + ) . [ comparing ] the double - trace color - ordered amplitude follows from : a_5;3(12;345 ) & = & i g^5 ( + _ 12345 i^(p ) ( 12345 ) -_14325 i^(p)(14325 ) + _ 13425 i^(p)(13425 ) 2.5 cm + & & -_12435 i^(p)(12435 ) -_14235 i^(p)(14235 ) + _ 13245 i^(p)(13245 ) + & & -_12543 i^(p)(12543 ) + _ 14523 i^(p)(14523 ) -_13524 i^(p)(13524 ) + & & + _ 12534 i^(p)(12534 ) + _ 14532 i^(p)(14532 ) -_13542 i^(p)(13542 ) [ doubletraceterm ] + & & + 2 + & & + 2 + & & + 2 ) .
carrasco and johansson derived a number of additional relations satisfied by the numerator functions @xmath59 and @xmath63 .
in addition to the symmetries noted above , @xmath60 is invariant under @xmath64 , and therefore under all permutations of @xmath65 , and so can be denoted simply as @xmath66 .
the box integral contribution to the double - trace amplitude ( [ doubletraceterm ] ) can therefore be expressed as a_5;3(12;345 ) |_box
= 2 i g^5 where @xmath67 represents the sum of box integrals i^(b)(12 ) i^(b ) ( 12;345 ) + i^(b ) ( 12;453 ) + i^(b ) ( 12;534 ) [ truncbox ] which is invariant under @xmath55 and all permutations of 345 .
in addition to the above symmetres , the @xmath68 obey _
j=1 ^ 5 _ ij = 0 [ sumgamma ] so that there are only six linearly - independent functions : @xmath17 , @xmath69 , @xmath70 , @xmath71 , @xmath72 , and @xmath73 .
the @xmath59 are expressible in terms of the @xmath68 : _ 12345 = 1 2 ( _ 12 + _ 13 + _ 14 + _ 23 + _ 24 + _ 34 ) [ betasumgamma ] and vice versa _ 12 = _
12345-_21345 [ betagamma ] which implies that the @xmath74 satisfy _ [ ij][kl]m = 0 . in the next section
, we will use these properties to prove a one - loop relation between subleading - color sym and supergravity five - point amplitudes .
given the @xmath1 sym amplitude in a form ( [ cjfirst ] ) satisfying color - kinematic duality , the bcj conjecture @xcite holds that the @xmath0 supergravity amplitude can be written as a double copy @xcite = - ( 2)^5 _ s_5 ( _ 12345 _ 12345 i^(p ) ( 12345 ) + _ 12;345 _ 12;345
i^(b ) ( 12;345 ) ) [ bcjsupergravity ] where @xmath75 and @xmath76 correspond to a second copy of the @xmath1 sym numerators .
it was verified in refs .
@xcite that is equivalent to the known five - point amplitude @xcite .
the additional symmetries of @xmath68 and @xmath77 , allow one to rewrite as = - ( 2)^5 ( _ s_5/_5 _ 2 _ 12345 _ 12345 i^(p ) ( 12345 ) + _
s_5/_2 s_3 _ 12 _ 12 i^(b ) ( 12 ) ) [ supergravity ] where @xmath67 is defined in .
the one - loop supergravity amplitude has a @xmath26 ir divergence @xcite , and so it is plausible that it may be expressed as a linear combination of the double - trace ( subleading - in-@xmath7 ) color - ordered amplitudes @xmath78 , which also have a @xmath26 leading ir divergence .
we now establish such a relation .
consider the following linear combination of double - trace coefficients _
s_5 _ 12345 a_5;3 ( 12;345 ) & = & _ s_5 1 2 ( _ 12345 -_12435 ) a_5;3 ( 12;345 ) + & = & _ s_5 1 2 ( _ 34512 -_43512 ) a_5;3 ( 12;345 ) + & = & _ s_5 1 2 _ 34 a_5;3 ( 12;345 ) + & = & _ s_5 1 6 ( _ 34+_45+_53 ) a_5;3 ( 12;345 ) where we have used as well as the symmetries of @xmath28 and @xmath59 .
since @xmath79 has the same symmetries as @xmath28 , we can restrict the sum to the ten independent double - trace coefficients _ s_5 _
12345 a_5;3 ( 12;345 ) = _
s_5/_2 s_3 2 ( _ 34+_45+_53 ) a_5;3 ( 12;345 ) .
[ result ] each of the ten subleading - color amplitudes @xmath80 appearing in can be rewritten in terms of pentagon integrals @xmath81 and ( sums of ) box integrals @xmath82 using .
first we extract the coefficient of @xmath83 in the double - trace amplitude @xmath84 .
these can be read off the first row of .
each of the following subleading - color amplitudes contain a factor of @xmath85 : & & a_5:3 ( 12;345 ) , a_5:3 ( 23;451 ) , a_5:3 ( 34;512 ) , a_5:3 ( 45;123 ) , a_5:3 ( 51;234 ) , + & & a_5:3 ( 13;245 ) , a_5:3 ( 24;351 ) , a_5:3 ( 35;412 ) , a_5:3 ( 41;523 ) , a_5:3 ( 52;134 ) .
hence the coefficient of @xmath86 in @xmath87 is & & 2 _ 12345 + & = & 10 ( _ 12 + _
14 + _ 23 + _
24 + _ 34 ) _ 12345 + & = & 20 _ 12345 _ 12345 where we used in the last two lines .
next we extract the coefficient of @xmath88 in .
the three independent subleading - color amplitudes that contribute to the coefficient of @xmath88 are a_5;3(12;345 ) |_box & = & 2 + a_5;3(25;341 ) |_box & = & 2 + a_5;3(51;342 ) |_box & = & 2 .
hence the coefficient of @xmath89 in @xmath87 is & & 4 _ 34 + & = & 12 _ 34 _ 34 + 4 ( _ 45 + _ 41 + _ 42 )
_ 34 + 4 ( _ 53 + _ 13 + _ 23 ) _ 34 + & = & 20 _ 34 _ 34 where we used in the last line . assembling all the pieces we have _ s_5 _
12345 a_5;3 ( 12;345 ) = 20 i g^5 ( _ s_5/_5 _ 2 _ 12345 _ 12345 i^(p ) ( 12345 ) + _
s_5/_2 s_3 _ 12 _ 12 i^(b)(12 ) ) . comparing this with the one - loop supergravity amplitude ( [ supergravity ] )
, we obtain the sym - supergravity relation = -120 i g^5 ( 2)^5 _ s_5 _ 12345 a_5;3 ( 12;345 ) [ generalization ] which is the five - point generalization of the one - loop sym - supergravity relation for the four - point amplitude @xcite .
we note that is distinct from the relation between the one - loop supergravity amplitude and the _ leading - color _ sym amplitudes = -1i g^5 ( 2)^5 _ s_5/_5 _ 2 _ 12345 a_5;1 ( 12345 ) [ bbj ] proposed in eq .
( 3.27 ) of ref .
indeed , substituting into one finds & = & - ( 2)^5 _ s_5 1 10 _ 12345 ( _ 12345 i^(p ) ( 12345 ) + 5 _ 12;345 i^(b ) ( 12;345 ) ) + & = & - ( 2)^5 _ s_5 .
then by using the symmetries of @xmath75 together with the relation ( [ betagamma ] ) one verifies that this is precisely equal to , thus providing a direct verification of the equivalence of eqs . (
3.27 ) and ( 3.28 ) of ref .
because the leading - color sym amplitudes goes as @xmath25 , the leading ir divergences must cancel between the terms of , whereas in the individual terms all go as @xmath90 .
in this paper , we have recast the carrasco - johansson representation of the one - loop @xmath1 sym five - point amplitude , which respects color - kinematic duality , into a trace basis consisting of leading- and subleading - color partial amplitudes .
we then proposed and proved a linear relation between the one - loop @xmath0 supergravity five - point amplitude and the subleading - color sym amplitude , guided by the fact that the leading ir divergence of each goes as @xmath26 .
this generalizes earlier one- and two - loop relations for four - point amplitudes @xcite . in all of these cases , the kinematic numerator factors are loop - momentum - independent and so we were able to obtain relations between integrated amplitudes , and not just integrands .
it is a challenging problem to find similar relations when numerator factors depend on loop momenta .
focusing on the ir behavior of the supergravity and subleading - color sym amplitudes may provide clues for discovering further supergravity - sym relations .
we would like to thank h. nastase for his collaboration on our previous work on sym - supergravity relations .
in this appendix , we rederive the result @xcite that the leading ir divergence of the one - loop subleading - color five - point amplitude @xmath28 goes as @xmath26 . recall from that the one - loop five - point coefficients satisfy a^_13 = a^_1 -a^_2 + a^_3 -a^_4 -a^_5 + a^_6 -a^_7 + a^_8 -a^_9 + a^_10 + a^_11 -a^_12 . using this relation ,
it was shown in ref .
@xcite that the subleading - color amplitude @xmath28 can be interpreted as sums of volumes of simple polytopes @xcite . we know from ref . @xcite that the one - loop planar five - point amplitude has a laurent expansion beginning a^_i = - 5 g^2 n32 ^2 ^2 a^_i + ( 1 ) , i=1 , 12 where @xmath91 is the tree - level amplitude , so a^_13 & = & - 5 g^2 n32 ^2 ^2 ( a^_1 -a^_2 + a^_3 -a^_4 -a^_5 + a^_6 + & & -a^_7 + a^_8 -a^_9 + a^_10 + a^_11 -a^_12 ) + ( 1 ) .
[ athirteenone ] the twelve tree - level coefficients appearing in this equation are not independent , however , but are related the kleiss - kuijf relations @xcite . a nice way to derive these relations is to write the tree - level amplitude in a color basis consisting of @xcite & & c_1 = ^a_1 a_2 b^b a_3 c^c a_4 a_5 , 0.8 cm c_2 = ^a_1 a_4 b^b a_3 c^c a_2 a_5 , 0.8 cm c_3 = ^a_1 a_3 b^b a_4 c^c a_2 a_5 , + & & c_4 = ^a_1 a_2 b^b a_4 c^c a_3 a_5 , 0.8 cm c_5 = ^a_1 a_4 b^b a_2 c^c a_3 a_5 , 0.8 cm c_6 = ^a_1 a_3 b^b a_2 c^c a_4 a_5 .
( these six basis elements are denoted @xmath92 in ref .
the other nine @xmath10 are related to these by jacobi identities . )
this color basis can be rewritten using a basis of single - trace terms c_i = _ = 1 ^ 12 _ i _ where @xmath93 are defined as in except that the factor of @xmath6 is omitted at tree level . then _
i = ( cccccccccccc 1 & 0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 1 & 1 + 0 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 0 & 0 + 0 & 0 & 1 & 0 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 0 + 0 & 0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 + 0 & 0 & 0 & 0 & 1 & 0 & 0 & 0 & 1 & 1 & 1 & 0 + 0 & 0 & 0 & 0 & 0 & 1 & 1 & 1 & 0 & 0 & 0 & 1 + ) .
this rank - six matrix possesses six null eigenvectors @xmath44 : _
= 1 ^ 12 _ i r_j = 0 , j = 1 , 6 . given that @xmath45 has the form @xmath94 , one sees that @xmath95 .
the existence of these null eigenvectors , together with , implies six relations among the color - ordered amplitudes _
= 1 ^ 12 a^_r_j = 0 , j = 1 , 6 which are precisely the kleiss - kuijf relations @xcite at five points .
these relations can be used to eliminate @xmath96 through @xmath97 , , @xmath98 . substituting these relations into
demonstrates that the leading @xmath25 ir divergence of @xmath28 vanishes .
z. bern , l. j. dixon , d. c. dunbar , m. perelstein and j. s. rozowsky , _
`` on the relationship between yang - mills theory and gravity and its implication for ultraviolet divergences '' _ , , ` hep - th/9802162 ` .
t. sondergaard , _
`` new relations for gauge - theory amplitudes with matter '' _ , , ` arxiv:0903.5453 ` ; c. r. mafra , _ `` simplifying the tree - level superstring massless five - point amplitude '' _ , , ` arxiv:0909.5206 ` ; s. h. henry tye and y. zhang , _ `` dual identities inside the gluon and the graviton scattering amplitudes '' _ , , ` arxiv:1003.1732 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , t. sondergaard and p. vanhove , _
`` monodromy and jacobi - like relations for color - ordered amplitudes '' _ , , ` arxiv:1003.2403 ` ; p. vanhove , _ `` the critical ultraviolet behaviour of n=8 supergravity amplitudes '' _ , ` arxiv:1004.1392 ` ; y. jia , r. huang and c .- y
. liu , _ `` u(1)-decoupling , kk and bcj relations in @xmath99 sym '' _ , , ` arxiv:1005.1821 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , b. feng and t. sondergaard , _ `` gravity and yang - mills amplitude relations '' _ , , ` arxiv:1005.4367 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , b. feng and t. sondergaard , _ `` new identities among gauge theory amplitudes '' _ , , ` arxiv:1006.3214 ` ; b. feng and s. he , _ `` klt and new relations for n=8 sugra and n=4 sym '' _ , , ` arxiv:1007.0055 ` ; h. tye and y. zhang , _ `` comment on the identities of the gluon tree amplitudes '' _ , ` arxiv:1007.0597 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , b. feng and t. sondergaard , _
`` proof of gravity and yang - mills amplitude relations '' _ , , ` arxiv:1007.3111 ` ; d. vaman and y .- p .
`` constraints and generalized gauge transformations on tree- level gluon and graviton amplitudes '' _ , , ` arxiv:1007.3475 ` ; c. r. mafra , _ `` towards field theory amplitudes from the cohomology of pure spinor superspace '' _ , , ` arxiv:1007.3639 ` ; c. r. mafra , _ `` pss : a form program to evaluate pure spinor superspace expressions '' _ , ` arxiv:1007.4999 ` ; b. feng , s. he , r. huang and y. jia , _ `` note on new klt relations '' _ , , ` arxiv:1008.1626 ` ; y. abe , _ `` holonomies of gauge fields in twistor space 3 : gravity as a square of n=4 theory '' _ , , ` arxiv:1008.2800 ` ; j .-
h . huang , r. huang and y. jia , _
`` tree amplitudes of noncommutative u(n ) yang - mills theory '' _ , ` arxiv:1009.5073 ` ; j. bjornsson , _ `` multi - loop amplitudes in maximally supersymmetric pure spinor field theory '' _ , , ` arxiv:1009.5906 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , t. sondergaard and p. vanhove , _ `` the momentum kernel of gauge and gravity theories '' _ , , ` arxiv:1010.3933 ` ; c. r. mafra , o. schlotterer , s. stieberger and d. tsimpis , _ `` six open string disk amplitude in pure spinor superspace '' _ , , ` arxiv:1011.0994 ` ; y .- x .
chen , y .- j .
du and b. feng , _
`` on tree amplitudes with gluons coupled to gravitons '' _ , , ` arxiv:1011.1953 ` ; h. nastase and h. j. schnitzer , _ `` on klt and sym - supergravity relations from 5-point 1-loop amplitudes '' _ , , ` arxiv:1011.2487 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , b. feng and t. sondergaard , _ `` unusual identities for qcd at tree - level '' _ , , ` arxiv:1101.5555 ` ; z. bern and t. dennen , _ `` a color dual form for gauge - theory amplitudes '' _ , ` arxiv:1103.0312 ` ; j. broedel and r. kallosh , _ `` from lightcone actions to maximally supersymmetric amplitudes '' _ , , ` arxiv:1103.0322 ` ; n. e. j. bjerrum - bohr , p. h. damgaard , h. johansson and t. sondergaard , _ `` monodromy like relations for finite loop amplitudes '' _ , , ` arxiv:1103.6190 ` ; c. r. mafra , o. schlotterer and s. stieberger , _
`` explicit bcj numerators from pure spinors '' _ , , ` arxiv:1104.5224 ` ; r. monteiro and d. oconnell , _ `` the kinematic algebra from the self - dual sector '' _ , , ` arxiv:1105.2565 ` ; y .- j .
du , b. feng and c .- h .
fu , _ `` bcj relation of color scalar theory and klt relation of gauge theory '' _ , ` arxiv:1105.3503 ` ; c. r. mafra , o. schlotterer and s. stieberger , _
`` complete n - point superstring disk amplitude i. pure spinor computation '' _ , ` arxiv:1106.2645 ` . | we derive a linear relation between the one - loop five - point amplitude of @xmath0 supergravity and the one - loop five - point subleading - color amplitudes of @xmath1 supersymmetric yang - mills theory .
bow - ph-151 + brx - th-639 * one - loop sym - supergravity relation for five - point amplitudes * stephen g. naculich@xmath2 and howard j. schnitzer@xmath3 @xmath4_department of physics + bowdoin college , brunswick , me 04011 , usa _ @xmath5_theoretical physics group + martin fisher school of physics + brandeis university , waltham , ma 02454 , usa _ |
recently , many efforts have been devoted to study quantum computation and information theory using different approaches .
these activities have brought new understanding of the fundamental physics associated with quantum theories @xcite .
it is shown that the quantum bit or the qubit , which is based on a superposition of the two possible states , is considered as the most basic building block of such quantum theories .
in fact , qubit and its generalizations are connected to many theories including string theory , d - branes , black holes , toric geometry and supermanifolds @xcite . a particular emphasis put on the link with the black holes and branes obtained from supergravity models in high dimensions .
more precisely , a nice interplay between the stu black holes having eight charges and three qubits have been proposed and developed in @xcite .
more recently , many works have suggested extended models using different ways . in connection with supersymmetry
, superqubits get developed using @xmath9 lie superalgebra @xcite . on the other hand ,
the qudit , which is a @xmath2-level quantum system , emerge naturally in the extended models by considering more than two bosonic states on which the qubit is built .
it is recalled that @xmath1 qudits can be characterized by a couple @xmath0 where @xmath1 and @xmath2 indicate the number of the qudits and the states respectively . here
, we refer to these systems as @xmath0 quantum systems .
inspection shows that there have been some attempts to dealt with such quantum systems but unfortunately with partial results only .
the aim of the paper is to enrich these activities by proposing a new way to deal the qudits using toric geometry and its relation to type ii d - branes placed at singularities of calabi - yau manifolds used in the string theory compactification . in particular , we relate the @xmath0 quantum systems to the holomorphic sections of the line bundles on @xmath1 dimensional projective spaces @xmath10 with degree @xmath4 .
using toric geometry technics , these sections are in one - to - one correspondence with @xmath5 integral points on @xmath1-dimensional simplex .
to give an explicit analysis , a particular interest has been on lower dimensional examples .
then , we explore the local mirror map to elaborate a linkage between the @xmath0 quantum system and type ii d - branes placed at toric singularities . @xmath6 and @xmath7 are analyzed in some details and are found to be related to the mirror of the ale space with the @xmath8 singularity and a generalized conifold respectively .
the present paper is organized as follows . in section 2
, we give a short review on the qubit and qudit systems .
section 3 concerns a toric description of the @xmath6 quantum systems while the generalization to @xmath0 ones is discussed in section 4 . in section 5
, we propose a string theory interpretation in terms of type ii d - branes placed at toric calabi - yau singularities using local mirror symmetry .
section 6 contains some including remarks .
inspired by toric varieties and motivated by the existence of combinatorial calculations in quantum information , we use toric geometry to handle multiple of qudit systems .
concretely , we elaborate a toric description in terms of holomorphic section of line bundles on the projective spaces @xmath3 . then , we propose a stringy interpretation using type ii d - branes placed at mirror calabi - yau singularities used in the string theory compactification . to end , let us first recall the qubit which has been extensively studied using different physical and mathematical approaches @xcite .
the qubit is a two state system which can be realized , for instance , by the position of the electron in the hydrogen atom .
the superposition state of a single qubit is generally given by the following dirac notation @xmath11 where @xmath12 are complex numbers satisfying the normalization condition @xmath13 it is observed that this equation can be interpreted geometrically in terms of the so called bloch sphere @xcite . the analysis can be extended to more than one qubit which has been used to discuss entanglement states .
in fact , the two qubits are four level systems . using the usual notation @xmath14 ,
the corresponding state superposition can be expressed as @xmath15 are complex numbers verifying now the following normalization condition @xmath16 @xmath1 qubits , in fact , are @xmath17 configuration states which can be represented by graphs sharing a strong resemblance with a particular class of adinkras formed by @xmath17 nodes connected with @xmath1 colored lines @xcite .
this observation has been explored to show a new similarity between @xmath1 qubit systems and @xmath17 cycles embedded in @xmath1-dimensional torii @xmath18 . in this way
, a quantum state has been interpreted as the poincar dual of the real homology cycle in @xmath18 on which type ii d - branes can wrap to generate black holes in the compactification of type iia superstring on @xmath18 . as in the case of qubits ,
the qudits represent , for instance , a physical system of spin @xmath19 particles with @xmath2 states . in the basis formed by the states @xmath20
, the general state can be written as follows @xmath21 where @xmath12 are complex numbers satisfying now the extended normalization condition @xmath22 for the organization reason , one may be interested in characterizing the multiple of qudits .
in fact , it should be evident to characterize them by a couple @xmath0 where @xmath1 and @xmath2 indicate the number of the qudits and the states respectively . using a similar qubit notation
, the general state can take the form @xmath23 in the handling of such physical quantities , it is remarked that there are some similarities between qudits and toric geometry .
here , thought , we will be concerned with them .
our main objective is to explore the toric geometry language , and related topics including string theory compactification , to deal with the @xmath0 quantum systems .
in this section , we borrow the idea of toric geometry to discuss qudits and make contact with other physics including the compactification of higher dimensional theories . we thus are considering a basic correspondence by replacing states by holomorphic sections of line bundles on toric varieties . before going ahead , we first start by giving some basic facts on toric geometry being one of the most useful mathematical tools used in superstrings , m and f - theories @xcite . indeed ,
a @xmath1-dimensional toric variety @xmath24 is a complex manifold which can be represented by a toric graph known by polytope @xmath25 consisting of @xmath26 vertices @xmath27 in an @xmath28 lattice satisfying @xmath29 where @xmath30 are called mori vectors .
it is observed that @xmath24 can be fixed by @xmath31 toric data .
it is interesting to note , in passing , that the local calabi - yau condition is satisfied by the following relation @xmath32 familiar examples of toric varieties , which have many applications in both mathematics and quantum physics , are the complex projective spaces @xmath33@xcite . for @xmath33 , @xmath30 reduce to a simple vector @xmath34 with @xmath35 and
the equation ( [ toric ] ) becomes a simple relation given by @xmath36 forming a @xmath1-dimensional simplex which will be explored to discuss holomorphic sections of the line bundles on @xmath33 .
having introduced the mathematical backgrounds , we move now to present a toric description of a single qudit . for a sake of simplicity ,
we consider quantum trit or qutrit associated with the couple @xmath37 .
it is a three level system which can be simulated by a physical system dealing with spin-1 particles .
it can be also realized by the degree of freedom of a photon living in five dimensional space - time , which can be obtained from the string theory compactification on a 5-dimensional compact sapce . in this way ,
a superposition state is a vector of a three dimensional hilbert space , given by latexmath:[\[\label{state1 } |\psi\rangle = a_0|0\rangle+a_1 |1\rangle+a_2 basis .
@xmath12 are complex coefficients satisfying the normalization condition @xmath39 where @xmath40 is the probability of measuring the qutrit in the state @xmath41 .
equation ( [ cp2 ] ) plays the same role as the bloch sphere in the single qubit model associated with the @xmath42 quantum system . to give a toric description of the qudit ,
the state @xmath41 is replaced by a laurent monomial @xmath43 as follows @xmath44 we naturally requires that ( [ state1 ] ) takes the following form @xmath45 where @xmath46 will be interpreted as the complex coordinate of one dimensional projective space @xmath47 .
it recalled that @xmath48 is the simplest example in toric geometry which turns out to play a primordial role in the elaboration of our basic correspondence .
it is considered a building block of higher dimensional toric manifolds explored in string theory compactification in particular in the blow up of ade singularities of calabi - yau manifolds producing four dimensional quantum field models @xcite .
it is known that @xmath47 has a u(1 ) toric action having two fixed points @xmath49 and @xmath50 describing respectively north and south poles of the real two sphere , identified with @xmath51 .
the corresponding toric graph is one dimensional polytope which is just the 1-dimensional simplex identified with the interval @xmath52 $ ] . using laurent polynomials , ( [ pol13 ] )
can be associated with holomorphic sections of line bundles on @xmath47 .
it is observed that such line bundles are associated with integral points on the interval @xcite . in this way
, the number of holomorphic sections of the bundle corresponds to the number of integral points on the interval .
indeed , let us assume that the interval @xmath52 $ ] goes from 0 to 2 .
the holomorphic sections can be identified now with the terms @xmath43 , where @xmath53 .
each integral point , specified by @xmath54 on the interval @xmath55 $ ] , corresponds to a section of the line bundle .
the state @xmath41 , appearing in the @xmath56 quantum system , is associated then with such a holomorphic section
. it should be evident that this extends to the @xmath6 quantum system . in this case
, the general state can be represented by the following polynomial @xmath57 this should be associated with a 1-dimensional simplex involving @xmath2 integral points . with this analogy , we are considering a similarity between holomorphic sections of line bundles on @xmath51 with degree @xmath58 and @xmath6 quantum system .
this correspondence can be illustrated in figure 1 .
quantum system .
, title="fig:",width=377 ]
we can also discuss , from the perspective of toric geometry , the more general case associated with the @xmath0 quantum system .
the analysis requires the incorporation of the @xmath1 dimensional projective space @xmath59 and holomorphic sections of line bundles on it .
these sections will be controlled by @xmath5 integral points placed at a @xmath1-dimensional simplex representing the toric description of the line bundles on @xmath59 with degree @xmath4 .
the general graphic representation is beyond the scope of the present work , though we will consider an explicit example corresponding to @xmath60 .
then , we give some remarks on the general case . for @xmath60 , ( [ qudit ] )
reduces to @xmath61 using the above basic correspondence , this state can be represented by the following polynomial with two complex variables @xmath62 and @xmath63 @xmath64 here @xmath62 and @xmath63 can be considered as local complex variables defining @xmath65 .
it is worth noting that the number of the complex variables is exactly the number of qudits ( here @xmath66 ) . roughly speaking
, @xmath65 is a complex two dimensional toric manifold with an u(1)@xmath67 toric action exhibiting three fixed points @xmath49 , @xmath50 and @xmath68 .
the corresponding polytope is a 2-dimensional simplex belonging to the @xmath69 square lattice .
it is obtained from the intersection of three @xmath47 complex lines defining a triangle ( @xmath70 ) in toric geometry language .
the holomorphic sections of the line bundles on @xmath65 are characterized by integral points on the triangle ( @xmath70 ) . to make contact with toric geometry
, we should regard @xmath71 as integral points placed on a 2-dimensional simplex .
then , we assign them to a monomial @xmath72 . in toric geometry language , this monomial can be interpreted as a holomorphic section of the line bundles on @xmath65 @xcite .
let us illustrate this model for @xmath73 .
the corresponding general state of this qudit reads @xmath74 the integers @xmath75 and @xmath76 should satisfy the following constrains @xmath77 producing holomorphic sections on the line bundles on @xmath78 associated with the @xmath79 quantum system .
they represent a collection of @xmath80 integral points in @xmath81 given by @xmath82 belonging to a 2-dimensional simplex representing a toric realization of the line bundles on @xmath65 with degree 4 , as illustrated in figure 2 . quantum system .
, title="fig:",width=302 ] for a physical system with @xmath2 states , the number @xmath75 and @xmath76 have to be lower than a maximal value in terms of @xmath2 .
these numbers are seen to satisfy this condition @xmath83 with the extra constraint @xmath84 the pairs @xmath71 generate now a toric description of holomorphic sections of the line bundles @xmath65 with degree @xmath85 associated with the @xmath7 quantum system .
it is represented by @xmath86 integral points on a 2-dimensional simplex .
these results can be extended to higher orders of qudits .
more precisely , we can do something similar for the more general case with @xmath87 corresponding to the @xmath0 quantum system using @xmath1 local complex variables defining @xmath59 .
the main point is to find all possible integral points of the @xmath1-dimensional lattice @xmath88 corresponding to the monomials @xmath89 quantum information requires that @xmath90 should satisfy @xmath91 together with the extra constraint @xmath92 producing holomorphic sections on the line bundles on a @xmath59 with degree @xmath4 .
these @xmath5 integral points give a toric description of the @xmath0 quantum system on a @xmath1-dimensional simplex .
toric geometry has been considered as a powerful tool to study mirror symmetry in the context of supersrtring theory compactification on global and local calabi - yau manifolds with singularities in the presence of type ii d - branes @xcite .
on the basis on these activities , we bridge qudits to brane configurations which are mirror to the singularities of local calabi - yau manifolds having toric representations .
in fact , it has been suggested that toric geometry can encode information concerning d - brane configurations in type ii superstrings @xcite .
this can be done by exploring the toric graph to encode the corresponding physical data .
we will see that this may produce a new link between @xmath1 qudits and d - branes moving on local calabi - yau geometries .
given a toric a toric manifold , the local mirror symmetry maps the toric data @xmath93 to an algebraic equation describing the mirror geometry given by @xmath94 where @xmath12 are complex coefficients and @xmath95 monomials satisfying the famous mirror constraint relations @xmath96 there are many ways to solve these constraint equations giving the type ii local mirror geometries @xcite .
these geometries have been extensively studied for bosonic and fermionic calabi - yau manifolds in connection with sigma model with four supercharges @xcite . in what follows
, we will show that the @xmath0 quantum system can be related with the local mirror geometry equations .
for the simplicity , we consider lower dimensional models but we expect that possible generalizations to higher dimensional cases could be done using a similar method . indeed , we consider the case of @xmath97 .
in fact , this model can be related to the mirror symmetry of m - theory on ale space with @xmath8 singularity given by the following algebraic equation @xmath98 where @xmath99 , @xmath100 and @xmath101 are complex variables @xcite .
this equation involves a singularity located at the @xmath102 origin @xmath103 .
the latter can be resolved in two ways either by deforming the complex structure of ( [ a1 ] ) or varying its kahler structure .
these two deformations are equivalent due to the self mirror property of the ale spaces considered as local versions of the k3 surfaces@xcite . using sigma model approach , it is shown that the kahler deformation consists on replacing the singular point by a collection of @xmath47 according to the dynkin diagram of the @xmath8 finite lie algebra . this nice connection between toric geometry and lie algebras
will be explored to build a bridge between local mirror geometries and the @xmath0 quantum systems .
roughly speaking , the above model is dual to the mirror of type iia superstring with @xmath2 d6-branes . identifying @xmath104 , up some details , with the cartan matrix of the finite lie algebra @xmath8 , the mirror geometry equations of the ale space with @xmath8 singularity read as @xmath105 these equations can be solved by the following monomials @xmath106 which can be identified with the holomorphic sections of the line bundles on @xmath48 as we have discussed previously . using mirror equations
( [ mirror ] ) , the polynomial representation of the @xmath6 quantum information system can be identified with the mirror geometry solution given by @xmath107 where @xmath108 and @xmath109 are auxiliary variables which have been introduced to recover the right dimension .
it is worth noting that the quadratic term @xmath110 has no physical importance since it does not affect the modouli space of the mirror geometry .
the relevant physical quantities are @xmath12 specifying the d6-brane locations associated with @xmath2 integral points on a 1-dimensional simplex as illustrated in figure 1 . in this scenario
, we associate the holomorphic sections of the line bundles on @xmath48 with d6-branes in type iia superstring .
the second example we want to discuss is the @xmath7 quantum system . in this case
, the @xmath71 can be associated with the following mirror geometry equation @xmath111 where @xmath71 have been explored to solve the mirror equations .
a close inspection shows that this equation describes the local mirror geometry of a generalized conifold given by @xmath112 where @xmath99 , @xmath100 , @xmath101 and @xmath113 are complex variables .
this geometry has been dealt with in connection with d - branes placed at toric singularities of local calani - yau manifolds as developed in @xcite .
inspired by these activities , the @xmath7 quantum system can be related to the d3-branes probing such a geometry . indeed , the complex parameters @xmath114 can be interpreted as the moduli space of the gauge theory living on the world volume of d3-branes . in this way , @xmath71 give the locations of the d3-branes as illustrated in figure 2 . according to @xcite ,
this model can be t - dual with d4-branes wrapping @xmath115 and stretched between ns - branes placed at various points on @xmath115 .
we believe that this connection deserves more deeper study .
we hope to come back to this issue in future .
in this paper , we have approached qudit systems using toric geometry language and its relation with type ii d - branes placed at calabi - yau singularities explored in string theory compactification .
we have refereed to these systems as @xmath0 quantum systems where @xmath1 and @xmath2 indicate the number of the qudits and the states respectively .
concretely , we have shown that the @xmath0 quantum systems can be related to the holomorphic sections of the line bundles on @xmath1 dimensional projective spaces @xmath3 .
these sections are in one - to - one correspondence with @xmath5 integral points on @xmath1-dimensional simplex associated with the toric realization of holomorphic section of line bundles with degree @xmath4 . using the local mirror map in the toric language ,
we have established a linkage between the @xmath0 quantum systems and type d - branes placed at singularities of local calab - yau manifolds used in the type ii superstring compactification .
it has been found that @xmath6 and @xmath7 can be linked with the mirror of the ale spaces with @xmath8 singularity and a generalized conifold respectively .
we expect that our approach can be adaptable to a broad variety of geometries represented by non trivial polytopes going beyond the simplex geometry .
we anticipate that other concepts used in quantum information could be discussed by implementing extra constraints on the integral points belonging on such polytopes .
we believe that the present study could be considered as a first steep for developing such concepts using toric geometry technics used in the string theory compactification . concretely
, we intend to discuss elsewhere gates and non separated states in future works .
a. belhaj , j. rasmussen , a. sebbar , m. b. sedra , _ on ade quiver models and f - theory compactification _ , j.phys .
* a39*(2006)9339 , hep - th/0512321 .
a. belhaj , _ f - theory duals of m - theory on g2 manifolds from mirror symmetry _ , j. phys . *
a36 * ( 2003 ) 4191 , hep - th/0207208 . | in this paper , we propose a new way to approach qudit systems using toric geometry and related topics including the local mirror symmetry used in the string theory compactification .
we refer to such systems as @xmath0 quantum systems where @xmath1 and @xmath2 denote the number of the qudits and the basis states respectively .
concretely , we first relate the @xmath0 quantum systems to the holomorphic sections of line bundles on @xmath1 dimensional projective spaces @xmath3 with degree @xmath4 .
these sections are in one - to - one correspondence with @xmath5 integral points on a @xmath1-dimensional simplex .
then , we explore the local mirror map in the toric geometry language to establish a linkage between the @xmath0 quantum systems and type ii d - branes placed at singularities of local calabi - yau manifolds . @xmath6 and @xmath7 are analyzed in some details and are found to be related to the mirror of the ale space with the @xmath8 singularity and a generalized conifold respectively .
* keywords * : qudits ; toric geometry , local mirror symmetry ; d - branes and string theory . |
here , we theoretically consider the thermal relaxation of electrons in a metallic grating and the consequences on the transient optical response of the grating .
the relaxation of photoexcited electrons in metals occurs in several stages .
commonly , the fastest stage implies thermalization of electrons resulting in a quasi - equilibrium electron distribution with the electron temperature exceeding that of the crystal lattice . this relaxation is due to efficient electron - electron scattering , and the corresponding times in metals are in the range of a few tens of fs ( for gold , this value is claimed to be about 30 fs ) .
once thermalization is reached , the most universal process is cooling of electrons due to emission of phonons , i.e. transfer of the excess energy to the crystal lattice . for gold ,
the characteristic time of this process , @xmath37 , is about 1 ps . for systems with a spatially inhomogeneous distribution of photoexcitation these processes are complemented by spatial electron energy transfer towards a balanced distribution .
it has a quasi - ballistic character until electron thermalization is reached , switching to thermal diffusion afterwards .
after the thermalization , the electron - lattice energy transfer can be described by the two - temperature model @xcite , which describes the electron - lattice energy balance where thermal diffusion occurs mainly due to electrons possessing superior thermal conductivity in comparison to the lattice .
the underlying equations for electron and lattice tempearture variation with time are : @xmath76 here @xmath77 are the electron and lattice heat capacity , @xmath39 is the electron heat conductivity , @xmath78 is the electron - phonon coupling factor , and @xmath79 describes the external power input into the electronic subsystem . the lattice thermal conductivity is neglected . the common approximation valid for @xmath80 exceeding the debye temperature is @xmath81 with the temperature - independent @xmath82 . in the linear - response regime , corresponding to small excitation fluence , we can neglect the temperature dependence of @xmath39 , @xmath77 .
it is important that in infinite or semi - infinite structures the thermal diffusion is not described by a characteristic time corresponding to an exponential decay , but rather obeys a power - like law .
the situation is essentially different for spatially confined structures .
taking into account that the excitation of plasmonic structure takes place very fast , @xmath79 can be assumed to be a @xmath83-like function of time .
the solution of eq.([eq:2 t ] ) can be written in terms of the initial electron temperature distribution @xmath84 .
after simple algebra we obtain that this solution can be presented as a series @xmath85 here @xmath86 is the electron - phonon energy relaxation time , @xmath87 , and @xmath88 are the eigenmodes of the helmholtz equation @xmath89 with the boundary condition of a zero gradient component of @xmath90 normal to the surface , corresponding to the boundary condition of a thermally isolated structure .
this problem provides eigenvalues @xmath91 , each of which enter into the following two decay rates @xmath92 where @xmath93 is the characteristic thermal diffusion relaxation time of the @xmath27-th mode .
the coefficients @xmath94 are determined by the following expression : @xmath95 where the @xmath96 are the coefficients in the expansion @xmath97 . for metals usually @xmath98 .
in this limit we obtain @xmath99 these expressions reflect the fact that due to the big difference of electron and lattice heat capacities hot electrons , at the first stage , lose their energy due to phonon emission and thermal diffusion reaching a quasi - equilibrium state with the lattice
. this stage is characterized by the times @xmath100 .
then , much slower relaxation occurs on the time scale @xmath101 , where both electron and lattice temperatures finally reach a uniform state .
since we are interested in the analysis of the fast stage of relaxation , we disregard the contributions of the @xmath102-terms to the electron temperature evolution .
let us provide explicit results for the thermal relaxation within a stripe of a plasmonic grating .
we chose the coordinate system such that the stripe is within the region @xmath103 , @xmath104 . from the general expressions for @xmath105 we obtain @xmath106 where @xmath107 the coefficients
@xmath31 are determined by @xmath105 : @xmath108 with @xmath109 for @xmath110 and @xmath111 otherwise .
as we see the relaxation of nonequilibrium electrons has a multi - exponential character , with the exponents strongly dependent on the geometry of the structure ( in the particular case of a stripe , on the dimensions of its cross - section ) .
for actual structures @xmath112 can be as small as a few hundreds of fs .
the mechanism of the optical response to electron heating is based on perturbation of the metal dielectric permittivity by heating of the electrons @xcite . in the linear response regime
we can assume that this perturbation is proportional to @xmath113 .
as a result , the perturbation of the dielectric perimittivity is given by @xmath114 where the derivative is calculated for ambient temperature .
in general , it is a complex number , and we present it in the form @xmath115 . to calculate the perturbation of the optical transmission of the system , @xmath116
, we introduce the coefficients @xmath34 , defined as @xmath117 where @xmath118 is the perturbation of transmission induced by that of the dielectric permittivity @xmath119 . in the linear response regime
we have @xmath120 the authors are grateful to natalia del fatti and fabrice valle for valuable discussions .
the work was supported by the deutsche forschungsgemeinschaft ( projects icrc trr 160 and ak40/7 - 1 ) , volkswagen foundation ( grant 90418 ) , the russian foundation for basic research ( grant no .
16 - 02 - 01065 ) , russian president grant ( id-5763.2015.2 ) .
vga acknowledges department of science and technology ( dst ) , india .
bergman and m.i .
stockman , _ surface plasmon amplification by stimulated emission of radiation : quantum generation of coherent surface plasmons in nanosystems _ , phys .
lett . * 90 * , 027402 ( 2003 ) . v. i. belotelov , i. a. akimov , m. pohl , v. a. kotov , s. kasture , a. s. vengurlekar , achanta venu gopal , d. r. yakovlev , a. k. zvezdin and m. bayer , _ enhanced magneto - optical effects in magnetoplasmonic crystals _ , nat . nanotechnol .
* 6 * , 370 ( 2011 ) . c. brggemann , a. v. akimov , b. a. glavin , v. i. belotelov , i. a. akimov , j. jger , s. kasture , a. v. gopal , a. s. vengurlekar , d. r. yakovlev , a. j. kent , and m. bayer , _ modulation of a surface plasmon - polariton resonance by subterahertz diffracted coherent phonons _ ,
b * 86 * , 121401 ( 2012 ) .
m. pohl , v.i .
belotelov , i.a .
akimov , s. kasture , a.s .
vengurlekar , a.v .
gopal , a.k .
zvezdin , d.r .
yakovlev , and m. bayer , _ plasmonic crystals for ultrafast nanophotonics : optical switching of surface plasmon polaritons _ , phys .
b * 85 * , 081401(r ) ( 2012 ) .
e. carpene , e. mancini , c. dallera , m. brenna , e. puppin and s. de silvestri , _ dynamics of electron - magnon interaction and ultrafast demagnetization in thin iron films _
b * 78 * , 174422 ( 2008 )
. g. r. stewart , _ measurement of low - temperature specific heat _
sci . instr . * 54 * , 1 ( 1983 ) .
anisimov , b.l .
kopeliovich , and t.l .
perelman , _ electron emission from metal surfaces exposed to ultrashort laser pulses _ , zh .
teor . fiz .
* 66 * , 776 ( 1974 ) ( sov .
phys.- jetp * 39 * , 375 ( 1974 ) ) .
m. g. moharam , d. a. pommet , e. b. grann , and t. k. gaylord , _ stable implementation of the rigorous coupled - wave analysis for surface - relief gratings : enhanced transmittance matrix approach _ , j. opt .
a * 12 * , 1077 ( 1995 ) .
h. harutyunyan , a. b. f. martinson , d. rosenmann , l.k .
khorashad , l.v .
besteiro , a.o .
govorov , and g.p .
wiederrecht , _ anomalous ultrafast dynamics of hot plasmonic electrons in nanostructures with hot spots _
, nat . nanotechnol .
* 10 * , 770 - 774 ( 2015 ) . | we demonstrate that in confined plasmonic metal structures subject to ultra - fast laser excitation electron thermal diffusion can provide relaxation faster than the energy transfer to the lattice .
this relaxation occurs due to excitation of nanometer - sized hot spots in the confined structure and the sensitivity of its optical parameters to the perturbation in these regions .
both factors become essential when the plasmonic resonance condition is met for both excitation and detection .
a pump - probe experiment on plasmonic gold lattices shows sub - picosecond relaxation with the characteristic times well - described by a two - temperature model .
the results suggest that dynamical optical response in plasmonic structures can be tuned by selection of the structural geometry as well as the choice of wavelength and polarization of the excitation and detection light .
currently fundamental plasmonic effects are intensely studied in pure and hybrid metallic structures , which possess also huge potential for numerous applications ranging from development of efficient nanolasers with plasmonic cavities to photodynamic cancer treatment with plasmonic nanoparticles @xcite . in this
regard , elucidating and understanding the relaxation dynamics and heat dissipation in plasmonic elements after excitation by a femtosecond light pulse is essential @xcite . in a homogeneous metal , hot electrons loose their energy by emission of phonons , following the ultrafast initial stage of photoexcited carrier thermalization by electron - electron scattering . at room temperature , in most metals phonon relaxation occurs on time scales of the order of a picosecond or less @xcite . in nanopatterned plasmonic structures ,
however , the light absorption becomes strongly inhomogeneous in space and shows in particular hot spots .
similarly , the reflection and transmission coefficients depend strongly on the spatial distribution of temperature . as a result ,
the optical response of the photoexcited structure is determined not only by electron cooling through phonon emission , but also by redistribution of energy in space via electron thermal diffusion ( etd ) . in this letter
we demonstrate that in plasmonic structures etd evolves on timescales comparable to or even shorter than phonon emission , modifying significantly the sub - ps cooling dynamics as manifested by the ultrafast optical response .
nanoplasmonic metal structures are typically embedded in an environment with low thermal conductivity , so that during times shortly after the optical excitation they behave as thermally isolated systems with a correspondingly decelerated heat removal .
nevertheless , we show that in plasmonic structures which support optical excitation of nanometer - sized hot spots , etd becomes feasible on ultrafast sub - ps time scales and provides a decisive contribution to the electron relaxation .
our results suggest that dedicated selection of material and geometry of a confined plasmonic structure allows tuning of its dynamical optical response by controlling the etd process parameters . for the etd studies we used a model system of a plasmonic nanostructure : lateral gold gratings on top of a dielectric garnet layer , as schematically shown in fig .
[ fig1 ] .
these structures can be considered as plasmonic crystals ( pcs ) which support surface plasmon polaritons ( spps ) .
the grating period is comparable to the wavelength of light in the visible / near - infrared spectral range which allows coupling of the far field radiation to the spps and vice versa .
the transmission spectra in fig .
[ fig1](a ) show clear resonances which correspond to propagating spps at the metal - air and metal - dielectric interfaces .
the investigated pcs combine two prime features in a single structure : high tunability of the plasmonic resonances and heat confinement in each of the grating stripes .
the tunability is achieved by variation of the incidence angle and proper choice of the polarization of the excitation light pulse .
this enables one to generate different hot spot distributions within a stripe and to modify the etd flow .
the thermal isolation is ensured by deposition of the pc on the dielectric substrate with low thermal conductivity .
most of the results presented below were obtained for a structure with stripe width @xmath0 nm , stripe height @xmath1 nm , and grating period @xmath2 nm .
in addition , measurements were performed on structures with @xmath3 nm , @xmath4 nm , and @xmath5 nm , to study the scaling of the etd relaxation with the lateral pc size .
additional details on the pc parameters and sample preparation are given in refs .
@xcite . as function of wavelength and incidence angle @xmath6 in pc with stripe width @xmath7 nm .
the spectra were taken with a white light source in @xmath8-pol .
dashed lines indicate spp dispersion at au - air ( 1 ) and au - dielectric ( 2,3 ) interfaces .
( b ) differential transmission transient @xmath9 measured at incidence angle of @xmath10 with @xmath8-polarized pump and @xmath11-polarized probe .
inset shows experimental geometry . ] for measuring the dynamics of the pc optical response to pulsed excitation we implemented a pump - probe technique based on a ti - sa oscillator with a repetition frequency of 80 mhz emitting a broad - band spectrum centered at 800 nm wavelength with a full width at half maximum of about 80 nm .
the optical pulses were chirp - compensated before entering the sample , providing an overall time resolution below 40 fs .
the pump beam hit the sample in the plane perpendicular to the slits , while the probe beam was incident in the slit plane .
the incidence angles of the pump @xmath6 , and the probe @xmath12 , could be tuned by rotating the sample around the axis parallel to the slits .
all measurements were done at room temperature ( 300 k ) .
an example of a differential transmission transient @xmath13 for @xmath14 is shown in fig .
[ fig1](b ) .
the negative spike at zero delay is due to instantaneous non - linear response of the dielectric @xcite .
it is followed by a rapid rise of the transmission with a time constant of about 50 fs which is attributed to the dynamical response of the imaginary part of the dielectric function due to the hot electron thermalization by electron - electron scattering right after the excitation pulse .
the subsequent decay of the differential transmission during about a ps is due to electron - phonon relaxation , affecting mainly the real part of the dielectric function .
finally , there is an offset @xmath15 which originates from an increase of the lattice temperature by the optical excitation . in the following we concentrate on the sub - ps dynamics and substract @xmath15 from the signals .
fig .
[ fig2 ] summarizes the data for different excitation conditions with plasmonic probe .
fig .
[ fig2](a ) compares two transients for which in one case the pump beam polarization is plasmonic ( @xmath8-pol , spps are excited ) and non - plasmonic in the other case ( @xmath11-pol , no spp excitation ) . by switching from plasmonic to non - plasmonic excitation
we observe a strong decrease of the signal due to less efficient excitation of hot carriers , in agreement with our previous work @xcite .
however , the most striking feature is the observation of double exponential decays of the optical transmission change @xmath16 induced by the pump .
we note that the transients scale linearly with pump excitation power ( up to 0.1 mj/@xmath17 ) which suggests that this behavior represents a linear response . in order to quantify the signal contributions we use the function @xmath18 to fit the decay with the first ( second ) term describing the slow ( fast ) component .
variation of polarization and incidence angles of pump and probe results in changes of the amplitudes @xmath19 only , while the times @xmath20 remain constant for a given structure . in the pc with @xmath7
nm fitting the experimental data provides @xmath21 fs and @xmath22 fs .
further , for plasmonic excitation @xmath23 is positive and strongly depends on the incidence angle .
otherwise , for @xmath11-polarized pump the ratio is negative and practically does not depend on @xmath6 ( see fig .
[ fig2](b ) ) .
similar measurements on the pc with @xmath5 nm show a double - exponential decay with @xmath24 fs and @xmath25 fs . for @xmath8- and @xmath11-polarized pump after substraction of @xmath15 .
solid lines are fits with double exponential decay with @xmath21 fs and @xmath22 fs .
( b ) dependence of @xmath23 on incidence angle @xmath6 for the two polarizations .
the data are presented for a pc with @xmath7 nm . dashed lines in ( a ) and ( b ) are guides to the eye . ]
the two - temperature model @xcite used for analysis of the cooling dynamics and its manifestation in the pump - probe experiment are given in the appendix [ a1 ] .
the evolution of the electron temperature @xmath26 has a complex behavior which can be analyzed by decomposing the spatial distribution of @xmath26 into the eigenmodes of the thermal problem .
each eigenmode is characterized by an individual relaxation time , resulting in general in a multi - exponential behavior .
the amplitude weight of each eigenmode is determined by its contribution to the initial spatial distribution of @xmath26 , initiated by the pump pulse .
for the rectangular pc stripe , these eigenmodes are the harmonic fourier components labeled by the indices @xmath27 and @xmath28 for the @xmath29- and @xmath30-directions ( see appendix [ a1 ] ) .
their initial amplitudes @xmath31 are given by their overlap integrals with the initial distribution of @xmath26 .
for the perturbation of the optical transmission by the pump @xmath16 , we obtain @xmath32 here we assume that the perturbation of the dielectric function @xmath33 , which is proportional to the rise of @xmath26 , is real , and @xmath34 are the perturbations of the transmission in response to the change of @xmath33 having the spatial structure of the @xmath35 harmonic .
the decay constant of each component @xmath36 , where @xmath37 is the time of electron energy relaxation due to phonon emission , and @xmath38 . here @xmath39 and @xmath40 are the thermal conductivity and thermal capacity of the electrons , respectively .
note that eq.([eq : perm ] ) reflects the main features of the relaxation process .
the spectrum of relaxation times is determined by geometry and material of the structure and not directly related with its electromagnetic properties . on the other hand , optical access to a particular eigenmode depends on the possibility to couple the corresponding spatial profile of the perturbation to the laser radiation , both in excitation and detection .
this coupling is determined by the plasmonic properties of the structure .
it can be controlled by tuning wavelength , polarization and incidence angle of the laser . in our case ,
the plasmonic resonance enhances the amplitudes of particular harmonics with @xmath41 such that they are comparable to the @xmath42 contribution . if the probe is also in the plasmonic resonance band , its sensitivity to the spatially - inhomogeneous pattern may be enhanced for particular harmonics with @xmath41 . the double exponential decay in fig .
2 suggests that mostly two fourier components contribute to the thermal perturbation .
let us provide some estimates in this respect .
we assume @xmath43 wk@xmath44m@xmath44 , @xmath45 , and @xmath46 jk@xmath47m@xmath48 @xcite , from which we obtain @xmath49 ps , @xmath50 ps .
obviously @xmath51 is very short and comparable to the electron scattering time @xcite .
this means that a thermal conduction approach can hardly be applied to modeling of the electron energy transfer along the @xmath30-direction .
as we are interested in the electron relaxation after the initial stage lasting a few tens of fs , we can analyze the subsequent dynamics using the etd approach solely for the @xmath29-direction considering the @xmath52 harmonics .
another peculiarity of the pc is that even for finite incidence angle the pump excitation pattern is close to an even function of @xmath29 with respect to the center of the stripe . as a result ,
the most important contributions arise from terms with even @xmath27 , corresponding to even harmonics . from calculations of the amplitudes @xmath31
we conclude that the essential fourier components are those with @xmath53 and @xmath54 , both with @xmath52 .
this is in line with the double - exponential behaviour in fig . 2 , so that we can identify @xmath20 with @xmath55 and @xmath56
to obtain quantitative agreement with the measured values , we have to assume a somewhat reduced @xmath39 of about 250 wk@xmath44m@xmath44 .
this reduction may be caused by additional scattering of electrons at the surface and interfaces of the stripes @xcite .
for the pc with @xmath3 nm , theoretical and experimental values match for @xmath39 about 150 wk@xmath44m@xmath44 . the stronger reduction of @xmath39 in this case may be due to the smaller stripe thickness , enhancing the role of surface electron scattering .
the amplitudes @xmath19 deserve special attention . to that end , we show in figs . [ fig3](a ) and ( b ) the spatial patterns of absorbed energy in a single stripe for plasmonic ( @xmath8-pol , top ) and non - plasmonic ( @xmath11-pol , bottom ) excitation .
the bright areas indicate hot spot formation .
the absorbed energy is proportional to @xmath57 , where @xmath58 is the complex amplitude of the electric field vector of the optical pump wave in the metal from which the initial electron temperature distribution arises .
the spatial distribution of @xmath58 was calculated applying a rigorous coupled - wave analysis @xcite .
the excitation pattern for plasmonic polarization is shown for the incidence angle @xmath59 , which is close to the plasmonic resonance ; for non - plasmonic excitation it is practically independent on @xmath6 .
-polarized light ( @xmath60 ) and @xmath11-polarized light , respectively .
bright areas correspond to hot spots . in ( c )
the @xmath29-dependence of absorbed energy averaged over the @xmath30-direction is shown for various incidence angles at @xmath8-pol . and normal incidence for @xmath11-pol .
for comparison also the harmonic with @xmath27=2 in the fourier analysis of the thermal conduction problem , which provides fast relaxation with time constant @xmath56 , is shown . in
( d ) the angle dependence of coefficients @xmath61 and @xmath62 is shown . ] for plasmonic excitation a pronounced hot spot structure is present scaling spatially with the wavelength of the uniform film spps , excited by the pump in the perforated structure .
in fig .
[ fig3](c ) the absorbed energy ( averaged over the @xmath30-direction ) is shown for different incidence angles @xmath6 demonstrating the absorption resonance for @xmath59 .
the hot spots contribute considerably to the modes with @xmath63 .
obviously , the excitation spatial profile can not be described by the combination of the uniform , @xmath53 , and @xmath54 harmonics ( the latter is shown in fig . [ fig3](c ) by the dashed line ) .
however , for the harmonics with @xmath64 , @xmath65 is very short .
similar to the case of vertical relaxation , the corresponding harmonics probably relax in a non - diffusive way being unresolvable for pump - probe delays exceeding a few tens of fs . for non - plasmonic excitation ( shown for normal incidence only )
the absorption is much weaker .
however , also here hot spots are formed at the sharp edges of the rectangular stripes , which can be well described by the @xmath53 and @xmath54 terms in the fourier series .
next , we have to consider how the local variation of the dielectric fucntion translates into changes of transmitted probe intensity
. the calculated angle dependencies of the coefficients @xmath61 and @xmath62 are shown in fig.[fig3](d ) .
the results explain the negative sign of the ratio @xmath23 observed for non - plasmonic excitation . indeed , the coefficients @xmath61 and @xmath62 have opposite signs for all considered angles .
@xmath66 , being the average absorbed excitation energy , is per se positive .
localization of the hot spots near the stripe edges for non - plasmonic pump ensures @xmath67 . for the plasmonic configuration
, we do not achieve quantitative agreement of experiment and theory .
the experimental data suggest negative @xmath68 , which would occur for intense hot spots in the middle of the stripes .
they do appear in the calculations due to the excitation of the plasmonic resonance ( see fig .
3(a ) ) , but the calculated hot spots near the wall edges provide larger contributions for incidence angles of @xmath69 0 , 8 , and 17@xmath70 .
thus @xmath68 should be positive there as well . according to our calculations , the absorbed energy near the edges is strongly sensitive to the stripe shape .
therefore , we tentatively attribute the discrepancy of experiment and theory to deviation of the stripe shape from the assumed rectangular one .
the developed model allows us to estimate the role of etd also in plasmonic structures with a shape differing from a stripe and/or fabricated from a metal other than gold .
etd becomes important if its characteristic relaxation rate @xmath71 , where @xmath72 is the typical size of the hot - spot pattern in the plasmonic structure , is comparable to or exceeds @xmath73 . from setting the two rates
equal , we can determine the required geometrical dimension of the plasmonic structure that needs to be undercut in order to reach these conditions , @xmath74 , which establishes a threshold for etd becoming efficient . in table
[ tab : matdata ] we list the @xmath75 as well as the @xmath37 for various metals at 300 k. by tailoring material and geometry of the structure it is possible to obtain etd relaxation faster than that due to phonon emission . in this case also shape design is important to provide a strongly inhomogeneous optical excitation pattern .
we would like to address the recent paper by harutyunyan et al .
@xcite in which optical excitation of hot spots in plasmonic nano - discs was attributed to the appearance of a short optical response in differential reflection transients . however , the relaxation time constant could not be measured due to insufficient time resolution .
moreover , the role of etd was not discussed as possible origin of the fast relaxation component . .electron - phonon cooling time and @xmath75 for different metals .
the material parameters are taken from @xcite [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] in conclusion , we demonstrate that the relaxation in a photoexcited plasmonic grating can be enhanced by thermal diffusion of hot electrons on a sub - picosecond time scale , essentially below the value established for electron cooling by phonon emission .
this fast relaxation is possible due to formation of nanometer - sized hot spots under ultra - fast optical excitation if the plasmonic resonance condition is met .
our analysis based on the two - temperature model provides good agreement with the experimentally observed relaxation times and allows us to formulate a geometric criterion for the importance of thermal diffusion which can be adjusted through the plasmonic structure material and design . |
a decade ago gould ( 1992 ) suggested that the earth s orbital motion may affect long duration events of gravitational microlensing , producing a periodic modulation of the event light curve .
the first such example of this behaviour was detected by alcock et al .
( 1995 ) , and a number of additional events have been reported by several groups ( e.g. , mao 1999 ; smith , mao & woniak 2002a ; bennett et al .
2002a ; bond et al .
2001 ) . however , most of these events were of relatively short duration and showed a strong asymmetry in the light curves , rather than a yearly modulation .
two events were long enough to make a yearly modulation clearly noticeable : ogle-1999-bul-32 = macho-99-blg-22 ( mao et al .
2002 ; bennett et al .
2002b ) and ogle-1999-bul-19 ( smith et al .
2002b ) .
while the earth s motion may generate a photometric parallax effect , a similar anomaly may be due to binary motion of the source ( sometimes referred to as ` xallarap ' ; griest & hu 1992 ; han & gould 1997 ; paczyski 1997 ) and/or the lens .
in fact such a binary - source modulation was detected in the microlensing event macho - smc-1 ( alcock et al .
1997 ; palanque - delabrouille et al . 1998 ; udalski et al . 1997 ) and macho 96-lmc-2 ( alcock et al . 2001 ) . while in these two cases
the binary source period was much shorter than the event time scale , an alternative situation may also be encountered .
in fact , smith et al .
( 2002b ) pointed out that even in the case of a very long event exhibiting an annual modulation of its light curve , a model based on a binary source moving with a one year orbital period and suitable inclination and orientation of the orbit may generate a light curve identical to that expected from the parallax effect .
while such a coincidence is very unlikely , it is also verifiable in that a binary source should display periodic variations in its radial velocities .
as most parallax microlensing events are not very long it is likely that the effects they show may be well described with a constant acceleration of the earth , the lens and/or the source , with no need to invoke a specific orbit .
the aim of this paper is to investigate this possibility and to present several consequences of the proposed approach . in section 2
we describe a simple model that incorporates a constant - acceleration term into the microlensing formalism .
we then test the efficacy of this model in section 3 by fitting a number of microlensing events that have been previously identified as exhibiting variations from the standard constant - velocity model of microlensing .
the events that we have selected were all identified as potential parallax microlensing events , and in this section we aim to test whether this constant - acceleration model is able to successfully reproduce this behaviour . in this section
we also mention that there is a generic degeneracy with the constant - acceleration model , and that there may be an equivalent degeneracy for parallax models . in section 4
we summarise our work and discuss some of the implications of our findings .
in particular , we note that the effect of constant - acceleration perturbations is likely to be of little importance for binary lenses , and such perturbations are therefore most - likely to be due to accelerations of the observer and/or the source .
in the simplest case of a microlensing event , the light curve is calculated assuming that the relative motion of the source , the lens , and the observer is linear and constant ( paczyski 1986 ) .
the next step in allowing for a more complicated relative motion is to expand it in a power series , and to retain the first two terms , i.e. the velocity , @xmath0 , and the acceleration , @xmath1 . for the purposes of this paper ,
we define the velocity , @xmath2 , to be the velocity of the lens relative to the observer - source line - of - sight .
this acceleration is therefore a relative acceleration , but since the following analysis is likely to be valid only for lenses in inertial motion ( see section [ disc ] ) , this acceleration can be thought of as the acceleration of the observer and/or the source .
an acceleration related velocity vector may be defined as @xmath3 , where @xmath4 is the event time scale : @xmath5 , @xmath6 is the einstein radius and @xmath7 is the velocity at the minimum separation ( peak magnification ) .
we expect that any ` anomaly ' in the light curve will depend on two dimensionless numbers : the ratio of the two velocities : @xmath8 , and the angle between them @xmath9 .
notice @xmath10 is simply the physical acceleration in units of @xmath11 .
in general , an acceleration will produce asymmetry in the microlensing light curve , and this is the effect which is used to select parallax events .
however , if the acceleration is mostly in the direction perpendicular to the velocity vector @xmath12 then there will be no asymmetry , just a change in the shape that may be difficult to notice .
[ fig : model ] shows four example light curves with different accelerations .
all four cases have a minimum impact parameter of @xmath13 , and the source is assumed to move along the @xmath14-axis in the absence of acceleration .
the dimensionless acceleration parameters along the @xmath14 and @xmath15 directions , @xmath16 and @xmath17 , are indicated in each panel .
the top left panel shows a light curve in which the acceleration is @xmath18 from the @xmath14-axis while the bottom left panel shows an example where the acceleration is in the same direction as the source motion . as expected , both light curves show clear asymmetries .
the two right panels show light curves in which the acceleration is perpendicular to the @xmath14-axis .
the light curves are symmetric in such cases .
the effect of acceleration may not be noticed , and a standard fit forced upon the event may generate incorrect values of the impact parameter and the event time scale ; we return to this issue briefly in the discussion .
a thorough analysis of all such possibilities in current experiments is beyond the scope of this paper . instead
, we concentrate on the readily available candidate parallax events , in order to check how many of them can be well fitted by adding an acceleration vector to the standard model .
in this section we implement the above constant - acceleration microlensing model and test its efficacy by fitting a number of microlensing events . by analysing previously identified candidate
parallax microlensing events , we aim to investigate whether this constant - acceleration model is able to suitably reproduce the observed ( i.e. seemingly parallactic ) behaviour .
we shall first outline the standard constant - velocity microlensing model . for the standard microlensing light curve , the source position ( in units of einstein radius )
is given by [ eq : linear ] x = t - x_0(t ) , y = u_0 , where @xmath19 is the time of maximum magnification and @xmath20 is the corresponding ( minimum ) impact parameter .
the magnification is given by ( paczyski 1986 ) [ eq : mag ] a(t ) = u^2 + 2 u , u=. with acceleration , the @xmath21 coordinates must be modified to [ eq : pos ] x = x_0(t ) + 1 2 a x^2_0(t ) , y = u_0 + 1 2 a x^2_0(t ) , where @xmath22 is defined in eq .
( [ eq : linear ] ) , @xmath10 is the physical acceleration in units of @xmath11 , and the magnification is again calculated using eq .
( [ eq : mag ] ) .
notice that , since the transverse velocity changes , @xmath23 should be understood as the einstein radius crossing time corresponding to the transverse velocity @xmath7 at time @xmath19 . also , in general , @xmath20 may no longer rigorously be the minimum impact parameter , but may still provide a useful approximation , particularly if @xmath10 is small . to fit a microlensing light curve with the standard model , one needs a minimum of four parameters , @xmath20 , @xmath19 , @xmath23 , and the baseline flux ( or magnitude ) , @xmath24 .
one also often requires an extra blending parameter , which accounts for the additional light contributed by the lens and/or other stars in the crowded field . to do this
, we use a parameter , @xmath25 , defined as the fraction of light contributed by the lensed source at the baseline . to incorporate the constant acceleration ,
we need two extra parameters , @xmath10 and the angle @xmath9 ( or equivalently , @xmath16 and @xmath17 ) .
we look for any degeneracy in the parameter space for this constant - acceleration model by fitting each event 1000 times , with the initial parameter guesses selected from wide range of randomly chosen values .
we find that two sets of parameters exist for each @xmath26 minimum .
the parameters are identical except for the time - scale , @xmath23 , and the acceleration parameters , @xmath27 and @xmath9 .
however , the magnitude of the physical acceleration in units of @xmath28 ( i.e. , @xmath29 , as defined in 2 ) is the same for both fits , although the direction of this vector ( i.e. , @xmath9 ) differs . in appendix a1 , we show that this degeneracy can be understood analytically .
an example of this degeneracy is illustrated in the inset of fig .
[ lc : sc6 ] for the microlensing event sc6_2563 ; even though the two trajectories look very different they both result in identical light curves .
the full analysis of the event sc6_2563 is described in section [ sc6 ] .
it should be noted that one may expect this degeneracy to manifest itself in parallax models . in particular , this could be important for short - duration events that exhibit weak parallax signatures , since in this case the earth s acceleration can be approximated to be constant ( i.e. the regime in which this degeneracy is expected to occur ) . by testing this hypothesis on the event sc6_2563 , we are able to verify the existence of this degeneracy
. however , it seems that this may only become apparent for weak parallax candidates , since a more convincing parallax candidate ( sc33_4505 ; section [ sc33 ] ) shows no such degeneracy ( see appendix a ) .
a full discussion of this degeneracy , including the application to events sc6_2563 and sc33_4505 , can be found in appendix a2 .
we proceed by analysing a selection of events that have previously been identified as candidate parallax microlensing events .
the parallax model used here requires 7 parameters : the 5 parameters from the standard model plus two additional parameters to describe the lens trajectory in the ecliptic plane , namely the einstein radius projected onto the observer plane , @xmath30 , and an angle @xmath31 in the ecliptic plane .
@xmath31 is defined as the angle between the heliocentric ecliptic @xmath14-axis and the normal to the trajectory ( this geometry is illustrated in fig . 5 of soszyski et al .
these candidate parallax events have been selected to provide a broad range of parallactic signatures , ranging from marginal cases to those displaying more prominent effects .
this event was detected by the ogle collaboration ( udalski et al .
its asymmetric nature was identified during a parallax search of the ogle - ii database ( woniak et al .
2001 ) , reported in smith et al .
( 2002a ) .
it was classified as a ` marginal ' candidate , since the deviations from the standard model are not particularly pronounced and the improvement in @xmath32 is only slight .
the best - fit parameters for the standard and parallax models are given in table [ tb : sc6 ] , and the corresponding light curves are plotted in fig .
[ lc : sc6 ] .
the inset of this figure shows the two degenerate trajectories for the constant - acceleration model ; as was discussed above , each @xmath26 minimum has two degenerate fits . both of these trajectories produce identical light curves , even though the parameter values for @xmath23 , @xmath10 and @xmath9 differ
this table also includes an additional parallax fit , which has a @xmath26 difference of 0.8 compared to the best - fit parallax model .
this is a manifestation of the parallax degeneracy that is analogous to this constant - acceleration degeneracy .
see section [ model ] and appendix a for a discussion of these degeneracies . it would be expected that a constant - acceleration fit may be suitable for this event since the duration is not very long , and the asymmetry shows no signs of modulation . indeed , when this event is fit with the above constant - acceleration model the best - fit @xmath32 value improves upon the best - fit parallax value from 347.4 to 332.6
. if the error bars are rescaled so that the best - fit @xmath26 per degree of freedom is renormalised to unity , then the difference in @xmath26 between these two fits is 9.5 for no additional free parameters , i.e. a significant 3-@xmath33 improvement .
the best - fit parameters and corresponding light curve for the acceleration model are given in table [ tb : sc6 ] and fig .
[ lc : sc6 ] , respectively . from this figure
it is clear that the parallax and constant - acceleration models produce similar fits .
however , both models seem to over - predict the flux around @xmath34 days , implying that there may be some non - constant contribution to the acceleration that is not parallactic in nature . notice that the standard microlensing model predicts a blending parameter @xmath35 while other models predict @xmath36 . correspondingly , the einstein radius crossing time is about 25% larger for the standard model than for the other models ; this has implications for the optical depth estimate in the experiments ( see section [ disc ] ) .
this event was detected in real - time by the ogle early warning system ( udalski et al .
1994 ) toward the carina spiral arm .
it has been found to exhibit systematic deviations from the standard model ( mao 1999 ) .
the duration for this event is well over 100 days , and it was concluded that these deviations are due to the parallax effect . as with sc6_2563 ,
the deviations from the standard model for ogle-1999-car-1 show no signs of modulation , and so one would expect that this event may be suitably approximated by a constant acceleration model .
this appears to be correct , since the best - fit constant - acceleration model is able to prove a very similar fit ( with a slightly worse @xmath26 value of 619.4 compared with the parallax best - fit value of 614.8 ) .
the best - fit parameters for all models are given in table [ tb : car ] , and the corresponding light curves are shown in fig .
[ lc : car ] .
it is interesting to note that both of the degenerate acceleration fits have values of @xmath37 greater than 1 ( @xmath38 , and @xmath39 ) .
this implies that @xmath40 , the change in velocity due to this constant acceleration in a time @xmath23 , is greater than the @xmath7 , the velocity at the point of closest approach . also of note is the difference in @xmath23 among the three models .
the values range from @xmath41 days for the parallax model , to as much as @xmath42 days for the constant - acceleration model .
these differences in the time - scale are closely connected with their differences of the blending parameter , due to potential degeneracy between these two parameters ( woniak & paczyski 1997 ) .
this ogle - ii event was also found during the search of smith et al .
( 2002a ) , but was classified as a ` convincing ' parallax candidate .
unlike sc6_2563 , the duration of this event was much longer ( @xmath43 days , cf .
@xmath44 days for sc6_2563 ) , and the data show clear signs of deviation from the standard model in both the rising and declining branch of the light curve .
therefore , if this asymmetry is indeed due to the parallax effect then one would suspect that a constant - acceleration model would be unable to reproduce this behaviour ( since the acceleration of the earth can not be considered constant during this time span ) . on the other hand , if this asymmetry is due to other causes , the constant - acceleration model may provide a better fit .
the best - fit parameters for the standard and parallax models are given in table [ tb : sc33 ] , and the corresponding light curves are shown in fig .
[ lc : sc33 ] .
a problem arises when this event is fit with the constant - acceleration model , in that two minima are identified , with @xmath45 and @xmath46 , respectively . due to the degeneracy discussed in 3 and appendix a1 ,
there are two degenerate fits corresponding to each of these two @xmath32 values .
the two degenerate fits that correspond to @xmath45 have extremely large time - scales ( @xmath47 and @xmath48 ) , and the lensed star contributes only 2.3 per cent of the total baseline light ( i.e. , @xmath49 ) .
these values do not seem physical to us , and therefore the fits corresponding to this @xmath26 minimum are disregarded .
the best - fit parameters for the two degenerate fits corresponding to @xmath46 are more ` feasible ' ; they are given in table [ tb : sc33 ] and the corresponding light curve is shown in fig .
[ lc : sc33 ] .
however , the discarded ` unfeasible ' constant - acceleration models have an overall best @xmath26 value ( 187.3 ) .
since this @xmath26 value is better than the best - fit parallax model , it highlights that there could be a deficiency in the parallax fit .
this deficiency in the parallax fit can be seen in the residual plot in fig .
[ lc : sc33 ] ; systematic residuals from the parallax model can clearly be seen , and the remaining ` feasible ' constant - acceleration fit is also unable to successfully reproduce this behaviour .
these residuals from the parallax model could imply that this event is exhibiting some weak binary - source signatures ( see , for example , alcock et al .
we performed a preliminary investigation to test this hypothesis , fitting the event with a simple binary - source model .
the basic details of this model can be found in smith et al .
( 2002b ) , but the version used here has two differences from the earlier work :
firstly , it incorporates the parallax motion of the earth ( which , in most cases , should have an effect on the light curve since the duration of this event is well over one year ) ; and secondly , only face - on elliptical orbits are considered .
the best binary - source plus parallax fit that we are able to identify has a much better @xmath26 value ( @xmath50 ) compared to all of the previous models .
however , we find a range of possible binary - source fits with @xmath26 less than 180 .
the single best - fit solution is presented in table [ tb : bin ] and fig .
[ lc : sc33 ] .
this model predicts a very large value for the einstein radius projected into the observer plane ( @xmath51 au ) , which implies that the parallax motion of the earth has little effect on the light curve , i.e. , the majority of the perturbations from the standard model arise due to the binary nature of the source .
it seems that this model is able to fit the systematic residuals from the parallax model , although it predicts peculiar bumps in the light curve ( most notably during the break between observing seasons ) .
we present this model to show that such fits are possible , but we do not attempt a more detailed analysis of this event in this work . if this event is indeed affected by the binary motion of the source then a spectroscopic followup should find the source to be a spectroscopic binary , and this would reduce the number of free parameters in our model .
furthermore , if the projected separation of the two sources can be estimated , then this would enable us to make a direct determination of the lens mass ( han & gould 1997 ) .
the source baseline magnitude ( including blending ) is about @xmath52mag , and so it should be within easy reach of modern large telescopes . [
cols="^,^,^,^,^,^,^",options="header " , ]
in this paper , we have illustrated the effect of acceleration on microlensing light curves .
we studied whether the known parallax microlensing events can be equally fitted by a model with a constant acceleration .
we found that for the shorter marginal parallax microlensing events , the constant acceleration is indeed sufficient .
however , for the longest events ( notably , ogle-1999-bul-32 and ogle-1999-bul-19 ) , the constant acceleration fails to provide a satisfactory model , so their parallax nature seems secure .
it is important to examine whether there is an empirical way of separating ` parallax events ' from ` microlensing events with acceleration ' . for binary lenses
it can be shown that such constant - acceleration perturbations are unlikely to be of importance .
this can be understood by considering three regimes of binary lenses separately : close binaries ( i.e. , those with separation @xmath53 ) , intermediate binaries ( @xmath54 ) , and wide binaries ( @xmath55 ) .
one would expect sufficiently large accelerations for close separation binaries .
however , unless the impact parameter is extremely small ( i.e. the source passes very close to the centre of mass of the lens ) , it is unlikely that the resulting light curve will be differentiable from a standard single - lens light curve ( gaudi & gould 1997 ) . even in the rare cases where such small impact parameters occur
, one would expect the primary deviations to be caused by the binary structure , rather than the acceleration ( for example , the binary - lens event macho-99-blg-47 , albrow et al .
2002 ) . for intermediate binaries where the separation is of the order of the einstein radius
, one would also expect the primary deviations to arise from the binary structure , rather than the acceleration . even in the absence of caustic crossings , the effect of acceleration
should be relatively small compared to the distortion of the light curve due to the other mass ( for example , typical lens geometries - with the source lying in the bulge at a distance of approximately @xmath56 and the lens lying half - way in between - predict @xmath57 , where @xmath58 is the total lens mass ; therefore , for face - on orbits with separations of approximately @xmath59 , a typical orbital period is @xmath60 , which implies that the resulting acceleration should produce only weak perturbations to the light curve compared to the distortion from the companion mass ) . for the case of wide separation binaries , it is clear that the accelerations should be small compared to the earth s acceleration , since binaries with separation much greater than @xmath61 are expected to have orbital periods of longer than 8 years .
therefore , we conclude that the effect of constant accelerations in binary lenses is unlikely to be of importance . in principle , there is a simple test to determine whether an acceleration is due to binary motion of the source or the observer : for the case where the acceleration is induced by the source s binary motion , the source must show ( periodic ) changes in radial velocity , which can be verified even after the event is over . in appendix b , we illustrate the expected radial velocities for the simple case of circular binary orbits .
we find that the expected radial velocity amplitude is of the order of @xmath62 , where @xmath63 is the inclination angle ; there is also some weak dependency on @xmath23 , the transverse velocity of the lens , @xmath7 , and the masses .
the expected orbital period depends on the unknown lens transverse velocity and the inclination angle , but is of the order of @xmath64 years .
this radial velocity can , in principle , be measured . in the right two panels of fig .
[ fig : model ] , we showed that the light curves are symmetric when the acceleration is perpendicular to the source trajectory .
in such cases , the effect of acceleration may be difficult to notice
. this may even be true for some cases for which the acceleration is not exactly perpendicular to the source motion , particularly since the real light curves have significant gaps due to bad weather , etc .
if we force the standard fit to the microlensing light curve , the time - scale obtained from the fit may be different from the true value . to illustrate the effect , we generate artificial light curves using monte carlo simulations .
we adopt @xmath65 days , and a sampling frequency of one observation per day ; the observations lasts from @xmath66 to @xmath67 centred around the peak of the light curve .
the observational errors are assumed to be 0.05 magnitudes at the baseline and scale according to photon poisson noise .
we find that a standard fit incorporating blending is significantly better than a fit without blending .
in other words , a microlensing light curve with constant acceleration can mimic light curves with blending .
interestingly , the fit with blending under - estimates the true @xmath23 by 20% for both acceleration values ( @xmath68 ) shown in the right panels of fig .
[ fig : model ] .
this can have important implications for the calculation of optical depth .
the optical depth , @xmath69 , is estimated from experiments using the following formula ( see , for example , han & gould 1995 ) , = _
i=1^n , where @xmath70 is the total number of monitored stars , @xmath71 is the experiment duration in years , @xmath72 is the detection efficiency , and @xmath73 is the einstein radius crossing time for the @xmath63-th event ( @xmath74 ) . obviously , if the time - scale is underestimated then this will lead to an underestimation in the optical depth , and vice versa .
however , a quantitative analysis requires a detailed knowledge about the binary parameters of the lens and the source , and sampling frequencies in observations .
this is clearly an area that deserves further study .
we thank simon white for an insightful question which prompted the study and ian browne for a careful reading of the paper .
we are indebted to the referee andy gould for a prompt and comprehensive report .
mcs acknowledges receipt of a pparc grant .
this project was supported by the nsf grants ast-9820314 , ast-1206213 , and the nasa grant nag5 - 12212 and funds for proposal # 09518 provided by nasa through a grant from the space telescope science institute , which is operated by the association of universities for research in astronomy , inc .
, under nasa contract nas5 - 26555 .
albrow m. d. et al . , 2002 ,
apj , 572 , 1031 alcock c. et al .
( macho collaboration ) , 1995 , apj , 454 , l125 alcock c. et al .
( macho collaboration ) , 1997 , apj , 491 , l11 alcock c. et al .
( macho collaboration ) , 2001 , apj , 552 , 259 bennett d. p. et al .
, 2002a , , 579 , 639 bennett d. p. , becker a. c. , calitz j. j. , johnson b. r. , laws c. , quinn j. l. , rhie s. h. , sutherland w. , 2002b , preprint ( astro - ph/0207006 ) bennett d. p. et al . , 1997 , baas , 191 , 8303 bond i. a. et al .
( the moa collaboration ) , 2001 , mnras , 327 , 868 gaudi b. s. , gould a. , 1997 , apj , 482 , 83 gould a. , 1992 , , 392 , 442 gould a. , 1994 , , 421 , l75 gould a. , 1995 , , 441 , l21 gould a. , miralda - escude j. , bahcall j.n . , 1994 , apj , 423 , l105 griest k. hu w. , 1992 , apj 397 , 362 han c. , gould a. , 1995 , , 449 , 521 han c. , gould a. , 1997 , , 480 , 196 mao s. , 1999 , , 350 , l19 mao s. et al .
( ogle collaboration ) , 2002 , mnras , 329 , 349 paczyski b. , 1986 , , 304 , 1 paczyski b. , 1997 , astro - ph/9711007 paczyski b. , 1996 , , 34 , 419 refsdal s. , 1966 , , 134 , 315 palanque - delabrouille n. et al .
( eros collaboration ) , 1998 , a&a , 332 , 1 smith m. c. , mao .
s. & woniak p. , 2002a ,
mnras , 332 , 962 smith m. c. et al .
( ogle collaboration ) , 2002b , mnras , 336 , 670 soszyski i. et al .
, 2001 , apj , 552 , 731 udalski a. , szymaski m. , kauny j. , kubiak m. , mateo m. , krzemiski w. , paczyski b. , 1994 , acta astron . , 44 , 227 udalski a. , szymaski m. , kubiak m. , pietrzyski g. , woniak p. , ebru k. , 1997 , aca , 47 , 431 udalski a. , ebru k. , szymaski m. , kubiak m. , pietrzyski g. , soszyski i. , woniak p. , 2000 ,
aca , 50 , 1 woniak p. , paczyski b. , 1997 , apj , 487 , 55 woniak p. , udalski a. , szymaski m. , kubiak m. , pietrzyski g. , soszyski i. , ebru k. , 2001 , acta astron . , 51 , 175
in this appendix we first describe the degeneracy that is found to arise in the constant - acceleration model .
this is followed by a discussion of an equivalent degeneracy in the parallax model , which can occur for short duration events where the earth s acceleration may be approximated as constant . in section 3
we discuss the degeneracy inherent in fitting microlensing light curves with the constant - acceleration model .
it is found that each @xmath26 minimum has two corresponding fits , each with differing values of time - scale , @xmath23 , and the acceleration parameters , @xmath10 and @xmath9 . however , the magnitude of the physical acceleration in units of @xmath28 ( i.e. , @xmath75 ) is the same for both fits , although the direction of this vector ( i.e. , @xmath9 ) differs . in this subsection
we show analytically why this degeneracy occurs . given a combination of parameters ( @xmath76 , @xmath20 , etc )
, the magnitude of the source displacement in the lens plane ( in units of the einstein radius ) at a time @xmath77 can be determined from equations ( [ eq : linear]-[eq : pos ] ) , u^2(t_0 + t ) = ( + ) ^2 + ( + u_0 ) ^2 , where @xmath78 is the magnitude of the physical acceleration in units of the einstein radius .
this equation can be expanded to give , @xmath79 for two degenerate solutions @xmath63 and @xmath80 , the magnitude of the source displacement , @xmath81 and @xmath82 , must be the same for all values of @xmath83 ( since this is required to produce identical light curves ) , i.e. , @xmath84 for all @xmath85 . since this is true for all values of @xmath83 , using equation ( [ aeq : pos ] ) we can equate powers of @xmath83 .
this gives the following four equations , @xmath86 from equations ( [ aeq:1 ] ) and ( [ aeq:4 ] ) we can see that both @xmath87 and @xmath20 must be constant for degenerate solutions ( the sign of @xmath20 is irrelevant due to the symmetry of the acceleration model ; by taking only @xmath88 one still obtains the full set of solutions ) .
also , from equations ( [ aeq:2 ] ) and ( [ aeq:3 ] ) we find that the quantities @xmath89 $ ] and @xmath90 $ ] must also be constant for degenerate solutions .
therefore , the following two equations must hold for _ every _ degenerate solution , @xmath91 where @xmath92 and @xmath93 are constants defined for any given @xmath26 minimum .
the parameter @xmath9 can then be eliminated from these two equations to give the following expression for @xmath23 , [ aeq : te ] f(w ) = ( u_0 c_1)^2 w^3 + ( c_2 ^ 2 - ^2u^2_0 ) w^2 - 2c_2 w + 1 = 0 , where @xmath94 .
we see that the above equation is a cubic equation in @xmath95 .
as the coefficients are all real , the number of real solutions is either one or three .
recall that since @xmath96 , physical solutions to equation ( [ aeq : te ] ) must clearly correspond to positive values of @xmath95 .
as @xmath97 , and @xmath98 , we must have one negative real solution for @xmath95 , which is unphysical and should be discarded .
it is easy to show that the number of positive ( i.e. physical ) solutions for @xmath95 must be either zero or two .
this follows because @xmath98 and @xmath99 .
if @xmath100 is everywhere positive for @xmath101 , then the number of positive solutions is obviously zero .
if , however , @xmath100 is negative at some point @xmath102 , then there must be one positive solution in the region 0 to @xmath103 , and another positive solution in the region @xmath103 to @xmath104 .
this therefore immediately implies that if there is one good constant - acceleration fit ( i.e. one physical solution for @xmath95 ) , then there must be another degenerate fit with the same @xmath32 , just as we have found numerically .
these two fits have identical values of @xmath20 and @xmath105 , but different values of @xmath10 , @xmath9 and @xmath23 .
a consequence of this constant - acceleration degeneracy is that a similar degeneracy should arise for short - duration events that exhibit weak parallax signatures .
this is because the acceleration in the parallax model can be approximated as constant provided that the duration of the event is sufficiently short . in the following subsection
we consider this new degeneracy and provide an example using the event sc6_2563 , which is introduced in section [ sc6 ] .
there are currently well - known degeneracies inherent in parallax microlensing : for example , gould , miralda - escude & bahcall ( 1994 ) discussed the continuous degeneracy that occurs for events with extremely weak parallax signatures where only one component of the relative lens velocity , @xmath106 , is measurable ; another four - fold discrete degeneracy was identified in the case of satellite parallax measurements ( refsdal 1966 ; gould 1994 ) , although it was shown that this problem can be resolved using higher - order effects ( gould 1995 ) .
as we have discussed above , in the case of the weak candidate parallax events , such as sc6_2563 ( see section [ sc6 ] ) , the best constant - acceleration fit mimics the best parallax fit .
we also found that each @xmath26 minima for the constant - acceleration model possesses two degenerate fits , each with identical values for impact parameter , @xmath20 and acceleration magnitude , @xmath105 , but different values of @xmath107 , @xmath9 and @xmath23 .
however , since the above constant - acceleration formalism is rotationally invariant , it is possible to rotate one of these degenerate constant - acceleration fits so that both degenerate fits have their acceleration vectors co - aligned .
therefore , if one of these best - fit constant - acceleration solutions mimics the existing parallax fit , then the degenerate counterpart ( once rotated through the required angle so that the acceleration is also in the same direction as the earth s acceleration ) should likewise mimic an equivalent parallax fit .
since this new parallax fit must be different from the existing one , this implies that there should be two degenerate parallax fits .
obviously , this argument only holds in the regime of constant - acceleration , i.e. , this would not apply to parallax events where the duration is sufficiently long to invalidate the constant - acceleration approximation .
we illustrate this new degeneracy by providing an example using the event sc6_2563 , which was introduced in section [ sc6 ] . to identify any degenerate solutions in the parallax parameter space we fit this event 1000 times , taking random values for the initial parameter guesses . from this
we identify another solution in the parameter space that has a slightly worse @xmath26 value than the previously identified best - fit parallax solution .
both fits are presented in table [ tb : sc6 ] .
if the @xmath26 values are renormalised to enforce the @xmath26 per degree of freedom to be unity for the best - fit parallax model , then the difference in @xmath26 is less than 0.5 . unlike the two degenerate constant - acceleration fits , these parallax fits have different values for the magnitude of the impact parameter , @xmath20 .
however , this is due to the difference in the definition of @xmath20 in the two models .
figure [ fig : traj ] shows the trajectories in the lens plane for the two parallax fits , along with the two degenerate constant - acceleration fits .
note that the @xmath108 trajectory has been rotated so that its acceleration is pointing in the same direction as the @xmath109 trajectory ( the rotation does not affect the light curve ) .
although each parallax trajectory deviates from its corresponding constant - acceleration counterpart , it can be seen that the distance from the origin ( i.e. , the quantity that determines the magnification ; the parameter @xmath110 in eq .
[ eq : mag ] ) is very similar .
as expected , this figure shows that the difference between the constant - acceleration and parallax trajectories only becomes apparent away from the point of closest approach , i.e. when the direction of the earth s acceleration has changed .
this corresponds to the wings of the light curve , but since the magnification is only weak in this region the resulting deviations in the light curve are difficult to detect .
we also investigated the event sc33_4505 , which is introduced in section [ sc33 ] .
this event exhibits more prominent deviations , and the constant - acceleration model is found to provide a significantly worse fit than the parallax model ( see table [ tb : sc33 ] ) .
again , we fit this event 1000 times to try and identify any degeneracy that might occur for the parallax model , but we are unable to find any such degenerate fits .
the next - best parallax fit has @xmath111 , which is more than @xmath112 away from the best - fit value of @xmath113 .
therefore , we conclude that there appears to be no such parallax degeneracy for this event , owing to the fact that the deviations in the light curve are relatively prominent , i.e. , these deviations are unable to be reproduced by a constant - acceleration fit .
here we estimate the expected radial velocity for the case where the acceleration is induced by the binary motion of the source . for simplicity , we consider the case where both orbits are circular .
let us denote the mass of the lensed source star as @xmath114 and that of its binary companion as @xmath115 , the separation between the stars as @xmath116 , and the inclination of the orbital plane as @xmath63 ( @xmath117 implies an edge - on orbit ) . in the centre - of - mass rest frame ,
the maximum radial velocity is given by @xmath118 the physical acceleration of the lensed source due to its companion is given by @xmath119 .
the component perpendicular to the line - of - sight is @xmath120 ; here @xmath121 is the angle of the position of the lensed star from the line where the orbital plane and the plane of sky intersect .
notice that our constant acceleration assumption requires that @xmath121 does not change significantly during the lensing event , i.e. for circular orbits @xmath122 implies that the period of the binary @xmath123 . combining the expression of @xmath124 with the definition of @xmath10 , we have @xmath125 cancelling out @xmath116 from eqs .
( [ eq : a1 ] ) and ( [ eq : a2 ] ) , we obtain the maximum radial velocity @xmath126 where @xmath127 is the lens transverse velocity perpendicular to the observer - source line . numerically , we have @xmath128 notice that the dependence on the masses , @xmath10 and the lens transverse velocity , @xmath7 , are fairly weak
. from kepler s third - law , the period of the binary is given by @xmath129 .
substituting @xmath116 from eq .
( [ eq : a2 ] ) into @xmath130 , we obtain @xmath131 the dependence of period on the unknown lens transverse velocity and the inclination is fairly sensitive . for events that show constant accelerations , the transverse velocity may be fairly low and hence the period can be of the order of years . | to generate the standard microlensing light curve one assumes that the relative motion of the source , the lens , and the observer is linear . in reality , the relative motion is likely to be more complicated due to accelerations of the observer , the lens and the source .
the simplest approximation beyond the linear - motion assumption is to add a constant acceleration .
microlensing light curves due to accelerations can be symmetric or asymmetric depending on the angle between the acceleration and the velocity .
we show that it is possible that some of the previously reported shorter marginal parallax events can be reproduced with constant - acceleration models , while the longer , multi - year parallax events are ill - fitted by such models .
we find that there is a generic degeneracy inherent in constant - acceleration microlensing models .
we also find that there is an equivalent degeneracy in parallax models , which manifests itself in short - duration events .
the importance of this new parallax degeneracy is illustrated with an example , using one of these marginal parallax events .
our new analysis suggests that another of these previously suspected parallax candidate events may be exhibiting some weak binary - source signatures .
if this turns out to be true , spectroscopic observations of the source could determine some parameters in the model and may also constrain or even determine the lens mass .
we also point out that symmetric light curves with constant accelerations can mimic blended light curves , producing misleading einstein - radius crossing time - scales when fitted by the standard ` blended ' microlensing model ; this may have some effect on the estimation of optical depth . #
1= 0u_0 0m_i,0 epsf gravitational lensing - galaxy : bulge - galaxy : centre - binaries : general |
recently , the higher - order two - loop corrections to the lamb shift have been studied rather intensively , both within the @xmath4-expansion ( see @xcite and references therein ) as well as within the nonperturbative ( in @xmath4 ) numerical approach , as described in refs . .
in the current note , we review some recent progress for the so - called @xmath5 coefficient , which is generated by the entire gauge - invariant set of two - loop diagrams , as depicted in fig .
. we also review some very recent progress @xcite regarding the @xmath5 coefficient for general excited hydrogenic states with nonvanishing angular momentum , with a special emphasis on states with @xmath1 symmetry .
feynman diagrams for the two - loop self - energy corrections , separated into subsets @xmath6@xmath7 according to ref .
subset @xmath6 is the pure two - loop self - energy , subset @xmath8 comprises the vacuum - polarization insertion into the virtual - photon line of the one - photon self - energy , subset @xmath9 contains vacuum - polarization corrections to the electron line in the one - photon self - energy , and subset @xmath7 contains remaining vacuum - polarization effects . ]
applications of the recent progress to high - precision spectroscopy are numerous .
as one example of current interest , the status of the @xmath2he@xmath3 lamb shift ( @xmath10 and @xmath11 states ) is summarized , based on the recent analytic and numerical results , and on information about further known contributions to the lamb shift from the literature ( see in particular @xcite and references therein ) .
the two - loop energy shift of an atomic level in a hydrogenlike atomic system reads ( in units with @xmath12 ) @xmath13 where @xmath14 is the electron mass , and @xmath15 is a dimensionless function . in the current section of this article ,
we are primarily concerned with recently obtained @xcite results for the normalized ( or `` weighted '' ) difference @xmath16 of @xmath0 states , whose importance for the determination of fundamental constants has been stressed in refs .
@xcite , and for individual @xmath1 states . for these states and/or combinations of states , the first nonvanishing terms in the semi - analytic expansion of the dimensionless function @xmath17 in powers for @xmath4 and @xmath18 read as follows , @xmath19 + b_{61 } \ln[(z\alpha)^{-2 } ] + b_{60 } \biggr\}.\ ] ] the first index of the @xmath20 coefficients marks the power of @xmath4 , whereas the second corresponds to the power of the logarithm @xmath21 $ ] .
for individual @xmath0 states , we only mention here the existence of a @xmath22 coefficient @xcite , which goes beyond the coefficients listed in ( [ defh ] ) .
for the normalized difference of @xmath0 states , we have @xcite @xmath23 where @xmath24 is euler s constant , and @xmath25 is the logarithmic derivative of the gamma function .
the normalized difference for @xmath26 reads @xcite @xmath27 \nonumber\\[2ex ] & & + \left [ \frac{304}{135 } - \frac{32}{9 } \ , \ln(2 ) \right ] \
, \left ( \frac34 - \frac{1}{n } + \frac{1}{4\,n^2 } + \gamma - \ln(n ) \ ! + \ !
\psi(n ) \right)\,.\end{aligned}\ ] ] the normalized difference of the nonlogarithmic term can be expressed as @xcite @xmath28 where @xmath29 is an additional contribution beyond the @xmath30-dependence of the two - loop bethe logarithm @xmath31 .
the result for @xmath32 is @xcite , @xmath33 - \frac{337043}{129600 } - \frac{94261}{21600n } + \frac{902609}{129600n^2 } \nonumber\\ & + \left ( \frac{4}{3 } - \frac{16}{9n } + \frac{4}{9n^2 } \right ) \ln^2 ( 2 ) % + \left ( -\frac{76}{45 } + \frac{304}{135n } - \frac{76}{135n^2 } \right ) \ln(2 ) % + \left ( - \frac{53}{15 } + \frac{35}{2n } - \frac{419}{30n^2 } \right ) \nonumber\\ % & \times \zeta(2)\ln(2 ) + \left ( \frac{28003}{10800 } - \frac{11}{2n } + \frac{31397}{10800n^2 } \right ) \zeta(2 ) % + \left ( \frac{53}{60 } - \frac{35}{8 n } + \frac{419}{120 n^2 } \right ) \zeta(3 ) \nonumber\\ % & + \left ( \frac{37793}{10800 } + \frac{16}{9}\ln^2(2 ) - \frac{304}{135}\ln ( 2 ) + 8\zeta(2)\ln(2 ) - \frac{13}{3}\zeta(2 ) - 2\zeta(3)\right ) \left [ \gamma + \psi(n ) - \ln(n ) \right ] .\end{aligned}\ ] ] here , @xmath34 is a nonlogarithmic term generated by a dirac-@xmath35 correction to a one - loop bethe logarithm , as calculated in ref . @xcite
of course , @xmath36 is the riemann zeta function . for @xmath1 states ,
we have the known results @xcite @xmath37 % b_{61}(np_{1/2 } ) = & \ ; \frac43\ , n(np ) + \frac{n^2 - 1}{n^2 } \left(\frac{166}{405 } -\frac{8}{27 } \ , \ln 2 \right ) \ , , \nonumber\\[2ex ] % b_{61}(np_{3/2 } ) = & \ ; \frac43\ , n(np ) + \frac{n^2 - 1}{n^2 } \left(\frac{31}{405 } -\frac{8}{27 } \ , \ln 2 \right ) \ , .\end{aligned}\ ] ] the results for the nonlogarithmic terms of @xmath1 can be inferred on the basis of eq .
( 8.1 ) of ref .
@xcite and the two - loop bethe logarithms for @xmath1 states ( see @xcite and table [ table2 ] ) , [ b60p ] @xmath38
n(np ) -\frac{27517}{25920 } - \frac{209}{288n } \nonumber\\[2ex ] % & + \frac{1223}{960 n^2 } + \frac{4}{27 } \frac{n^2 - 1}{n^2 } \ln^2(2 ) - \frac{38}{81 } \frac{n^2 - 1}{n^2 } \ln(2 ) + \left ( \frac{25}{6 } + \frac{3}{2 n } - \frac{9}{2 n^2 } \right ) \zeta(2 ) \ln(2 ) \nonumber\\[2ex ] % & + \left ( -\frac{9151}{10800 } -\frac{1}{4n } + \frac{1009}{1200 n^2 }
\right ) \zeta(2 ) + \left ( -\frac{25}{24 } -\frac{3}{8n } + \frac{9}{8 n^2 } \right ) \zeta(3)\ , , \\[2ex ] % \label{b60p32 } & b_{60}(np_{3/2 } )
= b_l(np ) + \beta_4(np_{3/2 } ) + \beta_5(np_{3/2 } ) + \left [ \frac{38}{45 } - \frac43\ln(2 ) \right ] n(np ) -\frac{73321}{103680 } + \frac{185}{1152n } \nonumber\\[2ex ] & + \frac{8111}{25920n^2 } + \frac{4}{27 } \frac{n^2 - 1}{n^2 } \ln^2(2 ) - \frac{11}{81 } \frac{n^2 - 1}{n^2 } \ln(2 ) + \left ( \frac{299}{80 } - \frac{3}{8 n } - \frac{53}{20 n^2 } \right ) \zeta(2 ) \ln(2 ) \nonumber\\[2ex ] & \quad + \left ( -\frac{24377}{21600 } + \frac{1}{16n } - \frac{3187}{3600 n^2 } \right ) \zeta(2 ) + \left ( -\frac{299}{320 } + \frac{3}{32n } + \frac{53}{80 n^2 } \right ) \zeta(3 ) \,.\end{aligned}\ ] ] in these formulas , @xmath39 and @xmath40 are low - energy spin - dependent contributions , defined in eq .
( 4.21 ) of ref .
@xcite , whose numerical values may be inferred from one - loop calculations @xcite .
the evaluation of the two - loop bethe logarithm for @xmath10 and @xmath11 has been discussed in ref .
@xcite , and for @xmath41@xmath42 in ref .
@xcite . for @xmath10 and @xmath11
, there is no ambiguity in the definition of the bethe logarithm , which can roughly be explained as follows : essentially , the two - loop bethe logarithm results from a renormalized integration over two photon energies
. both of these integrations are free of singularities for @xmath10 and @xmath11 . however , for all higher excited @xmath0 states and all @xmath1 states , one incurs real ( rather than imaginary ) contributions to the energy shift from the product of imaginary contributions due to singularities along both photon integrations ( these are `` squared decay rates '' in the sense of ref .
it is thus necessary to make a clear distinction between the singularity - free , principal - value part @xmath43 and a real part @xmath44 , which is incurred by `` squared '' ( or , more precisely , products of ) imaginary contributions from the pole terms .
we write @xmath45 where @xmath43 is obtained as the nonlogarithmic energy shift stemming from the nonrelativistic self - energy , with all integrations carried out by principal value , and @xmath44 is the corresponding contribution defined in refs .
@xcite , due to squared imaginary parts . for @xmath41@xmath42 states , the above separation is not really essential , because @xmath44 is a numerically marginal contribution as compared to @xmath43 ( see ref .
@xcite ) , and thus @xmath46 to a very good approximation . for @xmath1 states under investigation here ,
the distinction ( [ bldef ] ) , surprisingly , is already important ( see table [ table1 ] ) .
final numerical values of the weighted difference of @xmath5 for @xmath0 states , and for individual @xmath1 states , are summarized in table [ table2 ] .
c@c@c@c@c@c@c@c ' '' '' level & @xmath43 & @xmath44 & @xmath31 & level & @xmath43 & @xmath44 & @xmath31 + @xmath10 & @xmath47 & @xmath48 & @xmath47 & @xmath49 & @xmath49 & @xmath49 & @xmath49 + @xmath11 & @xmath50 & @xmath48 & @xmath50 & @xmath51 & @xmath52 & @xmath53 & @xmath52 + @xmath41 & @xmath54 & @xmath55 & @xmath56 & @xmath57 & @xmath58 & @xmath59 & @xmath60 + @xmath61 & @xmath62 & @xmath63 & @xmath64 & @xmath65 & @xmath66 & @xmath67 & @xmath68 + @xmath69 & @xmath70 & @xmath71 & @xmath72 & @xmath73 & @xmath66 & @xmath74 & @xmath75 + @xmath42 & @xmath76 & @xmath77 & @xmath78 & @xmath79 & @xmath80 & @xmath81 & @xmath82 + c@c@c@c@c@c@c@c ' '' '' level & @xmath83 & level & @xmath84 & level & @xmath85 + @xmath11 & @xmath86 & @xmath87 & @xmath88 & @xmath89 & @xmath90 + @xmath41 & @xmath91 & @xmath92 & @xmath93 & @xmath94 & @xmath52 + @xmath61 & @xmath95 & @xmath96 & @xmath97 & @xmath98 & @xmath58 + @xmath69 & @xmath99 & @xmath100 & @xmath97 & @xmath101 & @xmath58 + @xmath42 & @xmath102 & @xmath103 & @xmath58 & @xmath104 & @xmath105 + l@[email protected] & equation in ref .
@xcite & & + ' '' '' @xmath106 $ ] & eq . ( 59 ) [ part ] & 146724.762 & 18340.595 + ' '' ''
@xmath107 & eq . ( 59 ) [ part ] & -40796.296 & -4725.621 + ' '' '' @xmath108 & eq . ( 59 ) [ part ] & 16.295 & 2.037 + ' '' ''
@xmath109 & eq . ( 59 ) [ part ] & 0.029 & 0.004 + ' '' '' @xmath107 ( muonic vac . pol . ) & eq .
( 63 ) & -0.081 & -0.010 + ' '' '' @xmath107 ( hadronic vac . pol . ) & eq .
( 65 ) & -0.051 & -0.006 + ' '' '' @xmath110 ( se@xmath111vp ) & eq .
( 73 ) & 1827.214 & 228.402 + ' '' '' @xmath112 ( two one - loops ) & eqs .
( 74@xmath11176@xmath11179@xmath11180 ) & 0.492 & 0.061 + ' '' '' @xmath113 ( two - loop vp ) & eq .
( 75 ) & 0.704 & 0.088 + ' '' '' @xmath113 ( two - loop se ) & eq . ( 81 ) & -10.709(1 ) & -1.339 + ' '' '' @xmath113 ( two - loop sum ) & sum of 3 above & -9.513(1 ) & -1.189 + ' '' '' @xmath114 $ ] & eq . ( 84 ) [ part ] & -198.172 & -24.771 + ' '' '' @xmath115 $ ] & eq . ( 84 ) [ part ] & 123.906 & 16.985 + ' '' '' @xmath116 & ref .
@xcite & -82.542 & -10.621 + ' '' '' @xmath117 & ref .
@xcite & -1.685 & -0.276 + ' '' '' @xmath118 & eq . ( 101 ) & 0.157 & 0.020 + ' '' '' @xmath119 $ ] & ref .
@xcite & -1.153 & -0.144 + ' '' '' @xmath120 $ ] & ref .
@xcite & -0.213 & 0.037 + ' '' '' @xmath121 $ ] & ref .
@xcite & 2.666 & 0.279 + ' '' ''
@xmath122 & refs .
@xcite & -0.607(211 ) & -0.064(26 ) + ' '' '' @xmath123 & eq . ( 136 ) & 17.786 & 2.547 + ' '' '' @xmath124 & eq . ( 144 ) & -0.119 & -0.015 + ' '' '' @xmath125 & eq . ( 147 ) & -0.010 & -0.001 + ' '' '' @xmath126 & eq .
( 151)@xmath111eq .
( 46 ) of ref .
@xcite & -0.112 & -0.014 + ' '' '' @xmath127 m / m $ ] & eq . ( 155 ) & 0.018 & 0.002 + ' '' ''
@xmath128 & eq . ( 15 ) of ref .
@xcite & -0.053 & -0.007 + ' '' '' @xmath129 & approx . , eq .
( 152 ) & 0.019(19 ) & 0.002(2 ) + ' '' '' nucl .
size [ rel . , 1.680(5)fm ] & eq .
( [ resfs ] ) of this work & 70.865(422 ) & 8.860(53 ) + ' '' '' nucl . size [ rel . , 1.673(1)fm ] & eq .
( [ resfs ] ) of this work & 70.275(84 ) & 8.786(11 ) + ' '' '' sum [ 1.680(5)fm ] & & 107693.112(472 ) & 13837.031(59 ) + ' '' '' sum [ 1.673(1)fm ] & & 107692.522(228 ) & 13836.957(29 ) + l@[email protected] & equation in ref .
@xcite & & + ' '' '' @xmath106 $ ] & eq . ( 60 ) [ part ] & 0.000 & 0.000 + ' '' ''
@xmath107 & eq . ( 60 ) [ part ] & -206.095 & 200.725 + ' '' ''
@xmath108 & eq .
( 60 ) [ part ] & 0.414 & -0.207 + ' '' '' @xmath109 & eq .
( 60 ) [ part ] & -0.003 & 0.002 + ' '' '' @xmath107 ( muonic vac . pol . ) & ref .
@xcite & 0.000 & 0.000 + ' '' '' @xmath107 ( hadronic vac .
pol . ) & ref .
@xcite & 0.000 & 0.000 + ' '' '' @xmath110 ( se@xmath111vp ) & eq . ( 73 ) & 0.000 & 0.000 + ' '' '' @xmath112 ( two one - loops ) & eqs .
( 74@xmath11176@xmath11179@xmath11180 ) & 0.000 & 0.000 + ' '' '' @xmath113 ( two - loop vp ) & eq . ( 75 ) & 0.000 & 0.000 + ' '' '' @xmath113 ( two - loop se ) & eq .
( 81 ) & 0.000 & 0.000 + ' '' '' @xmath113 ( two - loop sum ) & sum of 3 above & 0.000 & 0.000 + ' '' '' @xmath114 $ ] & eq .
( 86 ) [ part ] & 0.000 & 0.000 + ' '' '' @xmath115 $ ] & eq . ( 86 ) & 1.677 & 0.944 + ' '' ''
@xmath116 & ref .
@xcite & -0.329 & -0.163 + ' '' '' @xmath117 & ref .
@xcite & -0.022 & -0.005 + ' '' '' @xmath118 & eq . ( 101 ) & 0.000 & 0.000 + ' '' '' @xmath119 $ ] & ref .
@xcite & 0.000 & 0.000 + ' '' '' @xmath120 $ ] & ref .
@xcite & 0.006 & 0.006 + ' '' '' @xmath121 $ ] & ref .
@xcite & 0.001 & -0.001 + ' '' '' @xmath122 & refs .
@xcite & -0.001 & -0.001 + ' '' '' @xmath123 & eq . ( 136 ) & -0.138 & -0.138 + ' '' '' @xmath124 & eq . ( 145 ) & 0.007 & 0.007 + ' '' ''
@xmath125 & eq . ( 147 ) & 0.000 & 0.000 + ' '' ''
@xmath126 & eq . ( 151 ) & 0.000 & 0.000 + ' '' ''
@xmath127 m / m $ ] & eq .
( 155 ) & 0.000 & 0.000 + ' '' ''
@xmath128 & e.g. , ref .
@xcite & 0.002 & -0.001 + ' '' '' @xmath129 & approx . ,
( 152 ) & 0.000 & 0.000 + ' '' '' nucl .
size & eq .
( [ resfs ] ) of this work & 0.000 & 0.000 + ' '' '' sum & & -204.481 & 201.168 +
in the current section ( we keep units with @xmath12 ) , we would like to use the results described above , in order to infer the current theoretical status of the lamb shift of @xmath10 and @xmath11 in the @xmath130he@xmath3 ion .
before we start our actual discussion , however , we should remember that an ideal way to carry out a related calculation would involve a full - featured least - squares adjustment according to ref .
@xcite , which includes all available data from relevant high - precision experiments ( see @xcite ) and which , in principle , allows for a deduction of the nuclear charge radius .
in order to infer an approximate theoretical prediction , though , one has to use a charge radius obtained from other sources , and we intend to follow this different route in the current work .
we partly base our evaluation on refs .
@xcite and choose a format as in table 1 of ref .
@xcite . in the evaluations described in tables [ contribsfors ] and [ contribsforp ] ,
the 2002 codata values of the fundamental constants @xcite were used . for the lamb shift @xmath131
, we use the implicit definition @xcite @xmath132 - \frac{m_{\rm r}^2}{2 ( m_{\rm e } + m ) } \left [ f(n , j ) - 1 \right]^2 + {
\cal l } + e_{\rm hfs}\,.\ ] ] here , @xmath133 is the energy level of the bound two - body system under investigation , and @xmath134 is the dimensionless dirac energy .
e.g. , we have @xmath135 , and @xmath136 for the @xmath10 and @xmath11 states , respectively .
furthermore , @xmath137 is the reduced mass of the system , @xmath138 is the nuclear mass , and @xmath139 is the energy shift due to hyperfine effects , which are absent for the spinless @xmath140 nucleus . in order to avoid confusion , we would like to include a few clarifying words regarding specific entries in tables [ contribsfors ] and [ contribsforp ] . in general , we have added the factor @xmath141 to all scales for the contributions listed in tables [ contribsfors ] and [ contribsforp ] . giving all contributions with an overall scaling of @xmath142 , in contrast to @xmath143 , which had been used in ref .
@xcite . regarding the contribution of order @xmath144 $ ] in the first row of table [ contribsfors ] , it is worthwhile to note that this term represents the leading logarithm of the lamb shift , given by @xmath145 \ , \left ( \frac { m_{\rm r } } { m_{\rm e } } \right)^3 \,.\ ] ] here , @xmath137 is the reduced mass of the system , given by @xmath146 , where @xmath147 is the mass of the nucleus .
the reduced - mass dependence of the argument of the logarithm itself , @xmath21 \to \ln[(z\alpha)^{-2 } \ , m_{\rm e}/m_{\rm r } ] = \ln[(z\alpha)^{-2 } ] + \ln[m_{\rm e}/m_{\rm r } ] $ ] , is being included here and in ref .
@xcite into the nonlogarithmic term of order @xmath107 .
one might wonder why there is a theoretical uncertainty associated to this contribution at all .
the reason is that the most accurate theoretical value for the quantity ( [ q ] ) is obtained by expressing it in terms of the 2002 codata rydberg constant , which has a relative uncertainty of @xmath148 , and the 2002 codata fine - structure constant , which has a relative uncertainity of @xmath149 .
the latter is responsible for the small theoretical uncertainty of the leading logarithmic contribution to the lamb shift of the @xmath10 level .
the term of order @xmath150 contains both contributions from the self - energy and the vacuum polarization , as indicated by the explanatory note `` se@xmath111vp . ''
the term of order @xmath127 $ ] corresponds to the self - energy coefficient @xmath151 , as given e.g. in ref .
@xcite , and the indicated term of order @xmath152 $ ] is the sum of a self - energy and a vacuum - polarization contribution in this order .
note that the latter distinction differs from the one used in ref .
@xcite , where the term of order @xmath127 $ ] denotes the sum of a double logarithmic , and a single logarithmic _ self - energy _ contribution , and the term of order @xmath152 $ ] was reserved exclusively for the _ vacuum - polarization _ contribution in this order .
the values of the self - energy remainder @xmath153 for @xmath0 and @xmath1 states are listed in @xcite .
the vacuum polarization remainder function @xmath154 is taken from @xcite and corresponds exclusively to the uehling part of the one - loop vacuum polarization .
it might be worthwhile to point out that at the current level of accuracy , it is entirely sufficient to consider the vacuum - polarization higher - order remainder for @xmath1 states via the formula @xmath155 where the analytic coefficients read [ resvp ] @xmath156 % & a^{\rm vp}_{70}(np_{1/2 } ) = \frac{41 \pi}{2304}\ , \frac{n^2 - 1}{n^2}\ , , \qquad a^{\rm vp}_{70}(np_{3/2 } ) = \frac{7 \pi}{768}\ , \frac{n^2 - 1}{n^2}\,.\end{aligned}\ ] ] these have been obtained in refs .
@xcite for general principal quantum number @xmath30 .
for the @xmath157-term for @xmath0 states of order @xmath120 $ ] , the result from ref .
@xcite was used , which supersedes the estimate given in eq .
( 101 ) of ref . @xcite . for the analytic @xmath26-term of order
@xmath158 $ ] , the result given in ref .
@xcite provides the most recent value . for the @xmath5 coefficient corresponding to the nonlogarithmic term of order @xmath159 for the ground state , two mutually contradictory results of @xmath160 ( ref .
@xcite ) and @xmath161 ( ref .
@xcite ) have been reported .
the latter is based on an extrapolation of an all - order ( in @xmath4 ) numerical calculation .
note , however , that there is a known missing piece in the analytic result reported in ref .
@xcite , which is currently under study and which will need to be evaluated before final conclusions can be drawn .
m. eides @xcite therefore suggested that a valid interim way of estimating the uncertainty of @xmath5 would consist in taking the arithmetic mean of these two results , and taking the half difference as an estimate for the theoretical uncertainty .
this uncertainty would comprise all higher - order analytic terms , as it involves a comparison to a nonperturbative ( in @xmath4 ) calculation . for the @xmath11 state
, we can use the result for @xmath10 and add the weighted difference listed in table [ table2 ] . for @xmath1 states , the results reported in sec .
[ twoloop ] of this paper ( see also @xcite ) provide enough information to eliminate all theoretical uncertainty at the current level of accuracy .
finally , let us remark that a term of order @xmath162 $ ] could be estimated in principle on the basis of taking a `` local '' lamb - shift potential that corresponds to the self - energy part of @xmath163 , namely , @xmath164\ , \frac{\pi \delta^3(r)}{m_{\rm e}^2},\ ] ] taken as an input for a dirac-@xmath35 induced correction to the one - loop self - energy .
the result of this approach could alternatively be used as an uncertainty estimate for all the higher - order terms .
this procedure leads to the estimate @xmath165\,,\ ] ] for @xmath10 , this leads to a value of @xmath166 mhz for the higher - order two - loop remainder .
however , the uncertainty due to the @xmath167-contribution is already contained in the uncertainty of the remainder term of order @xmath159 , because in determining the uncertainty of @xmath5 , a comparison was made to a nonperturbative numerical calculation for higher nuclear charge numbers .
the latter necessarily contains all contributions from the @xmath167 term and all higher - order remainders .
the above result of @xmath166 mhz therefore likely overestimates the uncertainty due to the higher - order two - loop remainder and is mentioned here only for illustrative purposes .
concerning radiative - recoil corrections , we note that the discrepancy between @xcite and @xcite concerning radiative insertions into the electron line in the @xmath168 radiative recoil correction was resolved in @xcite .
the analytical result from that work was used . according to eq .
( 46 ) of ref .
@xcite , the coefficient multiplying the non - vacuum - polarization part of order correction of order @xmath168 is @xmath169 .
the nuclear spin in @xmath2he@xmath3 is different as compared to atomic hydrogen .
the former is a spin-@xmath170spin - zero system , whereas the latter is a spin-@xmath170spin-@xmath170 system .
recoil corrections of first order in the mass ratio are unaffected by the different spin of the nucleus as compared to hydrogen .
however , recoil terms of order @xmath171 are nuclear spin - dependent . without carrying out a detailed analysis
, we approximately calculate the nuclear self - energy effects of order @xmath171 by leaving out the pauli form - factor correction from eq .
( 153 ) of ref .
@xcite , which is certainly absent for a spinless nucleus , and we conservatively take the dirac form factor contribution as an uncertainty , while we note that a more detailed analysis would be of interest and currently lacking in the literature . finally , we add the nuclear - spin dependent correction listed in eq .
( 15 ) of ref .
@xcite , @xmath172 which is of second order in the mass ratio , for the spin-@xmath170spin - zero system under investigation .
this term is connected to the absence of the zitterbewegung term in the breit hamiltonian for a spinless nucleus . for @xmath1 states ,
we also add terms of order @xmath173 , which do not depend on the zitterbewegung term and are given , e.g. , in ref .
@xcite . concerning the nuclear - size correction
, we would like to mention that a full integration of a nuclear potential with a fully relativistic wave function ( e.g. , within a hard - sphere approximation ) turns out to be quite essential to obtain reliable values for this correction .
we have carried out such an integration in the current investigation with the full dirac wave functions and obtain results in agreement with ref .
results obtained with two different values for the two different root - mean - square charge radii of @xmath174 @xcite and @xmath175 @xcite are given in table [ contribsfors ] .
note that the former charge radius of @xmath174 has been questioned ( see e.g. @xcite ) .
the uncertainty due to the shape of the nuclear charge distribution can be estimated to be much smaller than the uncertainty due to the nuclear size , based on experience with highly charged ions @xcite .
for the @xmath2he - nucleus , a spherically symmetric model is well justified ( closed shell , spin @xmath176 ) . the nuclear - size correction @xmath177 and @xmath178 , for low nuclear charge numbers , can be approximated very well by the first few terms of an expansion in the two small parameters @xmath4 and @xmath179 , with the result [ resfs ] @xmath180 & \left .
+ ( z\alpha)^2 \left\ { -\frac25\ , \left [ \ln\left ( \frac{2 \ , z\,\alpha\,m\ , r}{n } \right ) + 2 \gamma + \psi(n ) \right ] + \frac{227}{150 } + \frac{2}{5\,n } - \frac{9}{10 n^2 } \right\ } \right ] \ , , \\[2ex ] \label{resfsb } % \delta e_{\rm fs}(np_j ) = & \frac{1}{10 } \ , \frac{(z\alpha)^6 m^3 r^2}{n^3 } \ , \frac{n^2 - 1}{n^2 } \ , \delta_{j,1/2}\,.\end{aligned}\ ] ] here , @xmath24 is euler s constant , @xmath181 is the logarithmic derivative of the gamma function , and @xmath182 is the radius of the nucleus in a hard - sphere model , which is related to the root - mean - square radius @xmath183 by the following formula [ see eq . ( 7 ) of ref .
@xcite ] , @xmath184 for the @xmath87 state , we obtain an upward finite nuclear - size energy shift of @xmath185 which is barely significant on the khz level ( see table [ contribsforp ] ) . the results in eqs .
( [ resfs ] ) and ( [ r0corrected ] ) have been obtained in the approximation of an infinitely heavy nucleus , and with exact dirac wave functions for a point nucleus .
both of these approximations should be valid for @xmath186 .
in addition , it should be noted that both the results given in eqs .
( [ resfs ] ) and ( [ r0corrected ] ) are in excellent numerical agreement with a full numerical integration of the finite - size potential with dirac wave functions . the linear correction term in @xmath182 , i.e. the term @xmath187 in eq .
( [ resfsa ] ) , is a consequence of the exponential factor @xmath188 in the wave function , which should not be ignored in the evaluation of the finite - size effect , although this effect is primarily sensitive to the probability density at the origin ( at the nucleus ) .
any further effects that influence the finite - size effect like nuclear polarization are here absorbed into the uncertainty of the nuclear radius ( see , e.g. , ref .
@xcite for an illustrative discussion of some of the further aspects that are relevant to the finite - size effect ) .
in sec . [ twoloop ] , we summarize recent theoretical results for the higher - order two - loop binding corrections to the lamb shift .
these results are used , in sec .
[ helamb ] , to infer updated values for the lamb shift of low - lying states of the @xmath2he@xmath3 ion .
some of the analytic coefficients used in the evaluation are given in eqs .
( [ resvp ] ) and ( [ resfs ] ) .
the analytic expansion of the nuclear finite - size correction ( [ resfs ] ) might be useful in other contexts .
the recent progress in the field has allowed for an improvement of the theory beyond the limits set by the leading - order effects , and for some of the most accurate predictions in all of theoretical physics .
in particular , we reemphasize that for @xmath1 states , the results reported in sec .
[ twoloop ] of this paper ( see also @xcite ) provide sufficient information to eliminate all theoretical uncertainty at the khz level for the lamb shift in the @xmath2he@xmath3 ion .
helpful conversations with krzysztof pachucki , vladimir m. shabaev , vladimir a. yerokhin and peter j. mohr are gratefully acknowledged .
u.d.j . acknowledges support from the deutsche forschungsgemeinschaft ( heisenberg program ) . | we report on recent progress in the treatment of two - loop binding corrections to the lamb shift , with a special emphasis on @xmath0 and @xmath1 states .
we use these and other results in order to infer an updated theoretical value of the lamb shift in @xmath2he@xmath3 .
+ + pacs nos . : 12.20.ds , 31.30.jv , 06.20.jr , 31.15.-p [ 1]d .. #1 nous examinons le progrs rcent concernant le traitement des corrections lordre lev des diagrammes deux boucles contribuants aux dplacement de lamb , spcialement en ce qui concerne les tats @xmath0 et @xmath1 .
par consquent , on dduit de nouvelles prcises valeurs thoriques pour le dplacement de lamb en @xmath2he@xmath3 . |
the groups @xmath8 may be considered as a key to the study of a larger class of groups , the _ kleinian _ groups , which date back to work of henri poincar @xcite .
in fact , each non - co - compact arithmetic kleinian group is commensurable with some group @xcite .
a wealth of information on the groups can be found in the monographs @xcite , @xcite , @xcite . in the literature ,
there has been so far no example of @xmath4-torsion in the integral homology of the full bianchi groups , for @xmath4 a prime greater than the order of elements of finite order in the bianchi group ( a recent survey of relevant calculations has been given in @xcite ) .
in fact , the numerical studies that have been made so far , were carried out in the range where the quotient space of hyperbolic @xmath9-space @xmath10 by the bianchi group is often homotopy equivalent to a wedge sum of @xmath11-spheres , @xmath11-tori and mbius bands @xcite .
we make use of serre s decomposition @xcite of the homology group @xmath12 into the direct sum of the free abelian group with one generator for each element of the class group of @xmath2 and the group @xmath13 computed in figures [ small absolute values ] and [ greater absolute values ] .
the first compuations of @xmath14 by swan @xcite were on a range of bianchi groups with vanishing cusp - complementary homology @xmath13 .
the first example where @xmath13 is non - zero , occurred in an unpublished calculation of .
swan s manual computations of group presentations have been extended on the computer by riley @xcite ; and later vogtmann @xcite and scheutzow @xcite systematically computed @xmath15 for a large range of bianchi groups .
but they were still in in the range where @xmath16 admits no homological torsion .
aranes @xcite has computed cell complexes for the bianchi groups for all @xmath17 , and yasaki @xcite has obtained @xmath18-cell complexes ( with the vorono model ) for the same range as well as all cases where @xmath19 is of class number @xmath20 or @xmath11 .
this includes two cases , @xmath21 and @xmath22 , where some @xmath11-torsion appears in @xmath13 , but the latter two authors have not yet provided homology computations . when the absolute value of the discriminant gets greater , torsion in the integral homology of the quotient space appears ( see figure [ greater absolute values ] ) at prime numbers as high as for instance @xmath23 at the discriminant @xmath7 , whereas the order of elements of finite order in @xmath3 is at most @xmath5
. a growth of the torsion in the abelianization of the bianchi groups with respect to the covolume can be observed , which is in concordance with the predictions of @xcite .
we can also observe that the occurring torsion subgroups are quite likely to occur as squares , but this is no general principle , because the discriminant @xmath24 produces a counterexample to this phenomenon . in order to obtain the results of figures [ small absolute values ] and [ greater absolute values ] , in section [ realization ] we fill out swan s concept @xcite and elaborate algorithms to compute a fundamental polyhedron for the action of the bianchi groups on hyperbolic @xmath9-space .
other algorithms based on the same concept have independently been implemented by cremona @xcite for the five cases where @xmath2 is euclidean , and by his students whitley @xcite for the non - euclidean principal ideal domain cases , bygott @xcite for a case of class number 2 and lingham ( @xcite , used in @xcite ) for some cases of class number 3 ; and finally arans @xcite for arbitrary class numbers .
the algorithms presented in subsection [ realization ] come with an implementation @xcite for all bianchi groups ; and we make explicit use of the cell complexes it produces .
the provided implementation @xcite has been validated by the project plume of the cnrs , and is subject to the certificate c3i of the genci and the cpu .
other results obtained with the employed implementation are described in @xcite and @xcite . on the computing clusters of the weizmann institute of science ,
this implementation has been applied to establish a database of cell complexes for over @xmath25 bianchi groups , using over fifty processor - months .
this database includes all the cases of ideal class numbers @xmath9 and @xmath26 , most of the cases of ideal class number @xmath27 and all of the cases of discriminant absolute value bounded by the number @xmath28 .
a computational advantage is the shortcut that we obtain in section [ connection ] by linking the borel
serre compactification of the quotient space with flge s compactification in a long exact sequence , based on the recent paper @xcite .
flge s compactification admits a computationally easier cell structure , and we can explicitly calculate the equivariant leray serre spectral sequence associated to it . in section [ specseq ] ,
we describe how to assemble the homology of the borel
serre compactified quotient space and the farrell cohomology of a bianchi group to its full group homology with trivial @xmath29coefficients . here ,
we divide by the center of @xmath3 , consisting of plus and minus the identity matrix , yielding @xmath30 . as the center of @xmath3 is the kernel of its action on hyperbolic @xmath9-space
, this does not change the quotient space . and for @xmath31 , general formulae for its farrell cohomology have been given @xcite ( based on @xcite ) .
we print the isomorphism types of @xmath13 that were obtained in figures [ small absolute values ] and [ greater absolute values ] .
the homology group @xmath12 is a direct sum of the former and the free abelian group with rank the cardinality of the class group of @xmath2 , which we also print .
there is an inclusion of @xmath12 into the group homology @xmath32 ; and the latter group homology is a quotient of the direct sum of @xmath12 and the farrell supplement that has been computed and printed in a separate column . in section [ the bianchi fundamental polyhedron ]
, we define the bianchi fundamental polyhedron , which induces our cell structure on @xmath16 .
we use it in section [ floege cellular complex ] to obtain the flge cellular complex , which we connect in section [ connection ] to the borel
serre compactification of @xmath16
. then we proceed to @xmath33 in section [ specseq ] , describe swan s concept in section [ swan s concept ] and its realization in section [ realization ] .
let @xmath1 be a squarefree positive integer and consider the imaginary quadratic number field @xmath0 with ring of integers @xmath35 , which we also just denote by @xmath36 .
consider the familiar action by fractional linear transformations ( we give an explicit formula for it in lemma [ operationformula ] ) of the group on hyperbolic three - space , for which we will use the upper - half space model @xmath10 . as a set ,
@xmath37 the bianchi humbert theory @xcite , @xcite gives a fundamental domain for this action . we will start by giving a geometric description of it , and the arguments why it is a fundamental domain .
+ a pair of elements @xmath38 is called _ unimodular _ if the ideal sum @xmath39 equals @xmath36 .
the boundary of @xmath10 is the riemann sphere @xmath40 ( as a set ) , which contains the complex plane @xmath41 .
the totally geodesic surfaces in @xmath10 are the euclidean vertical planes ( we define _ vertical _ as orthogonal to the complex plane ) and the euclidean hemispheres centred on the complex plane .
[ hemispheres ] given a unimodular pair @xmath42 , @xmath43 with @xmath44 , let @xmath45 denote the hemisphere given by the equation @xmath46 .
this hemisphere has centre @xmath47 on the complex plane @xmath41 , and radius @xmath48 .
[ b ] let + @xmath49 : the inequality @xmath50 is fulfilled for all unimodular pairs @xmath42 , @xmath43 with @xmath44 @xmath51 .
then @xmath52 is the set of points in @xmath10 which lie above or on all hemispheres @xmath53 .
[ gammabequalsh ] the set @xmath52 contains representatives for all the orbits of points under the action of @xmath54 on @xmath10 .
the action extends continuously to the boundary , which is a riemann sphere .
+ in @xmath55 , consider the stabiliser subgroup @xmath56 of the point @xmath57 . in the cases @xmath58 and @xmath59 , the latter group contains some rotation matrices like @xmath60 , which we want to exclude .
these two cases have been treated in @xcite , @xcite and others , and we assume @xmath61 , @xmath62 throughout the remainder of this article
. then , @xmath63 which performs translations by the elements of @xmath19 with respect to the euclidean geometry of the upper - half space @xmath10 .
+ [ fundamentalrectangle ] a fundamental domain for @xmath56 in the complex plane ( as a subset of @xmath64 ) is given by the rectangle @xmath65 and a fundamental domain for @xmath56 in @xmath10 is given by @xmath66 we define the _ bianchi fundamental polyhedron _ as @xmath67 it is a polyhedron in hyperbolic space up to the missing vertex @xmath68 , and up to missing vertices at the singular points if @xmath19 is not a principal ideal domain ( see subsection [ singular ] ) .
as lemma [ gammabequalsh ] states @xmath69 , and as yields @xmath70 , we have @xmath71 .
we observe the following notion of strictness of the fundamental domain : the interior of the bianchi fundamental polyhedron contains no two points which are identified by @xmath72 . + swan proves the following theorem , which implies that the boundary of the bianchi fundamental polyhedron consists of finitely many cells .
[ finitenumber ] there is only a finite number of unimodular pairs @xmath73 such that the intersection of @xmath53 with the bianchi fundamental polyhedron is non - empty .
swan further proves a corollary , from which it can be deduced that the action of @xmath72 on @xmath10 is properly discontinuous .
in order to obtain a cell complex with compact quotient space , we proceed in the following way due to flge @xcite . the boundary of @xmath10 is the riemann sphere @xmath64 , which , as a topological space , is made up of the complex plane @xmath41 compactified with the cusp @xmath68 .
the totally geodesic surfaces in @xmath10 are the euclidean vertical planes ( we define _ vertical _ as orthogonal to the complex plane ) and the euclidean hemispheres centred on the complex plane .
the action of the bianchi groups extends continuously to the consider the cellular structure on @xmath10 induced by the @xmath72-images of the bianchi fundamental polyhedron .
the cellular closure of this cell complex in @xmath74 consists of and the @xmath3orbit of a cusp @xmath75 in @xmath76 corresponds to the ideal class @xmath77 $ ] of @xmath2 .
it is well - known that this does not depend on the choice of the representative @xmath75 .
we extend our cell complex to a cell complex @xmath78 by joining to it , in the case that @xmath2 is not a principal ideal domain , the @xmath3orbits of the cusps @xmath75 for which the ideal @xmath73 is not principal . at these cusps
, we equip @xmath78 with the `` horoball topology '' described in @xcite .
this simply means that the set of cusps , which is discrete in , is located at the hyperbolic extremities of @xmath78 : no neighbourhood of a cusp , except the whole @xmath78 , contains any other cusp .
we retract @xmath78 in the following , @xmath3equivariant , way . on the bianchi fundamental polyhedron ,
the retraction is given by the vertical projection ( away from the cusp @xmath68 ) onto its facets which are closed in @xmath74 .
the latter are the facets which do not touch the cusp @xmath68 , and are the bottom facets with respect to our vertical direction .
the retraction is continued on @xmath10 by the group action .
it is proven in @xcite that this retraction is continuous .
we call the retract of @xmath78 the _ flge cellular complex _ and denote it by @xmath79 .
so in the principal ideal domain cases , @xmath79 is a retract of the original cellular structure on @xmath10 , obtained by contracting the bianchi fundamental polyhedron onto its cells which do not touch the boundary of @xmath10 . in @xcite , it is checked that the flge cellular complex is contractible .
let @xmath72 be a bianchi group with @xmath19 admitting as only units @xmath80 , i.e. we suppose @xmath19 not to be the gaussian or eisenstein integers .
in the latter two cases , the problem of the singular cusps treated here does not occur anyway .
let @xmath81 be the torus attached at the cusp @xmath82 of @xmath72 , and let @xmath83 and @xmath84 denote the cycles generating @xmath85 .
let @xmath86 be the bianchi fundamental polyhedron of @xmath72 .
write `` hyp .
cells '' for cells in the interior of hyperbolic space .
consider the short exact sequence of chain complexes obtained from collapsing the singular tori , \(m ) [ matrix of math nodes , row sep=2em , column sep=1.4em , text height=2.5ex , text depth=0.95ex ] 0 & 0 & p & p & 0 + 0 & _
s^singular _ s & _
c^any cusp _
c 2-cells & _ 2-cells & 0 + 0 & _ s^singular x_s , y_s & _ c^any cusp x_c , y_c & x _ , y _ & 0 + 0 & 0 & _ c^any cusp c & _ c^any cusp
+ ; ( m-1 - 1 ) edge ( m-1 - 2 ) ( m-1 - 2 ) edge ( m-2 - 2 ) ( m-1 - 2 ) edge ( m-1 - 3 ) ( m-1 - 3 ) edge node[descr , xshift=1.4ex ] @xmath87(m-2 - 3 ) ( m-1 - 3 ) edge ( m-1 - 4 ) ( m-1 - 4 ) edge node[descr , xshift=1.4ex ] @xmath88 ( m-2 - 4 ) ( m-1 - 4 ) edge ( m-1 - 5 ) ( m-2 - 1 ) edge ( m-2 - 2 ) ( m-2 - 2 ) edge node[descr , yshift=1.2ex ] @xmath89 ( m-2 - 3 ) ( m-2 - 3 ) edge ( m-2 - 4 ) ( m-2 - 4 ) edge ( m-2 - 5 ) ( m-2 - 2 ) edge node[descr , xshift=1.2ex ] @xmath90 ( m-3 - 2 ) ( m-2 - 3 ) edge node[descr , xshift=3.0ex ] @xmath91 ( m-3 - 3 ) ( m-2 - 4 ) edge node[descr , xshift=3.0ex ] @xmath92 ( m-3 - 4 ) ( m-3 - 1 ) edge ( m-3 - 2 ) ( m-3 - 2 ) edge node[descr , yshift=1.4ex ] @xmath89 ( m-3 - 3 ) ( m-3 - 3 ) edge ( m-3 - 4 ) ( m-3 - 4 ) edge ( m-3 - 5 ) ( m-3 - 2 ) edge ( m-4 - 2 ) ( m-3 - 3 ) edge node[descr , xshift=3.0ex ] @xmath93 ( m-4 - 3 ) ( m-3 - 4 ) edge node[descr , xshift=3.0ex ] @xmath93 ( m-4 - 4 ) ( m-4 - 1 ) edge ( m-4 - 2 ) ( m-4 - 2 ) edge ( m-4 - 3 ) ( m-4 - 3 ) edge ( m-4 - 4 ) ( m-4 - 4 ) edge ( m-4 - 5 ) ; poincar s theorem on fundamental polyhedra tells us that @xmath94 , and hence @xmath95 . from @xcite , we see that for every cusp @xmath96 , there is a chain of hyperbolic @xmath11-cells that we denote by @xmath97 and which is mapped to the cycle @xmath98 by @xmath99 . and
furthermore , @xmath100 is in the cokernel of @xmath99 ( of course , this holds up to the appropriate permutation of the labels @xmath98 and @xmath100 ) .
this implies that @xmath101 and @xmath102 is in the cokernel of @xmath103 .
as the quotient space is path - wise connected , the cokernel of @xmath104 is isomorphic to @xmath29 .
the above information tells us that the long exact sequence induced on integral homology by the map @xmath89 concentrates in \(m ) [ matrix of math nodes , row sep=1em , column sep=2.5em , text height=1.7ex , text depth=0.25ex ] 0 & _ s^singular _ s & ( _ c^any cusp _
c ) /__c _ c _ 2^ & _ 2^ _ s ch(x_s ) + & _ s x_s , y_s & _ c y_c _ 1^ & _ 1^ y _ + & 0 & z & z & 0 , + ; ( m-1 - 1 ) edge ( m-1 - 2 ) ( m-1 - 2 ) edge node[descr , yshift=1.2ex ] @xmath105 ( m-1 - 3 ) ( m-1 - 3 ) edge ( m-1 - 4 ) ( m-1 - 4 ) edge[out=355,in=175 ] ( m-2 - 2 ) ( m-2 - 2 ) edge node[descr , yshift=1.15ex ] @xmath106 ( m-2 - 3 ) ( m-2 - 3 ) edge ( m-2 - 4 ) ( m-2 - 4 ) edge[out=355,in=175 ] ( m-3 - 2 ) ( m-3 - 2 ) edge ( m-3 - 3 ) ( m-3 - 3 ) edge ( m-3 - 4 ) ( m-3 - 4 ) edge ( m-3 - 5 ) ; where the maps without labels are the obvious restriction maps making the sequence exact ; and where @xmath107 and @xmath108 are generated by cycles from the interior of @xmath16 . note that @xmath109 is non - naturally isomorphic to @xmath110 , namely collapsing a torus @xmath111 moves its @xmath11-cycle into a bubble @xmath112 emerging adjacent to the singular cusp @xmath113 in the flge complex .
let @xmath31 , and let @xmath79 be the flge cellular complex of section [ floege cellular complex ] , the cell structure of which we subdivide until the cells are fixed pointwise by their stabilisers .
we describe now how to assemble the homology of the borel serre compactified quotient space ( issue of the previous section ) and the farrell cohomology of @xmath72 , for which general formulae have been given in @xcite ( based on @xcite ) , to the full group homology of @xmath72 with trivial @xmath29coefficients .
we proceed following ( * ? ? ?
* vii ) and @xcite .
let us consider the homology @xmath114 of @xmath72 with coefficients in the cellular chain complex @xmath115 associated to @xmath79 ; and call it the @xmath72-equivariant homology of @xmath79 .
as @xmath79 is contractible , the map @xmath116 to the point @xmath117 induces an isomorphism @xmath118 denote by @xmath119 the set of @xmath4-cells of @xmath79 , and make use of that the stabiliser @xmath120 in @xmath72 of any @xmath4-cell @xmath121 of @xmath79 fixes @xmath121 pointwise . then from @xmath122 shapiro s lemma yields @xmath123 and the equivariant leray / serre spectral sequence takes the form @xmath124 converging to the @xmath72-equivariant homology of @xmath79 , which is , as we have already seen , isomorphic to @xmath125 with the trivial action on the coefficients @xmath29 .
as in degrees above the virtual cohomological dimension , which is @xmath11 for the bianchi groups , the group homology is isomorphic to the farrell cohomology , we obtain the isomorphism type from the above mentioned general formulae . in the lower degrees @xmath126
, the following terms remain on the @xmath127-page , which is concentrated in the columns @xmath128 : @xmath129_{d^2_{2,0 } } { \operatorname{h}}_2(_\gamma\backslash{x } ; \thinspace { \mathbb z } ) } \ ] ] where the `` farrell supplement '' is the cokernel of the map @xmath130 induced by inclusion of finite cell stabilisers . as the cells
are fixed pointwise by their stabilisers , we see that for @xmath131 , the @xmath132-terms are concentrated in the two columns @xmath133 and @xmath134 . we compute the bottom row ( @xmath135 ) of the above spectral sequence as the homology of the quotient space @xmath136 .
then we infer from section [ connection ] that the rational rank of the differential @xmath137 is the number of non - trivial ideal classes of @xmath2 . using serre s decomposition of the homology group @xmath12 into the direct sum of the free abelian group with one generator for each element of the class group of @xmath2 and the group @xmath13 , and using the long exact sequence of section [ connection ] ,
we see that @xmath138 .
this has made it possible ot compute @xmath13 from the quotient space of the flge cellular complex in figures [ small absolute values ] and [ greater absolute values ] .
furthermore , we have an inclusion of @xmath12 into the group homology @xmath32 ; and the latter group homology is a quotient of the direct sum of @xmath12 and the farrell supplement .
this section recalls richard g. swan s work @xcite , which gives a concept from the theoretical viewpoint for an algorithm to compute the bianchi fundamental polyhedron .
the set @xmath52 which determines the bianchi fundamental polyhedron has been defined using infinitely many hemispheres .
but we will see that only a finite number of them are significant for this purpose and need to be computed .
we will state a criterion for what is an appropriate choice that gives us precisely the set @xmath52 .
this criterion is easy to verify in practice .
suppose we have made a finite selection of @xmath139 hemispheres .
the index @xmath82 running from @xmath20 through @xmath139 , we denote the @xmath82-th hemisphere by @xmath140 , where @xmath141 is its centre and given by a fraction @xmath142 in the number field @xmath143 . here
, we require the ideal @xmath144 to be the whole ring of integers @xmath36 .
this requirement is just the one already made for all the hemispheres in the definition of @xmath52 .
now , we can do an approximation of notation [ b ] , using , modulo the translation group @xmath56 , a finite number of hemispheres .
[ spaceabovehemispheres ] let @xmath145 : the inequality @xmath50 is fulfilled for all unimodular pairs with @xmath146 , for some @xmath147 and some @xmath148 @xmath51 .
then @xmath149 is the set of all points in @xmath10 lying above or on all hemispheres @xmath150 , @xmath151 ; for any @xmath148 .
the intersection @xmath152 with the fundamental domain @xmath153 for the translation group @xmath56 , is our candidate to equal the bianchi fundamental polyhedron .
we will give a method to decide when + this gives us an effective way to find @xmath52 by adding more and more elements to the set @xmath154 until we find @xmath155 .
[ delb ] we consider the boundary @xmath156 of @xmath149 in @xmath157 .
it consists of the points @xmath158 satisfying all the non - strict inequalities @xmath50 that we have used to define @xmath149 , and satisfy the additional condition that at least one of these non - strict inequalities is an equality .
we will see below that @xmath156 carries a natural cell structure .
this , together with the following definitions , makes it possible to state the criterion which tells us when we have found all the hemispheres relevant for the bianchi fundamental polyhedron .
[ strictlybelow ] we shall say that the hemisphere @xmath53 is _ strictly below _
the hemisphere @xmath159 at a point @xmath160 if the following inequality is satisfied : @xmath161 this is , of course , an abuse of language because there may not be any points on @xmath159 or @xmath53 with coordinate @xmath162 . however , if there is a point @xmath163 on @xmath53 , the right hand side of the inequality is just @xmath164 .
thus the left hand side is negative and so of the form @xmath165 .
clearly , @xmath166 and @xmath167 .
we will further say that a point @xmath158 is _ strictly below _
a hemisphere @xmath53 , if there is a point @xmath168 with @xmath167 .
we call _ cusps _ the elements of the number field @xmath169 considered as points in the boundary of hyperbolic space , via an embedding we write @xmath170 , which we also consider as a cusp .
it is well - known that the set of cusps is closed under the action of sl@xmath171 on @xmath64 ; and that we have the following bijective correspondence between the sl@xmath171-orbits of cusps and the ideal classes in @xmath19 .
a cusp @xmath75 is in the sl@xmath171-orbit of the cusp @xmath172 , if and only if the ideals @xmath173 and @xmath174 are in the same ideal class .
it immediately follows that the orbit of the cusp @xmath170 corresponds to the principal ideals .
let us call _ singular _ the cusps @xmath75 such that @xmath174 is not principal . and
let us call _ singular points _ the singular cusps which lie in @xmath175 .
it follows from the characterisation of the singular points by bianchi that they are precisely the points in @xmath176 which can not be strictly below any hemisphere . in the cases where @xmath19 is a principal ideal domain , @xmath177 consists of only one sl@xmath171-orbit , so there are no singular points .
we use the following formulae derived by swan , to compute representatives modulo the translations by @xmath56 , of the singular points .
the singular points of @xmath178 , are given by + @xmath179 , where @xmath180 , @xmath181 , @xmath182 , and * if @xmath183 or @xmath184 , + @xmath185 , @xmath186 , the numbers @xmath4 and @xmath113 are coprime , and @xmath4 is taken @xmath187 ; * if @xmath188 , + @xmath113 is even , @xmath189 , @xmath190 , the numbers @xmath4 and @xmath191 are coprime ; @xmath4 is taken@xmath192 .
the singular points need not be considered in swan s termination criterion , because they can not be strictly below any hemisphere @xmath53 .
we observe that the set of @xmath160 over which some hemisphere is strictly below another is @xmath41 or an open half - plane . in the latter case ,
the boundary of this is a line .
[ agreeingline ] denote by @xmath193 the set of @xmath160 over which neither @xmath159 is strictly below @xmath53 nor vice versa .
this line is computed by turning the inequality in definition [ strictlybelow ] into an equation .
swan calls it the line over which the two hemispheres _ agree _ , and we will see later that the most important edges of the bianchi fundamental polyhedron lie on the preimages of such lines .
we now restrict our attention to a set of hemispheres which is finite modulo the translations in @xmath56 .
[ collection ] consider a set of hemispheres @xmath150 , where the index @xmath82 runs from 1 through @xmath139 , and @xmath194 runs through @xmath19 .
we call this set of hemispheres a _ collection _ , if every non - singular point @xmath195 is strictly below some hemisphere in our set .
now consider a set @xmath149 which is determined by such a collection of hemispheres .
[ cancelcriterion ] we have if and only if no vertex of @xmath156 can be strictly below any hemisphere @xmath53 .
in other words , no vertex @xmath196 of @xmath156 can lie strictly below any hemisphere @xmath53 .
+ let us call _ height _ the coordinate @xmath197 of the upper - half space model introduced at the beginning of section [ the bianchi fundamental polyhedron ] . with this criterion , it suffices to compute the cell structure of @xmath156 to see if our choice of hemispheres gives us the bianchi fundamental polyhedron .
this has only to be done modulo the translations of @xmath56 , which preserve the height and hence the situations of being strictly below .
thus our computations only need to be carried out on a finite set of hemispheres .
we will in a first step compute the image of the cell structure under the homeomorphism from @xmath156 to @xmath41 given by the vertical projection . for each 2-cell of this structure
, there is an associated hemisphere @xmath53 .
the interior of this 2-cell consists of the points @xmath160 , where all other hemispheres in our collection are strictly below @xmath53 .
swan shows that this is the interior of a convex polygon .
the edges of these polygons lie on real lines in @xmath41 specified in notation [ agreeingline ] .
a vertex is an intersection point @xmath162 of any two of these lines involving the same hemisphere @xmath53 , if all other hemispheres in our collection are strictly below , or agree with , @xmath53 at @xmath162 .
now we can lift the cell structure back to @xmath156 , using the projection homeomorphism onto @xmath41 .
the preimages of the convex polygons of the cell structure on @xmath41 , are totally geodesic hyperbolic polygons each lying on one of the hemispheres in our collection .
these are the 2-cells of @xmath156 .
the edges of these hyperbolic polygons lie on the intersection arcs of pairs of hemispheres in our collection . as two euclidean 2-spheres intersect , if they do so non - trivially , in a circle centred on the straight line which connects the two 2-sphere centres , such an intersection arc lies on a semicircle centred in the complex plane .
the plane which contains this semicircle must be orthogonal to the connecting line , hence a vertical plane in @xmath10 .
we can alternatively conclude the latter facts observing that an edge which two totally geodesic polygons have in common must be a geodesic segment .
lifting the vertices becomes now obvious from their definition .
this enables us to check swan s termination criterion .
we will now sketch swan s proof of this criterion .
let @xmath86 be one of the convex polygons of the cell structure on @xmath41 .
the preimage of @xmath86 lies on one hemisphere @xmath140 of our collection .
now the condition stated in theorem [ cancelcriterion ] says that at the vertices of @xmath86 , the hemisphere @xmath140 can not be strictly below any other hemisphere .
the points where @xmath140 can be strictly below some hemisphere constitute an open half - plane in @xmath41 , and hence can not lie in the convex hull of the vertices of @xmath86 , which is @xmath86 .
theorem [ cancelcriterion ] now follows because @xmath41 is tessellated by these convex polygons .
from now on , we will work on putting swan s concept into practice .
we can reduce the set of hemispheres on which we carry out our computations , with the help of the following notion .
write @xmath203 and @xmath204 with @xmath205 .
we take any point @xmath163 strictly below @xmath53 and show that it is also strictly below @xmath206 . in terms of notation [ spaceabovehemispheres ] , this problem looks as follows : we assume that the inequality @xmath207 is satisfied , and show that this implies the inequality @xmath208 .
the first inequality can be transformed into + @xmath209 .
hence , @xmath210 .
we will insert this into the triangle inequality for the euclidean distance in @xmath211 applied to the three points @xmath163 , @xmath212 and @xmath213 , which is @xmath214 so we obtain @xmath215 . by definition [ everywherebelow ] ,
the expression on the right hand side is smaller than or equal to @xmath216 .
therefore , we take the square and obtain @xmath217 , which is equivalent to the claimed inequality .
[ liftonhemisphere ] let @xmath160 be a point lying within the vertical projection of @xmath53 .
define the _
lift on the hemisphere _
@xmath53 of @xmath162 as the point on @xmath53 the vertical projection of which is @xmath162 .
denote by the _ hemisphere list _ a list into which we will record a finite number of hemisphere s @xmath218 , @xmath219 , @xmath200 .
its purpose is to determine a set @xmath149 in order to approximate , and finally obtain , the bianchi fundamental polyhedron .
* initial step . *
we begin with the smallest value which the norm of a non - zero element @xmath220 can take , namely 1 .
then @xmath221 is a unit in @xmath19 , and for any @xmath222 , the pair @xmath223 is unimodular .
and we can rewrite the fraction @xmath75 such that @xmath224 .
we obtain the unit hemispheres ( of radius 1 ) , centred at the imaginary quadratic integers @xmath222 .
we record into the hemisphere list the ones which touch the bianchi fundamental polyhedron , i.e. the ones the centre of which lies in the fundamental rectangle @xmath225 ( of notation [ fundamentalrectangle ] ) for the action of @xmath56 on the complex plane .
* step a. * increase @xmath226 to the next higher value which the norm takes on elements of @xmath19 .
run through all the finitely many @xmath221 which have this norm . for each of these @xmath221 ,
run through all the finitely many @xmath227 with @xmath75 in the fundamental rectangle @xmath225 . check that @xmath228 and that the hemisphere @xmath53 is not everywhere below a hemisphere @xmath159 in the hemisphere list .
if these two checks are passed , record @xmath223 into the hemisphere list .
* step b. * we compute the lines @xmath229 of definition [ agreeingline ] , over which two hemispheres agree , for all pairs @xmath159 , @xmath53 in the hemisphere list which touch one another .
+ then , for each hemisphere @xmath159 , we compute the intersection points of each two lines @xmath229 and @xmath230 referring to @xmath231 . + we drop the intersection points at which @xmath159 is strictly below some hemisphere in the list . + we erase the hemispheres from our list , for which less than three intersection points remain .
we can do this because a hemisphere which touches the fundamental polyhedron only in two vertices shares only an edge with it and no 2-cell .
+ now , the vertices of @xmath152 are the lifts of the remaining intersection points .
thus we can check swan s termination criterion ( theorem [ cancelcriterion ] ) , which we do as follows .
we pick the lowest value @xmath232 for which @xmath233 is the lift inside hyperbolic space of a remaining intersection point @xmath162 .
+ if @xmath234 , then all ( infinitely many ) remaining hemispheres have radius equal or smaller than @xmath197 , so @xmath163 can not be strictly below them .
so swan s termination criterion is fulfilled , we have found the bianchi fundamental polyhedron , and can proceed by determining its cell structure .
+ else , @xmath197 becomes the new expected value for @xmath235 .
we repeat step * a * until @xmath226 reaches @xmath236 and then proceed again with step * b*. * input : a square - free positive integer @xmath1 . *
output : the hemisphere list , containing entries @xmath218, ,@xmath200 such that @xmath155 .
let @xmath19 be the ring of integers in @xmath0 .
let @xmath237 be the class number of @xmath19 .
compute @xmath237 . estimate the highest value for @xmath226 which will occur in notation [ spaceabovehemispheres ] by the formula @xmath238
@xmath239 @xmath240 .
swans_cancel_criterion_fulfilled @xmath241 false .
* * execute algorithm [ record_hemispheres ] with argument @xmath242 .
increase @xmath242 to the next greater value in the set @xmath243 of values of the norm on @xmath19 .
compute @xmath197 with algorithm [ get_minimal_proper_vertex_height ] .
all ( infinitely many ) remaining hemispheres have radius smaller than @xmath197 , * the value @xmath197 is the minimal height of the non - singular vertices of the cell complex @xmath156 determined by the hemisphere list @xmath246 . +
all the hemispheres which are not in the list , have radius smaller than @xmath244 . by remark [ expectedlimit ] ,
the inequality @xmath247 will become satisfied ; and then no non - singular vertex of @xmath156 can be strictly below any of them . hence by theorem [ cancelcriterion ] , @xmath155 ; and we obtain the bianchi fundamental polyhedron as @xmath152 . * we now consider the run - time . by , the set of hemispheres @xmath248 is finite .
so , there exists an @xmath53 for which the norm of @xmath221 takes its maximum on this finite set .
the variable @xmath242 reaches this maximum for @xmath226 after a finite number of steps ; and then swan s termination criterion is fulfilled .
the latter steps require a finite run - time because of propositions [ runtime2 ] and [ runtime3 ] .
swan explains furthermore how to obtain an a priori bound for the norm of the @xmath220 occurring for such hemispheres @xmath53 .
but he states that this upper bound for @xmath226 is much too large .
so instead of the theory behind theorem [ finitenumber ] , we use swan s termination criterion ( theorem [ cancelcriterion ] above ) to limit the number of steps in our computations .
we then get the following . [ boundsformu ] we can give bounds for @xmath226 in the cases where @xmath249 is of class number 1 or 2 ( there are nine cases of class number 1 and eighteen cases of class number 2 , and we have done the computation for all of them ) .
they are the following : @xmath250 where @xmath251 is the discriminant of @xmath252 , i.e. , @xmath253 * input : the value @xmath242 , and the hemisphere list ( empty by default ) .
* output : the hemisphere list with some hemispheres of radius @xmath244 added .
let @xmath254 .
let @xmath255 be the length of the hemisphere list .
everywhere_below : = false , @xmath256 .
let @xmath159 be the @xmath257th entry in the hemisphere list ; everywhere_below : = true . increase @xmath257 by @xmath20 .
record @xmath53 into the hemisphere list .
we recall that the notion `` everywhere below '' has been made precise in definition [ everywherebelow ] ; and that the fundamental rectangle @xmath225 has been specified in notation [ fundamentalrectangle ] . * * [ runtime2 ] algorithm _ [ record_hemispheres ] _ finds all the hemispheres of radius @xmath244 , on which a 2-cell of the bianchi fundamental polyhedron can lie .
this algorithm terminates within finite time .
* directly from the definition of the hemispheres @xmath53 , it follows that the radius is given by @xmath235 .
so our algorithm runs through all @xmath221 in question . by construction of the bianchi fundamental polyhedron @xmath258 , the hemispheres on which a 2-cell of @xmath258 lies must have their centre in the fundamental rectangle @xmath225 . by proposition [ everywherebelowhemispheres ]
, such hemispheres can not be everywhere below some other hemisphere in the list .
* now we consider the run - time of the algorithm .
there are finitely many @xmath220 the norm of which takes a given value . and
for a given @xmath221 , there are finitely many @xmath222 such that @xmath75 is in the fundamental rectangle @xmath225 .
therefore , this algorithm consists of finite loops and terminates within finite time .
[ runtime3 ] algorithm _ [ get_minimal_proper_vertex_height ] _ finds the minimal height occurring amongst the non - singular vertices of @xmath156 .
this algorithm erases only such hemispheres from the list , which do not change @xmath156 .
it terminates within finite time .
* the heights of the points in @xmath10 are preserved by the action of the translation group @xmath56 , so we only need to consider representatives in the fundamental domain @xmath153 for this action .
our algorithm computes the entire cell structure of as described in subsection [ cellstructure ] .
the number of lines to intersect is smaller than the square of the length of the hemisphere list , and thus finite . as a consequence ,
the minimum of the height has to be taken only on a finite set of intersection points , whence the first claim . *
if a cell of @xmath156 lies on a hemisphere , then its vertices are lifts of intersection points .
so we can erase the hemispheres which are strictly below some other hemispheres at all the intersection points , without changing @xmath156 .
* now we consider the run - time .
this algorithm consists of loops running through the hemisphere list , which has finite length . within one of these loops , there is a loop running through the set of pairs of lines @xmath229 . a ( far too large ) bound for the cardinality of this set is given by the fourth power of the length of the hemisphere list .
+ the steps performed within these loops are very delimited and easily seen to be of finite run - time .
* input : the hemisphere list @xmath246 . *
output : the lowest height @xmath197 of a non - singular vertex of @xmath156 . and the hemisphere list with some hemispheres removed which do not touch the bianchi fundamental polyhedron .
compute the line @xmath229 of notation [ agreeingline ] .
compute the intersection point of @xmath229 and @xmath230 , if it exists .
drop the intersection points at which @xmath159 is strictly below some hemisphere in the list .
erase the hemispheres from our list , for which no intersection points remain .
now the vertices of @xmath152 are the lifts ( specified in definition [ liftonhemisphere ] ) on the appropriate hemispheres of the remaining intersection points .
pick the lowest value @xmath232 for which @xmath233 is the lift on some hemisphere of a remaining intersection point @xmath162 .
return @xmath197 . *
* with the method described in subsection [ cellstructure ] , we obtain a cell structure on the of the bianchi fundamental polyhedron .
the cells in this structure which touch the cusp @xmath68 are easily determined : they are four 2-cells each lying on one of the euclidean vertical planes bounding the fundamental domain @xmath153 for @xmath56 specified in notation [ fundamentalrectangle ] ; and four 1-cells each lying on one of the intersection lines of these planes .
the other in this structure lie each on one of the hemispheres determined with our realization of swan s algorithm .
as the bianchi fundamental polyhedron is a hyperbolic polyhedron up to some missing cusps , its boundary cells can be oriented as its facets .
once the cell structure is subdivided until the cells are fixed pointwise by their stabilisers , this cell structure with orientation is transported onto the whole hyperbolic space by the action of @xmath72 . from this operation formula , we establish equations and inequalities on the entries of a matrix sending a given point @xmath163 to another given point @xmath263 in @xmath10 .
we will use them in algorithm [ getvertexidentifications3mod4 ] to compute such matrices . for the computation of the vertex stabilisers , we have @xmath264 , which simplifies the below equations and inequalities as well as the pertinent algorithm .
first , we fix a basis for @xmath19 as the elements @xmath20 and @xmath265 as we have put @xmath61 and @xmath62 , the only units in the ring @xmath19 are @xmath266 .
we will use the notations @xmath267 and @xmath268 for @xmath269 .
* input : the points @xmath270 , @xmath271 in the interior of @xmath10 , where @xmath162 , @xmath272 and @xmath273 , @xmath274 .
* output : the set of matrices @xmath260 @xmath275 sl@xmath276 , @xmath277 , with nonzero entry @xmath96 , sending the first of the input points to the second one .
@xmath96 will run through @xmath19 with write @xmath96 in the basis as @xmath278 , where @xmath279 .
@xmath281 ; write @xmath282 in the basis as @xmath283 with @xmath284 .
@xmath285 will run through @xmath19 with @xmath286 .
write @xmath285 in the basis as @xmath287 , where @xmath288 .
@xmath290 ; @xmath291 .
do the following for both @xmath292 and @xmath293 if @xmath294 .
@xmath295 ; @xmath296 .
@xmath297 is determined by the determinant @xmath20 : @xmath298 .
check that @xmath260@xmath299 .
return @xmath300 . *
* from the operation equation @xmath260@xmath307 , we deduce and conclude whence the first inequality .
we insert @xmath308 @xmath309 into it , and obtain @xmath310 we observe that @xmath311 is a quadratic function in @xmath312 , taking its values exclusively in @xmath313 . hence its graph has the shape of a parabola , and the negative values of @xmath311 appear exactly on the interval where @xmath314 is between its two zeroes , this implies the third and fourth claimed inequalities .
as @xmath314 is a real number , @xmath251 must be non - negative in order that @xmath311 be non - positive .
hence @xmath317 , which gives the second claimed inequality .
let @xmath334 with @xmath335 .
our first step is to show that consider the case @xmath336 .
then @xmath337 and we have shown our claim . else
consider the case @xmath188 .
then , @xmath338 and our claim follows for @xmath339 .
now we have shown that @xmath340 ; and we use some embedding of @xmath249 into @xmath41 to verify the equation @xmath341 . * the first claim is easily established using the bounds and formulae stated in lemmata [ jandk ] and [ sandq ] .
* now we consider the run - time .
this algorithm consists of three loops the limits of which are at most linear expressions in @xmath344 . for @xmath345 , we use lemma [ simpleinequation ] and @xmath346 to see this ( we get a factor @xmath347 here , which we can neglect ) . finally , it should be said that the scope of computations one can do with geometric models for the bianchi groups does not stop once the integral homology of the full group is known .
there is further interest in homology with twisted coefficients , congruence subgroups and modular forms ( see for instance @xcite , @xcite ) .
currently , page @xcite is working on optimizing algorithms in order to obtain more cell complexes for bianchi groups and other kleinian groups .
the author would like to thank bill allombert ( pari / gp development headquarters ) and philippe elbaz - vincent ( ujf grenoble ) for invaluable help with the development of _ bianchi.gp_. he is grateful to the weizmann institute of science for providing him its high perfomance computation clusters in order to establish the cell complexes database ; and to stephen s. gelbart and graham ellis for support and encouragement .
nicolas bergeron and akshay venkatesh , _ the asymptotic growth of torsion homology for arithmetic groups _
, journal of the institute of mathematics of jussieu , + http://dx.doi.org/10.1017/s1474748012000667 eduardo r. mendoza , _ cohomology of @xmath349 over imaginary quadratic integers _ , bonner mathematische schriften , * 128 * , dissertation , rheinische friedrich - wilhelms - universitt , mathematisches
institut , 1979 .
alexander d. rahm and mehmet haluk engn , _ on level one cuspidal bianchi modular forms _ , preprint , http://hal.archives-ouvertes.fr/hal-00589184/en/ robert riley , _ applications of a computer implementation of poincar s theorem on fundamental polyhedra _ , math
* 40 * ( 1983 ) , no . 162 , 607632 .
mr 689477 ( 85b:20064 ) dan yasaki , _ hyperbolic tessellations associated to bianchi groups _ , algorithmic number theory , 9th international symposium , nancy , france , ants - ix , july 19 - 23 , 2010 , proceedings , 385396 .
mr 2721434 ( 2012g:11069 ) | denote by @xmath0 , with @xmath1 a square - free positive integer , an imaginary quadratic number field , and by @xmath2 its ring of integers .
the _ groups _ are the groups @xmath3 . in the literature ,
there has been so far no example of @xmath4-torsion in the integral homology of the full bianchi groups , for @xmath4 a prime greater than the order of elements of finite order in the bianchi group , which is at most @xmath5 . however ,
extending the scope of the computations , we can observe examples of torsion in the integral homology of the quotient space , at prime numbers as high as for instance @xmath6 at the discriminant @xmath7 . |
all fields considered in this paper are of characteristic zero .
let @xmath8 be a ring .
a map @xmath9 satisfying @xmath10 and @xmath11 for all @xmath12 is called _
derivation_. we will denote @xmath13 by @xmath14 and @xmath15 by @xmath16 . a _ differential ring _
@xmath8 is a ring with a specified derivation .
a differential ring which is a field will be called a _
differential field_. let @xmath1 be a differential field extension and @xmath6 .
let us denote by @xmath17 the differential subfield of @xmath0 generated by @xmath3 and @xmath7 .
if @xmath18 , then element @xmath7 is said to be _
primitive_. an element @xmath19 of the differential ring @xmath8 is said to be _ constant _ if @xmath20 .
kolchin proved ( @xcite ) a differential analogue of the primitive element theorem : let @xmath21 and @xmath2 .
assume also that @xmath3 contains a nonconstant element .
then , there exists @xmath22 such that @xmath23 .
let @xmath21 and @xmath2 .
assume also that @xmath0 contains a nonconstant element .
then , there exist @xmath24 such that @xmath25 . in @xcite kolchin considered a more general case , i.e. fields equipped with a set of commuting derivations .
we restrict ourselves to the ordinary case . in @xcite babakhanian constructed primitive elements for several specific extensions @xmath1 with @xmath3 consisting of constant elements .
the goal of the present paper is to prove the primitive element theorem for the case @xmath26 for all @xmath27 .
[ th : density ] let @xmath28 , @xmath29 , and @xmath30
. then , there exists @xmath31 $ ] such that @xmath32 . unlike kolchin s proof it is not sufficient to consider elements of the form @xmath35 ( @xmath36 ) .
for example , let @xmath37 be a differential field with the derivation defined by @xmath38 and @xmath39 .
there is no primitive element of the form @xmath40 ( @xmath41 ) , but @xmath42 . without loss of generality
, we can assume that @xmath59 .
let us introduce algebraically independent variables @xmath60 .
we extend the derivation from @xmath0 to @xmath61 $ ] by @xmath62 .
this construction can be explained by the following observation : let us fix a polynomial @xmath63 $ ] ; the formulas @xmath64 define a homomorphism of differential @xmath47-algebras @xmath65 \to e$ ] .
let @xmath66 and @xmath67 . since @xmath68 , @xmath69 .
let nonzero @xmath70 $ ] satisfy @xmath71 .
notice that it depends on @xmath72 .
multiplying by the suitable element of @xmath73 $ ] , we obtain a polynomial in both @xmath74 and @xmath75 over @xmath47 .
let us denote it by @xmath76 .
moreover , we assume that @xmath77 satisfies the following conditions : assume that @xmath81 .
let us rewrite @xmath77 as a polynomial in @xmath82 : @xmath83 , where @xmath84 are polynomials over @xmath47 in @xmath85 .
since @xmath81 , @xmath86 and @xmath82 is transcendental over @xmath87 .
thus , @xmath88 implies @xmath89 for all @xmath90 .
we obtained a contradiction with minimality of @xmath52 .
let @xmath63 $ ] .
applying @xmath99 to @xmath100 , we obtain an algebraic dependence for @xmath101 $ ] over @xmath102 .
it yields to an algebraic dependence for @xmath103 over @xmath104 .
the goal is to find an appropriate @xmath105 to make this dependence nontrivial .
let us compute its derivation with respect to @xmath103 : by lemma [ lem : part_dev_nonzero ] the polynomial @xmath107 is nonzero .
since @xmath66 , we can rewrite @xmath108 as a nonzero polynomial in @xmath109 over @xmath110 .
let us denote the derivation on @xmath0 by @xmath111 .
then , @xmath112 is also a derivation on @xmath0 .
obviously , @xmath113 .
hence , we can apply lemma [ lem : nonzero ] to the differential field @xmath0 with respect to @xmath114 , nonconstant element @xmath22 , and the polynomial @xmath108 in variables @xmath115 .
therefore , we obtain such @xmath116 $ ] that @xmath117 .
let @xmath130 be a differential extension field of @xmath131 . by an _ isomorphism of @xmath130 with respect to @xmath131
_ we will mean an isomorphic mapping of @xmath130 onto a differential field @xmath132 such that [ lem : invariant ] let @xmath0 be an extension of @xmath3 and @xmath133 .
a necessary and sufficent condition for @xmath134 is that no isomorphism of @xmath0 with respect to @xmath3 other then the identity leaves @xmath135 invariant .
let @xmath136 ( shortly , @xmath137 ) have a solution @xmath138 and @xmath139 ( shortly , @xmath140 ) have a solution @xmath138 , @xmath141 .
we will show that there exist elements @xmath142 such that @xmath143 takes different values for different solutions of @xmath144 . then
, certainly @xmath145 will satisfy the requirements on @xmath135 from lemma [ lem : invariant ] .
let @xmath149 be the decomposition of @xmath150 into essential prime differential ideals ( see @xcite ) , and suppose the subscripts have been assigned so that each @xmath151 , @xmath152 , @xmath153 contains both @xmath154 and @xmath155 , whereas @xmath156 , @xmath152 , @xmath157 each fails to contain either @xmath154 or @xmath155 .
consider @xmath158 with @xmath159 .
if @xmath160 , then also @xmath161 .
thus , @xmath161 for all @xmath159 .
let @xmath162 be a generic solution of @xmath158 ( see @xcite ) .
differentiating the equation for the sake of simplicity let us consider @xmath183 .
the proof for the other cases is analogous .
the last row of the corresponding matrix is a sum of three rows : @xmath184 , @xmath185 and @xmath186 where @xmath187 $ ] for all @xmath90 .
thus , the determinant @xmath188 can be expressed as a sum : since all these differential polynomials are symmetric in @xmath175 and @xmath176 , it suffices to prove that at least one of them depends on either @xmath197 or @xmath198 .
assume the contrary , that @xmath199 . by the cramer s rule these fractions are solutions of the following system of linear equations in @xmath200 :
we are going to assign values from a particular differential field to @xmath175 and @xmath176 .
precisely , let @xmath204 be a field of rational functions equipped with standard derivation ( @xmath205 ) and @xmath206 be a primitive @xmath207-th root of unity .
let @xmath208 and @xmath209 .
the matrix of the system ( [ eq : system ] ) is nondegenerate , because its determinant equals @xmath210 , which is nonzero since @xmath211 , @xmath152 , @xmath212 are linearly independent over constants ( ( * ? ? ?
clearly , the unique solution of the system ( [ eq : system ] ) in this case is @xmath213 .
but the equalities ( [ eq : cons ] ) do not hold .
this contradiction proves the lemma .
1 . @xmath220 .
let @xmath221 be the index from the corollary [ cor : alg_relation ] .
since @xmath222 , @xmath223 provides us an algebraic relation between @xmath7 , @xmath152 , @xmath224 , @xmath225 , @xmath152 , @xmath226 .
hence , @xmath227 is algebraic over @xmath228 for all @xmath229 .
thus , @xmath230 is algebraic over @xmath228 . since @xmath231
, we are done .
2 . @xmath232 . in this case both @xmath233 and @xmath234 can be computed directly : + @xmath235 + @xmath236 + if both @xmath237 and @xmath238 vanish , either @xmath239 or the determinant of the system of linear equations @xmath240 in variables @xmath241 and @xmath242 vanishes , i.e. @xmath243 . both cases are impossible since @xmath244 .
1 . @xmath248 .
thus , at least one of @xmath249 minors of the matrix : + @xmath250 + is nondegenerate .
let @xmath251 .
multiplying by the inverse matrix , we obtain the formulas which express @xmath252 as a rational function in @xmath7 , @xmath103 , @xmath225 , @xmath253 and their derivations , @xmath254 , @xmath255 for all @xmath256 and @xmath257 .
+ hence , @xmath258 . by lemma [ lem : trdeg_estim1 ] , @xmath259 .
thus + @xmath260 2 . @xmath261 . in this case
there exists @xmath263 such that @xmath264 . by the same argument as above @xmath265 .
the desired inequality is now obvious .
lemma [ lem : trdeg_estim2 ] implies that @xmath266 , @xmath152 , @xmath267 are algebraically dependent over @xmath110 .
let us denote this dependence by @xmath268 $ ] .
consider the polynomial @xmath269 .
let @xmath128 satisfy @xmath270 .
then , @xmath271 for any solution of @xmath272 other than @xmath273
. therefore , the proof of the theorem is complete . | in this paper we strengthen kolchin s theorem ( @xcite ) in the ordinary case .
it states that if a differential field @xmath0 is finitely generated over a differential subfield @xmath1 , @xmath2 , and @xmath3 contains a nonconstant , i.e. an element @xmath4 such that @xmath5 , then there exists @xmath6 such that @xmath0 is generated by @xmath7 and @xmath3 .
we replace the last condition with the existence of a nonconstant element in @xmath0 .
differential field , primitive element , derivation 12h05 , 13n15 |
superstring theories enjoy a number of interesting properties at first sight not shared by ordinary four - dimensional field theories . for a superstring theory , ( 1 )
the basic entity is a string of size @xmath0 ) cm , with massless levels and excitation energies @xmath1 ) gev ; ( 2 ) loop corrections are ultraviolet finite and quantum gravity is well defined ; ( 3 ) the fundamental dynamical variables consist of the spacetime @xmath2 and the internal @xmath3 fields , all as functions of the worldsheet coordinates @xmath4 and @xmath5 ; ( 4 ) these variables propagate as independent free fields throughout the worldsheet , in a manner dependent on the topology but not on the geometry of the worldsheet ( reparametrization and conformal invariance ) ; ( 5 ) an external photon of momentum @xmath6 and wave function @xmath7 is inserted into the string through a vertex operator @xmath8\exp[ip\!\cdot\!x(\sigma,\tau)]$ ] , a form which is fixed by conformal invariance ; ( 6 ) conformal invariance leads to local ( veneziano ) duality of the scattering amplitude .
a scattering process is described by one or very few string diagrams . in particular , for elastic scattering in the tree approximation , one string diagram ( the veneziano amplitude ) gives rise simultaneously to all the @xmath9-channel and the @xmath10-channel exchanges . in constrast , in ordinary four - dimensional field theories ( qft ) , ( @xmath11 ) the basic entities are point particles ; ( @xmath12 ) loop corrections contain ultraviolet divergences and quantum gravity is non - renormalizable ; ( @xmath13 ) the fundamental dynamical variables are fields @xmath14 of the four - dimensional spacetime coordinates @xmath15 ; ( @xmath16 ) these fields propagate freely only between vertices , where interactions take place ; ( @xmath17 ) external photons are inserted into a feynman diagram through the photon operator @xmath18 $ ] ; ( @xmath19 ) there are many distinct feynman diagrams contributing to a scattering amplitude . for elastic scattering , @xmath9- and @xmath10-channel exchanges
are given by different diagrams that must be added up together .
in the ` low - energy ' limit when @xmath20 gev , a string of dimension @xmath21 cm is indistinguishable from a point , energy levels @xmath1 ) gev are too high to matter , so one would expect a string theory to reduce to an ordinary field theory of zero masses .
in fact , explicit calculations have been carried out @xcite to show that a one - loop @xmath22-gluon amplitude in a string theory does reduce to corresponding results in field theory . in order to go beyond one loop or the @xmath22-gluon amplitude , where no simple string expression is available
, it is better to proceed in a different way .
since the string properties ( 3)(6 ) are quite different from the field - theory properties ( @xmath13)(@xmath19 ) , the equivalence of string and qft at low energies is not immediately obvious .
the purpose of this series of papers is , among other things , to make these connections .
there are two motivations for doing so . at a theoretical level
, one can hope to gain new insights into gauge and gravitational theories from the string arrangement .
for example , according to ( 3 ) and ( @xmath13 ) , spacetime and internal coordinates are treated on an equal footing in a string , but not so in qft .
this asymmetry makes _
local _ gauge transformations in qft awkward to deal with and practical calculations difficult to obtain
. it would therefore be nice to be able to reformulate gauge and a gravitational theories in the symmetrical way of a string .
conversely , one can hope that the successful multiloop treatment of qft , when written in a string - like manner , can give hints useful for multiloop string calculations . on the practical side , string - like organization of an field - theoretical amplitude
allows the spinor helicity technique @xcite to be used on multiloop diagrams as easily as for tree diagrams @xcite ; it also enables color , spin , and momentum degrees of freedom to propagate independently , a separation that leads to efficient simplifications in actual calculations , so much so that amplitudes not computable by ordinary means can be obtained when organized in this novel way
. examples of this includes the parke - taylor @xmath22-gluon amplitude @xcite , the factorization of identical - helicity photons produced in the @xmath23 annihilation into @xmath24-pairs @xcite , the computation of the one - loop @xmath22-photon amplitude of identical helicities @xcite , and the calculation of the @xmath22-gluon one - loop amplitudes @xcite . in the low energy limit ,
the variable @xmath4 along the string is frozen , so the dynamical fields in ( 3 ) are effectively functions of @xmath5 alone . to convert ( @xmath13 ) to ( 3 ) , one must be able to free the dependence of @xmath25 on @xmath26 , to make them both independent functions of some _ proper time _ @xmath5 .
this can be accomplished by using the schwinger - parameter representation for the field - theoretic scattering amplitude @xcite . in this representation ,
every vertex @xmath27 in a feynman diagram is assigned a proper time @xmath28 , and each propagator @xmath29 is assigned a schwinger parameter @xmath30 . if @xmath31 is a line connecting vertices @xmath27 to @xmath32 , then @xmath33 .
if we regard the feynman diagram as an electric circuit with resistances @xmath30 , then @xmath34 can be interpreted as ( the four - dimensional ) voltage at the vertex @xmath27 @xcite , thus freeing it to be an independent function of @xmath5 .
spacetime flow is thus analogous to the current flow of an electric network , or equivalently the change of the electrical potential from point to point .
color flows can be isolated by creating color - oriented vertices in feynman diagrams @xcite .
one color - oriented vertex may be related to another by twising , thus creating twisted feynman diagrams in much the same way like twisted open strings .
the color subamplitudes so isolated with these color - flow factors are gauge invariant . with massless fermions , chirality and helicities
are conserved and this allows free and unobstructed spin flows .
this is the essence of the spinor helicity technique which is applicable to photons and gluons as well , for a spin-1 particle can kinematically be regarded as the composite of two spin-@xmath35 particles .
the schwinger proper - time formalism allows the spinor helicity technique to continue to work in loop diagrams @xcite ; otherwise the loop momentum would be there to break chirality conservation and to obstruct the free flow of spins .
the spacetime , color , and spin flows thus obtained approximate the properties ( 3 ) and ( 4 ) of the string .
moreover , using _
differential circuit identities _
@xcite , external electromagnetic vertices can be converted from ( @xmath17 ) to ( 5 ) .
what remains to be considered in completing the string organization of field theory is then the property of _ local duality_. historically it was this unusual feature , first seen in the veneziano model @xcite , that marked the beginning of string theory . to what extent this interesting property
is retained in field theory is therefore an interesting topic to study .
we propose to make a first attempt in that direction in the present paper .
duality for tree diagrams is considered in sec . 2 , where its exact meaning in field theory
is also discussed .
for the sake of simplicity we shall confine that section to a scalar theory where a scalar photon field @xmath36 is coupled to a charged scalar meson field @xmath37 via the lagrangian @xmath38 , but it turns out that once we solve the duality problem here it is solved in quantum electrodynamics as well . to prepare for discussions of qed and multiloop amplitudes , we review in sec .
3 the schwinger - parameter representation for a field - theoretic scattering amplitude which provides the main tool for further discussions . duality for the multiloop scalar theory is discussed in sec .
4 , and duality for qed is discussed in sec .
the problem of qcd is much harder and we shall defer that to a future publication . finally , a concluding section appears in sec .
duality in a string theory follows from its conformal invariance , which in the special case of a four - point amplitude can be used to fix the ( complex ) worldsheet positions of three of the external lines to be @xmath39 and @xmath40 , and the fourth one to be @xmath41 $ ] .
a particularly simple example is the veneziano amplitude @xcite in the mandelstam variables @xmath9 and @xmath10 , @xmath42 this amplitude has poles when either @xmath9 or @xmath10 is a non - negative integer ( in units of @xmath43 gev]@xmath44 ) , but there are no simultaneous @xmath9- and @xmath10-channel poles .
the amplitude can be expanded either as a sum of @xmath9-channel poles , represented purely by @xmath9-channel exchange feynman diagrams , _ or _ a sum of @xmath10-channel poles , represented purely by @xmath10-channel exchange diagrams .
there is no need to _ add _ both the @xmath9-channel and the @xmath10-channel diagrams , as is necessary in ordinary quantum field theories .
the @xmath10-channel amplitude is obviously equal to the @xmath9-channel amplitude , and both are equal to the single integral in ( 2.1 ) .
duality_. these results can be obtained directly from a pole expansion of the euler gamma functions , or from the integral representation for the beta function . in the latter case ,
an expansion of the integrand about @xmath45 gives rise to the @xmath10-channel poles , and an expansion of the integrand about @xmath46 gives rise to the @xmath9-channel poles .
at present energies , @xmath47 ( in units of @xmath48 ) , so only the massless poles contribute , giving rise to @xmath49 the @xmath10-channel pole comes from the divergence of the integral near @xmath45 when @xmath50 , and the @xmath9-channel pole comes from the divergence of the integral near @xmath46 when @xmath51 . in this form ,
the amplitude does not _ appear _ to be ` dual ' anymore , because _ both _
the @xmath9-channel and the @xmath10-channel poles are _ summed _ , instead of having a _ single _ expression like ( 2.1 ) , where only a sum of the @xmath9-channel _ or _ a sum of the @xmath10-channel poles are present .
nevertheless , appearances are deceiving , because ( 2.2 ) follows mathematically from ( 2.1 ) , which is dual .
in other words , the @xmath10-channel poles in ( 2.2 ) can be formally obtained by an infinite sum of massive @xmath9-channel poles , and ( 2.2 ) is as dual as it can be at low energies .
it is instructive for later discussions to obtain ( 2.2 ) directly from the integral representation of ( 2.1 ) . for that purpose ,
divide the integral in @xmath26 into two halves at the midpoint @xmath52 .
since @xmath53 , the contribution of the integral comes mainly from @xmath45 , so we can put @xmath54 there .
similarly the term @xmath55 can be ignored in the second integral .
next , make the transformation @xmath56 in the first integral , so that @xmath57 $ ] there , and the transformation @xmath58 $ ] in the second integral , so that @xmath59 $ ] there . then for @xmath53 , the integral in ( 2.1 ) becomes @xmath60 which is the same as ( 2.2 ) .
-2 cm 4.7 truein -7.5 cm.1 cm if one were to start directly from a massless scalar field theory @xmath61 ( where all fields are scalar and massless ) , then the ` compton scattering ' amplitude given by fig . 1 is identical to ( 2.2 ) . in that sense
the field - theoretic amplitude is already dual , or as dual as it can be at the present energy range .
one might still be unsatisfied with this remark about duality , and points out that the original dual amplitude in ( 2.1 ) is given by a _
single _ integral , whereas in ( 2.3 ) this is given by the sum of _ two _ integrals . since ( 2.3 ) comes from ( 2.1 ) it must be able to write it as a single integral as well . all that we have to do is to define a function @xmath62 , then @xmath63 which is of course equivalent to the veneziano integral ( 2.1 ) at the present energy range .
we shall refer to expressions of this type , where the sum of a _ number of diagrams _ is represented by a _
single _ ( possibly multi - dimensional ) integral , as _
dual expressions_. what allows the dual expression ( 2.4 ) to be written is not so much the explicit form of the integrands in ( 2.3 ) , but that the two integrals there have non - overlapping ranges in @xmath64 . given that , it is always possible to define a common integrand @xmath65 so that the two integrals can be combined into one .
similar reasoning shows that dual expressions can be written for other processes in a scalar - photon theory . since we are imitating the low energy limit of strings
, we may simplify writings by assuming all particles to be massless . in that case
, the schwinger - parameter representation for a scalar propagator is @xmath66 and the variable @xmath67 is called a schwinger proper - time parameter . using this ,
we can obtain ( 2.3 ) from ( 2.5 ) simply by letting @xmath68 , @xmath69 in the first term , and @xmath70 , @xmath71 in the second term .
the fact that we can get the veneziano integral representation ( 2.3 ) from the schwinger representation confirms the claim that the schwinger - parameter representations are string - like .
consider now a more complicated example , fig . 2 , in which charged particles scatter to produce @xmath72 ( scalar ) photons from one charged line and @xmath22 photons from another .
we have drawn only one of the @xmath73 possible diagrams ; others are obtained from it by permuting the photon lines .
.5 cm 4.7 truein -10.5 cm let us first establish some common notations .
assign to each vertex of a feynman diagram a _ proper time _
@xmath5 , as illustrated in fig . 2 . the schwinger proper - time parameters
@xmath67 are then given by differences of the proper times .
specifically , if @xmath31 is an internal line between vertices @xmath27 and @xmath32 , then @xmath74 .
in this way , all the proper - time differences are determined but translational invariance prevents the origin of proper time to be fixed .
let @xmath75}&&\equiv \int_a^bd\tau_1\int_{\tau_1}^bd\tau_2 \cdots\int_{\tau_{n-1}}^bd\tau_n\ , \nonumber\\ \int_a^b d\tau_{[12(345)678]}&&\equiv \int_a^b d\tau_{[12 ] } \int_{\tau_2}^bd\tau_{[678 ] } \left(\prod_{i=3}^5\int_{\tau_2}^{\tau _ 6}d\tau_i\right)\ , \nonumber\\ \langle\int d\tau_{[12\cdots n]}\rangle&&\equiv\lim_{t\to\infty}{1\over 2 t } \int_{-t}^{t}d\tau_{[12\cdots n]}\ , \nonumber\\ \langle\int d\tau_{[12(345)678]}\rangle&&\equiv\lim_{t\to\infty}{1\over 2 t } \int_{-t}^t d\tau_{[12(345)678]}\ , \nonumber\\ \langle\int d\tau_{(12\cdots n)}\rangle&&\equiv \langle\int d\tau_{[(12\cdots n)]}\rangle \ .\end{aligned}\ ] ] in short , the integration variable enclosed between square brackets are ordered , and those between round brackets are unordered .
we can now return to fig .
its amplitude is @xmath76}\rangle\langle\int d\tau'_{[12\cdots 0\cdots n]}\rangle\exp(ip)\ , \ ] ] for some quadratic function @xmath77 of the external momenta obtained by using ( 2.5 ) on each propagator .
the detailed form of @xmath77 does not concern us at the moment .
the other diagrams are obtained by permuting the photon lines in fig . 2 , so their amplitudes are all given by something like ( 2.7 ) , but with the @xmath5 and @xmath78 integration regions permuted separately .
the detailed form of the quadratic function @xmath77 may change from region to region but again we do not have to worry about it . since the integration regions of these different diagrams do not overlap ,
it is possible to define a common @xmath79 equal to the individual @xmath77 s in their respective regions . in this way
, all the @xmath73 diagrams can be summed up to get a single dual expression @xmath80 in which the integration regions are completely unordered .
the reasoning can obviously be extended to any tree diagram in a scalar - photon theory . in every case
, each charged line provides a platform for ordering the photon lines attached to it .
different diagrams correspond to different permutations of these photon lines , so they correspond to different integration regions in the proper times .
a sum into a single dual expression in which the proper time integration regions are unordered is clearly possible .
furthermore , with minor modifications to be discussed in sec .
4 , essentially the same consideration works for multiloop diagrams as well .
there are three remarks to be made .
first of all , it is interesting to note that the relation between a dual expression and a feynman diagram is very much like the relation between a feynman diagram and an old - fashioned diagram . recall that a feynman diagram with @xmath22 vertices is made up of a sum of @xmath81 old - fashioned diagrams , corresponding to the @xmath81 possible ( real ) time orderings of its vertices .
similarly , a dual ` diagram ' consists of a sum of feynman diagrams , and they differ from one another by the _ proper - time _ orderings of their vertices . secondly ,
proper - time orderings in qed diagrams are very simple and natural , because there are the conserved charged lines along which the photon vertices can be proper - time ordered .
in contrast , in a neutral scalar @xmath82 theory , or in a pure gluon qcd , all lines are equivalent and there are no obvious ways to proper - time order -.5 cm 4.7 truein -13 cm .1 cm the vertices , especially in multiloop diagrams .
this is one of the difficulties one encounters in qcd .
thirdly , there is the question of how useful these dual expressions are . that naturally depends on the details of the diagrams one tries to sum , and how simple the resulting integrand @xmath83 is .
for example , consider the emission of scalar photons shown in fig .
3 , where @xmath6 and @xmath84 are massless but @xmath85 may be offshell . using ( 2.5 ) , and expressing the schwinger parameters @xmath86 as differences of the proper - time parameters @xmath87 , the exponent @xmath77 in the integrand @xmath83 becomes @xmath88 by using momentum conservation , this can also be written in a more symmetric form : latexmath:[\ ] ] where @xmath164 is given by the product of the non - type - a vertices and terms generated from their mutual contractions .
moreover , it is true that @xcite @xmath165 for a type - a vertex , so it does not matter whether @xmath130 in ( 3.14 ) is put before or after the @xmath152 s .
another useful relation to know is @xmath166 provided @xmath167 .
this relation can be used to show how the string - like vertex changes under a guage transformation , when @xmath168 is replaced by @xmath169 . in that case , remembering that @xmath170 , ( 3.3 ) and ( 3.16 ) give @xmath171
in the scalar - photon theory with interaction @xmath172 , the most general amplitude is given by ( 3.2 ) , with @xmath173 and @xmath77 given by ( 3.3 ) and ( 3.7 ) .
the schwinger parameters @xmath86 will be expressed as differences of the proper - time parameters @xmath174 , and the proper times will again be ordered along the charged lines as in the tree cases .
the only new problem here is where to begin the ordering in the case of a charged - scalar loop .
-0 cm 3.0 truein -5 cm since the origin of the proper time is never determined by the @xmath86 s , it can be chosen arbitrarily , say at the position marked ` 0 ' in figs . 5 and 6 .
we must now insert into ( 3.2 ) a factor @xmath175 for every charged loop , where the sum is taken over all the @xmath86 s in the loop .
so for fig . 5 ,
the @xmath86-integrations can be replaced by @xmath176}\equiv \langle\int d \tau _ { [ 12\cdots n]}\rangle\ , \ ] ] -0 cm 4.7 truein -9.5 cm .1 cm where @xmath177 for @xmath178 , and @xmath179 .
strictly speaking , there is an inconsistency in the definition of @xmath180}\rangle$ ] between ( 4.2 ) and ( 2.6 ) , but in practice ( 4.2 ) is always used for closed charged loops and ( 2.6 ) is always used for open charged lines .
one can now obtain dual amplitudes by summing over all photon permutations in exactly the same way as before .
for example , for fig . 5 and its permuted diagrams , the sum is proportional to @xmath181 for fig . 6 and its permuted diagrams , the sum is proportional to @xmath182 as in tree amplitudes , how useful such dual expressions are depends on the complexity of @xmath77 and @xmath130 in each case . in the eikonal approximation ( 2.11)(2.16 ) ,
the integrals can be carried out because @xmath98 is common for all integration regions _ and _ because it has a simple dependence on the proper times .
now the first condition is not hard to meet , if the diagrams to be summed have a high degree of symmetry .
for example , for the well - studied case of fig .
5 @xcite , eqs .
( 3.6 ) and ( 3.7 ) show that @xmath183 , and @xmath184 which has a symmetric form in the sense that it has the same functional form in all integration regions . however , one is still unable to evaluate the integral ( 4.3 ) analytically because of its relatively complicated @xmath5-dependences .
for the reason discussed at the end of sec . 2 ,
@xmath77 can be made symmetric or partially symmetric in many cases .
one way of seeing this is the following . if we modify fig . 3 by adding on a charged - scalar propagator at each end , then analogous to ( 2.10 ) one can produce a form of @xmath77 which is completely symmetrical in all the photon lines .
we have already seen that the @xmath77 in fig .
5 is completely symmetrical in all its photon lines as well .
now every feynman diagram can be built up from a number of open charged lines with its attached photons , and a number of one - charged - loop diagrams with its attached photons , by joining together pairs of photon lines .
mathematically , one obtains the resulting amplitudes by multiplying these @xmath77 s of the components , together with the propagators of the joined photon lines in the form of ( 2.5 ) , then carries out the momentum integrations of the joined photon lines .
since the dependences on these joined momenta are gaussian , such integrations can be carried out , and one again obtains a result of the form ( 3.2 ) , with @xmath185 , and with @xmath77 of ( 3.2 ) a function of the component @xmath77 s .
since the component @xmath77 s are symmetric in all the photon lines , they will still be symmetric in the remaining , unjoined , external photon lines , so in this way one can obtain a symmetric form for the final @xmath77 .
this mechanism for obtaining a symmetric form has been discussed recently @xcite in a slightly different language .
however , the symmetric form obtained this way is usually much more complicated than those obtained directly from ( 3.3 ) , ( 3.6 ) , and ( 3.7 ) .
it is so complicated that it is unlikely to be integrated analytically , nor will it lead to simpler numerical evaluations in most cases . see the end of sec . 2 for more discussions along these lines .
though one can generally produce other simpler symmetric forms , they are still not simple enough for the integrations to be carried out explicitly
. for these reasons there seems to be no particular advantage of having a symmetric form and we will not do so most of the time .
the most important ingredient for obtaining a dual expression is the presence of a conserved charged line along which to order the interaction vertices .
this is independent of the spin of the particles involved , hence one can obtain dual amplitudes in qed just as easily as those in the scalar - photon theory .
the factor @xmath186 in ( 3.2 ) is no longer 1 for qed . for scalar qed , for example
, it is made up of the product of the vertex factors @xmath187 and @xmath188 and their contractions .
( 3.11 ) and the paragraph above it .
this however makes no difference to the construction of the dual amplitude .
another minor complication is the presence of the seagull vertex @xmath188 .
this simply adds a dirac-@xmath189 function contribution to the integrand .
for example , in the compton scattering diagram fig . 7 , all that we have to do to accommodate the seagull vertex in fig .
7(c ) is to define the integrand of the dual amplitude to be @xmath190 where @xmath191 are the respectively factors for the three diagrams figs .
7(a ) , 7(b ) , and 7(c ) .
-2.5 cm 4.7 truein -5.5 cm .1 cm what distinguishes the dual expressions of qed from the scalar theory is gauge invariance , in that the diagrams in qed to be summed are connected by gauge invariance .
this means that the gauge - dependent parts of each diagram should no longer be there in the dual sum .
how the dual expression can be mathematically manipulated to achieve this purpose unfortunately depends on the details . at this moment
two general techniques are available to aid us .
one is the spinor helicity technique @xcite , where the reference momenta of the photons can be chosen appropriately to reduce the amount of gauge - dependent contributions .
the other is the integration - by - parts technique @xcite used in connection with the string - like operators appearing in ( 3.11 ) and ( 3.14 ) .
we mentioned previously that the relation between a dual amplitude and a feynman amplitude is analogous to the relation between a feynman amplitude and an old - fashioned amplitude . in each case
the former is not time - ordered , and the latter is ; the only difference being that it is proper - time ordering for the first pair and real - time ordering for the second . now with gauge theories , there is another parallel between these two cases : an old - fashioned diagram is not relativistically invariant but a feynman diagram is .
similarly , a feynman diagram is not gauge invariant but a dual diagram is .
let us consider two simple examples to illustrate these points .
first , consider the compton amplitude fig . 7 in scalar qed .
a propagator is added to each external charged line so that the vertices 1 and 3 in figs .
7(a ) and 7(b ) are of type - a , to enable the string - like vertex ( 3.11 ) to be used . without the seagull term fig .
7(c ) , the amplitude is not gauge invariant , so it definitely contains a non - trivial gauge - dependent part . as we shall see below
, one can actually manipulate the dual expression so that the seagull vertex seems to disappear , and the gauge - dependent parts from these three diagrams are no longer present .
the expression @xmath77 for diagrams ( a ) ,
( b ) , and ( c ) are respectively @xmath192 using ( 3.14 ) , the vertex factors in both cases become @xmath193\left[\epsilon ( p_3)\!\cdot\!d_3\right ] \equiv s\ .\ ] ] in their respective regions , one can replace @xmath194 and @xmath195 by @xmath196 however , since @xmath174-differentiations are involved in @xmath197 of ( 5.3 ) , @xmath198 and @xmath199 are not identical to @xmath200 and @xmath201 .
it can be checked by explict calculation that the former already contains the seagull vertex in fig .
hence the dual expression for fig . 7 is @xmath202\left [ \epsilon(p_3)\!\cdot\!d_3 \right]\exp(ip')\ , \ ] ] where @xmath203 the fact that the seagull vertex _ seems _ to have disappeared suggests that we have eliminated the gauge - dependent contributions altogether . to see that explicitly , use ( 3.17 ) . then under a gauge transformation , @xmath204 ( @xmath205 )
is changed into something proportional to @xmath206 .
the integral over @xmath207 in ( 5.5 ) can then be carried out to yield the boundary contributions at @xmath208 and @xmath209 , and hence the ward - takahashi identity .
since there is no trace of explicit cancellations needed at @xmath210 , it must mean that the gauge - dependent terms in the individual diagrams have now been eliminated .
another simple example one can mention is qed in the eikonal approximation , fig .
3 . in the soft photon limit for scalar qed ,
diagrams involving seagull vertex are not dominant because of the presence of one less propagator @xmath211 in the amplitude .
the cubic vertex factor is trivial in the soft photon limit , yielding @xmath212 .
the dual amplitude can therefore be read off from ( 2.17 ) to be @xmath213 it is gauge invariant to leading order in the photon momenta .
in the schwinger - parameter representation , qed diagrams differing from one another by the permutation of photon lines correspond to different proper - time ordering of the vertices , and can be formally summed into a single integral over a hypercubic region .
this sum is referred to as a dual sum because it is the field - theoretic counter part of a dual amplitude in string theory .
the relation between individual feynman diagrams and their dual sum is analogous to the relation between individual old - fashioned diagrams and their sum into a single feynman diagram . among other things , individual feynman diagrams in qed
are not gauge invariant but the dual sum is .
similarly , the individual old - fashioned diagrams are not lorentz invariant but their sum is . the dual sum allows formal manipulations between different feynman diagrams to be carried out , _
e.g. , _ by the integration - by - parts technique on string - like vertices . with appropriate approximations , such as the eikonal approximation
, explicit results may sometimes be obtained from the dual expression as well .
dual expressions for qcd are much more complicated to deal with and are not discussed in the present paper .
this research is supported in part by the natural sciences and engineering research council of canada and the qubec department of education.s e - mail address : [email protected] .
e - mail address : [email protected] . 1 z. bern and d.k .
kosower , phys .
b66 * ( 1991 ) , 1669 ; nucl . phys .
* b379 * ( 1992 ) , 451 .
2 z. bern and d.c .
dunbar , nucl .
b379 * ( 1992 ) , 562 . 3 p.
de causmaecker , r. gastmans , w. troost , and t.t .
wu , phys .
* 105b * ( 1981 ) , 215 ; nucl . phys .
* b291 * ( 1987 ) , 392 ; m.l .
mangano and s.j .
parke , phys .
200 ( 1991 ) , 301 ; r. gastmans and t.t .
wu , ` the ubiquitous photon ' , international series of monographs on physics , vol .
80 ( clarendon press , oxford , 1990 )
. 4 z. xu , d .- h .
zhang , and l. chang , tsinghua university preprints , beijing , china , tutp-84/4 , tutp-84/5 , tutp-84/6 . 5 c.s .
lam , nucl .
b397 * ( 1993 ) , 143 ; phys . rev .
* d48 * ( 1993 ) , 873 ; can .
j. phys . to appear .
6 s. parke and t. taylor , phys .
56 ( 1986 ) , 2459 .
7 g. mahlon , fermilab preprint fermilab - pub-93/327-t ( 1993 ) . 8 z.
bern , l. dixon , and d.a .
kosower , phys .
b70 * ( 1993 ) , 2677 ; z. bern , l dixon , d. c. dunbar , and d. a. kosower , slac preprint slac - pub-6415 . 9 c.s . lam and j.p .
lebrun , nuovo cimento * 59a * ( 1969 ) , 397 .
g. veneziano , nuovo cimento * 57a * ( 1968 ) , 190 .
m. strassler , nucl . phys . *
b385 * ( 1992 ) 145 ; slac preprint slac - pub 5978 ( 1992 ) .
schmidt and c. schubert , desy preprint desy - hd - thep-94 - 7 ( hep - ph/9403158 ) ( 1994 ) . | string theories should reduce to ordinary four - dimensional field theories at low energies .
yet the formulation of the two are so different that such a connection , if it exists , is not immediately obvious . with the schwinger proper - time representation , and the spinor helicity technique , it has been shown that field theories can indeed be written in a string - like manner , thus resulting in simplifications in practical calculations , and providing novel insights into gauge and gravitational theories .
this paper continues the study of string organization of field theories by focusing on the question of local duality .
it is shown that a single expression for the sum of many diagrams can indeed be written for qed , thereby simulating the duality property in strings .
the relation between a single diagram and the dual sum is somewhat analogous to the relation between a old - fashioned perturbation diagram and a feynman diagram .
dual expressions are particularly significant for gauge theories because they are gauge invariant while expressions for single diagrams are not
. = -1 cm = -1 cm mcgill/94 - 21 + ( may , 1994 ) + hep - th/9409083 epsf = cmr10 = cmbx10 * string organization of field theories : + duality and gauge invariance * + + + |
the fact that exciton condensation and lasing are fundamentally similar can be seen by looking at the basic mathematics of exciton condensation . in second quantization language , the creation operator for an exciton with center - of - mass momentum @xmath0 is ( see , e.g. , ref .
@xcite , section 11.2.1 ) , @xmath1 where @xmath2 is the fermion creation operator for an electron in band @xmath3 with momentum @xmath4 , and @xmath5 is the orbital wave function for the electron - hole relative motion . in other words ,
the creating an exciton corresponds to creating a superposition of electron - hole pairs in the conduction band and valence bands .
a bose - einstein condensate corresponds to a coherent state of excitons in a single @xmath0-state , which we will assume is @xmath6 : @xmath7 where @xmath8 is the phase of the condensate , and @xmath9 is the number of excitons in the condensate .
let us look at this formalism in the limit of frenkel excitons .
frenkel excitons correspond to excitons with electron and hole on the same lattice site , as opposed to wannier excitons , in which the exciton and hole orbit each other at some distance .
we write the fermion creation operator as the fourier transform of a spatial creation operator for a single lattice site @xmath10 : @xmath11 where @xmath12 is the number of lattice sites .
the creation operator ( [ ccreate ] ) is then @xmath13 the frenkel limit corresponds to picking @xmath14 constant @xmath15 , in which case the term in the parentheses , which is the real - space wave function for the electron - hole relative motion , equals @xmath16 .
putting this into our definition of @xmath17 gives us @xmath18 i.e. , a superposition of excitations at all lattice sites .
creating a macroscopic number of @xmath6 excitons in one state involves terms of the form @xmath19 this is a superposition of all possible ways to have @xmath9 excitons in @xmath12 lattice sites .
the expectation value @xmath20 gives the probability of a lattice site @xmath10 being in the excited state .
if we pick site @xmath10 , then in the product ( [ product ] ) there are @xmath3 terms which will create an exciton at that site . for each of these terms the number of ways of picking the remaining terms so as to not have double fermion creation operators ( which vanish ) is @xmath21 in the limit @xmath22 , we ignore the possibility of double occupation and approximate @xmath23 each product term created in @xmath24 will have @xmath25 matching terms in the term with the same order of @xmath3 in @xmath26 .
the expectation value is therefore @xmath27 in other words , the probability of a given lattice site being in the excited state is equal to the number of excitons in the condensate divided by the total number of lattice sites.if the number of excitons increases , at some point the possibility of double occupation becomes significant , and will cause this relation to become sublinear with @xmath9 .
this is known as `` phase space filling . ''
the calculation for the dipole moment term @xmath28 is similar , except that this term changes the @xmath3 term in the coherent state sum to the @xmath29 term .
we therefore have @xmath30 this is proportional to the amplitude of the condensate wave function .
the total dipole moment of the solid is @xmath12 times this .
this corresponds to a macroscopic radiating dipole moment which will emit coherent light .
let us now switch to thinking about a standard laser .
in the theory of lasing in an ensemble of independent two - level oscillators , we start with a single bloch oscillator coupled to a coherent electromagnetic wave : @xmath31 where @xmath32 .
the coefficient @xmath33 determines the degree of excitation , which is typically much less than unity . for an ensemble of identical oscillators
, we repeatedly create identical states at all the different sites : @xmath34 the expectation value @xmath35 is @xmath36 , which we can set equal to @xmath37 . as shown in ref .
@xcite , section 11.2.3 ( and discussed earlier in ref .
@xcite ) , the state ( [ laserbcs ] ) , which has the form of a bcs wave function , is equivalent to a coherent state of the form ( [ coh ] ) , where we equate @xmath38 if we choose @xmath39 and @xmath40 , in the limit where @xmath41 when phase space filling is negligible , we have exactly the same state as our frenkel exciton example , above , with probability of excitation of a single lattice site equal to @xmath37 . the case of a coherent electromagnetic wave passing through a transparent medium with a dielectric constant ( index of refraction ) produces exactly the same state of matter . as with a laser , we model the system as an ensemble of two - level oscillators in phase with each other ( see , e.g. , sections 7.1 and 7.3 of ref .
@xcite ) . in this case , the frequency of the wave is well below the electronic resonance , so that the radiation from the polarized medium gives a phase shift to the electromagnetic wave , and a change in phase velocity , but no absorption or gain .
we thus have three physical scenarios which all lead to a state of matter which consists of an ensemble of electronic excitations in phase with each other : excitonic condensation , lasing , and linear refraction of a wave in a dielectric medium . the first two involve spontaneous coherence , while the last has electronic coherence because it is driven by an external electromagnetic field .
some people have used the term `` driven condensate '' to refer to coherence of a medium driven by an external field , but this terminology is not very helpful , because it would have to also include every sound wave produced by a loudspeaker and every radio wave driven by an antenna .
it is important to notice , though , that the polarization state of the underlying medium is fundamentally similar in all three cases ; all are states of coherent polarization of matter .
the differences arise in other properties , namely , how we get to the state , how stable the state is , the degree of inversion , and what is going on with other electronic states not participating in the coherence .
we will discuss the differences of exciton - polariton bec and standard lasing below , but before we do that , let us consider another connection . as discussed above ,
every bec state is a coherent state of bosons of the form ( [ coh ] ) .
a general property of coherent states with large @xmath9 is that they have exactly the same properties as classical wave states ; in fact , in the modern understanding , all classical waves such as sound waves and electromagnetic waves are actually quantum - mechanical coherent states with high occupation number of bosons ( phonons and photons , respectively ) . in the case of a standard bec ,
the behavior of the coherent condensate is described by the gross - pitaevskii equation , @xmath42 which becomes the ginzburg - landau equation if the time derivative is replaced by @xmath43 .
this is also called the nonlinear schrdinger equation , because it has exactly the same form as a single - particle schrdinger equation with wave function @xmath44 and a nonlinear potential energy that is proportional to the square of the amplitude of the wave function , @xmath45 . in the context of bec ,
the wave function @xmath44 is normalized by @xmath46 , where @xmath9 is the total number of particles in the condensate , and @xmath47 is the dimensionality .
the gross - pitaevskii equation is often viewed as a heuristic assumption , but it can be justified fairly rigorously in terms of the underlying microsopics , as discussed in ref .
@xcite , section 11.1.3 .
it is simply a consequence of the condensate being a coherent state , and the particles have a two - body interaction that leads to local potential energy proportional to the particle density .
the fact that a condensate can be described by a classical wave equation with a nonlinear term has also led to other qualms that a polariton bec is no different from a coherent classical optical system .
we can see that the two have a fundamental connection by showing that we can also derive the gross - pitaevskii equation starting with maxwell s equations .
we start with maxwell s wave equation in a nonlinear isotropic medium , @xmath48 where @xmath49 is the standard nonlinear optical constant , and we ignore frequency - mixing terms in the general @xmath50 nonlinear response .
we write a solution @xmath51 where @xmath44 is an amplitude which may vary in time and space .
we write this envelope amplitude suggestively as @xmath44 because we will see that it plays the same role as the matter - wave @xmath44 in the gross - pitaevskii equation . keeping only leading terms in frequency ( known as the slowly varying envelope approximation )
, we have for the time derivative of @xmath52 , @xmath53 and for the time derivative of the nonlinear term @xmath54 the standard polariton structure used a planar or nearly - planar cavity to give one confined direction of the optical mode .
we therefore distinguish between the component of momentum @xmath55 in the direction of the cavity confinement and the momentum @xmath56 in the two - dimensional plane perpendicular to this direction .
we therefore write @xmath57 the full maxwell wave equation ( [ maxwell ] ) then becomes @xmath58 since @xmath59 , this becomes @xmath60 near @xmath61 , we can approximate @xmath62 which gives an effective mass for the photon motion in the plane .
this does not take into account the renormalization of the effective mass due to the coupling with the exciton states , which typically makes the effective mass about a factor of two heavier . to account for this
, we would have to take into account the dependence of the index of refraction @xmath3 on frequency due to the exciton resonance , as discussed in ref .
@xcite , section 7.4 . neglecting this correction ,
for the first term on the right - hand side , we approximate @xmath63 so that we have @xmath64 which we can rewrite as @xmath65 this is the gross - pitaevskii equation , or nonlinear schrdinger equation . note that although the maxwell wave equation is second order in the time derivative , this equation is first order in time derivative , as in a typical schrdinger equation .
one useful result of this derivation is that we can write the nonlinear @xmath49 value in terms of experimentally measured parameters . to do this we first need to calibrate the electric field amplitude in terms of the density of particles . from basic photon theory ( see ref .
@xcite , section 4.4 ) the number of photons is related to the electric field amplitude by @xmath66 which for our two - dimensional system can be written as @xmath67 where @xmath68 is the area and @xmath47 is the effective thickness .
if we switch @xmath44 in ( [ gp2 ] ) from standing for electric field amplitude to standing for a number density wave function , the nonlinear term therefore corresponds to a potential energy linearly proportional to the density , with constant of proportionality equal to @xmath69 we can use this result to equate the measured coefficient of linear shift of the energy of the polaritons , measured from the shift of the spectral line , with an effective @xmath49 value .
typical values for polariton condensates are 1 mev of shift for a density of @xmath70 @xmath71 . using a typical microcavity length of 400 nm and material constants for gaas , we obtain @xmath72 which is equivalent to @xmath73 esu . by comparison ,
typical nonlinear coefficients for strongly nonlinear media are around three orders of magnitude lower @xcite .
this is one of the main appeals of polariton condensates as optical systems they have world - record nonlinearities , allowing , for example , optical parametric oscillator ( opo ) behavior with continuous laser pump powers of a milliwatt or less , compared to standard broadband opo systems which require pulsed lasers with instantaneous power at least six orders of magnitude higher .
the tradeoff is that polariton systems only work for a narrow range of wavelengths ; the huge nonlinearity is obtained by operating near a sharp exciton resonance .
this can be done with resonances of single atoms , also , but to do that requires ultra - high vacuum and extremely good laser frequency stability .
the fact that that polariton condensate gross - pitaevskii equation maps to the nonlinear maxwell wave equation in this system should not be cause to doubt the validity of the bec description . every condensate is described by a classical wave equation , because the gross - pitaevskii equation is a classical wave equation .
it does tell us , however , that some experimental results with polariton condensates do not intrinsically distinguish between condensation and lasing , or more generally , a coherent optical field .
for example , quantized vortices are fundamentally a result of having a single - valued , coherent wave .
since this occurs also in the optical modes of a laser , quantized vortices can also be seen in a standard laser or vcsel @xcite . in the same way , in - plane phase coherence of the wave function
is also seen in spatially extended laser modes from a vcsel @xcite .
however , the central core radius of a vortex of a polariton bec , in which the condensate is depleted , is determined by the healing length of the condensate @xmath74 , which is determined by the interaction strength of the particles stronger interaction gives shorter healing length ( see ref .
@xcite , section 11.1.4 ) . in vortices of laser modes , the vortex structure
is determined by the spatial dimensions of the laser modes .
to see the difference between a polariton condensate coherent state and a standard laser , we start by examining how these states arise . figure [ laser1](a ) shows the standard band picture of a semiconductor laser .
electrons from the valence band are pumped high up into the conduction band through what is typically an incoherent process , and then both the electrons and holes created by the pumping fall in energy to near the band edges ( holes float upwards as they lose energy , in this picture ) .
this energy relaxation is an incoherent process , so that even if a coherent pump was used to excite the carriers , by the time they arrive in the lowest energy states , they have lost all memory of the pump laser coherence .
fundamentally , the semiconductor laser is no different from an ensemble of isolated atoms , e.g. atoms in a gas or dopant atoms in a transparent matrix , with two occupied levels and two unoccupied levels .
the pumping of the electrons into the upper states leads to _ inversion _ , that is , higher probability of carriers in the upper state than in the lower state of the two levels involved in the light emitting transition ( so - called `` negative temperature '' ) .
this leads to a spectral range for electromagnetic waves with _ gain _ , that is , amplification of coherent electromagnetic waves . as shown in ref .
@xcite , section 11.3.1 , a simple model of a laser as a homogeneous ensemble of pumped two - level atoms leads to the result @xmath75 where @xmath76 is the dielectric constant and @xmath68 and @xmath77 are two terms which depend on the pumping rate and oscillator strength ; @xmath68 is positive for net pumping rate exceeding the spontaneous emission rate , and @xmath77 has the same sign as @xmath68 . if the system is initially incoherent , but there is some small fluctuation that gives a coherent seed , this coherence will be amplified by the @xmath68 term until there is a macroscopic coherence .
the amplification will eventually be limited by the nonlinear @xmath77 term , which arises from the physical constraint that the upper level of the two - level atoms can not have higher than unity occupation , and rabi flopping will depopulate the upper level if the coherent field grows too strong .
this is equivalent to saying that there is a nonlinear term due to _ phase space filling _ , that is , filling of the upper states due to the pauli exclusion principle for the fermion electrons .
one can see from this that the onset of lasing is a type of spontaneous symmetry breaking , since we can not predict the exact value of the complex amplitude of the seed that will be amplified by ( [ lasegain ] ) .
the same type of spontaneous symmetry breaking occurs in bose - einstein condensation .
a calculation to be presented elsewhere @xcite gives for the complex amplitude @xmath78 of a condensate @xmath79 \delta(e_{k}+e_{k_2}-e_{k_3}-e_{k_4}),\nonumber\\ \label{bosegain } \end{aligned}\ ] ] where the @xmath80 are the average occupation numbers of the excited @xmath4-states , the same occupation numbers that go into the quantum boltzmann equation for the increase of the population in the condensate ( see ref .
@xcite , section 4.8 ) .
the first term is normally small relative to the second , since it involves a product of two occupation numbers while the second involves only one , but when the number of particles in low - energy states becomes large , it can lead to gain of the coherent amplitude of the condensate .
( the topic of the onset of phase coherence in bec has been a longstanding problem in the theory of bec , addressed by numerous authors ( e.g. , refs .
it is not essential that a collisional interaction be the means of establishing condensation ; condensation can also occur by emission and reabsorption of phonons ( see ref .
@xcite , section 8.2.3 ) .
but in every case , the stimulated scattering @xmath81 factors , which connect the occupation of one particle state to another , drives the onset of the phase coherence . as illustrated in fig .
[ laser1](b ) , the gain in the system occurs in the coupling of the continuum of excited states to the emitting states , not in the emission process itself . these two different relations ,
( [ lasegain ] ) and ( [ bosegain ] ) , tell us a lot about both the similarities and the differences between lasing and excitonic condensation .
we see that both are examples of spontaneous coherence . in the spontaneous symmetry breaking of a laser ,
the control parameter is not the temperature @xmath82 , as the transition is not a thermodynamic phase transition ; instead the control parameter is the incoherent pumping rate which goes into the gain constant @xmath68 .
nevertheless , spontaneous symmetry breaking is not unique to bec .
the _ onset _ and _ stability _ of standard lasing arise from the net gain from incoherent pumping , and the coherence arises from the stimulated emission of photons , which is taken into account in ( [ lasegain ] ) by the @xmath83 term .
this implies that a standard laser must have _ inversion _ , and the stimulated emission occurs _ only in the coherent state_ whatever state has many photons will be amplified if there is gain . in a standard laser , that photon state is defined by the mirors of the cavity , which recirculates photons to suppress the loss rate .
electronic excitations not resonant with that state will undergo transitions to pump that state only incoherently . in a polariton condensate , as with any condensate , the onset and stability of the coherent amplitude come fundamentally from the _ interaction of the particles in nearby states_. the two terms which amplify the coherence in ( [ bosegain ] ) come from the occupation numbers of the non - condensed states .
if all of the occupation numbers @xmath84 , @xmath85 , and @xmath86 are all small compared to unity , then there will always be net dephasing . if @xmath85 and @xmath86 are large compared to unity , however , while @xmath84 is small , then there will be amplification .
onset of condensation therefore requires not just that the condensate state itself is highly occupied , but that nearby states also have large ( though not necessarily macroscopic ) occupation .
the coherence of the system is shared as a collective property of particles in many states .
figure [ pileup ] shows how the states near to the condensate state in a polariton system are also highly occupied .
factors not included .
green curve : full calculation including final states @xmath87 factors .
there is not only a peak ak @xmath88 , there is also a pileup of particles in low - momentum states . from ref .
@xcite.,scaledwidth=80.0% ] in a condensate of particles with finite lifetime , there is a pump that incoherently creates new particles to compensate for the loss of the particles , but the stimulated process that gives rise to coherence is in the scattering of these incoherently created particles with each other , not in gain in the recombination process .
inversion is not only not necessary for polariton condensation , it actually kills polariton condensation , because inversion corresponds to pauli state filling of the bands , which prevents pairing of the carriers into bosonic excitons , and thus shuts off the process of bosonic stimulated scattering which leads to condensation .
one can think of the two limits of the bec - lasing continuum in terms of what drives the coherence . in the case of lasing
, we can imagine starting with a very small coherent electromagnetic field which drives the two - level oscillators , similar to the way that an electromagnetic wave drives the oscillators which give the dielectric response of the medium .
this small coherent electromagnetic wave drives a small coherent polarization of the medium , which then radiates , and this radiation is amplified by the gain due to the inversion . in an excitonic bec
, the coherent polarization of the medium drives the electromagnetic radiation .
the coherence in the polarization of the medium leads to emission of coherent radiation , but would still occur even if there were no radiation ; which can be the case , for example , when the emission of radiation is forbidden by the symmetry of the semiconductor bands @xcite or by a barrier between the electrons and holes , as in the case of excitons in double quantum wells @xcite .
we imagine starting with a small coherent polarization of the medium , which then is amplified by pulling in carriers from the incoherent excited population of the medium , and then this amplified coherent polarization can emit radiation .
in other words , in both systems there are electromagnetic degrees of freedom and electronic degrees of freedom , but in a laser the coherence of the electronic states of the medium is driven by the electromagnetic coherence , as in a linear coherent wave passing through a dielectric medium . in an excitonic condensate , the coherence of the medium arises first , and coherent radiation is emitted by this polarization as a byproduct in some cases .
the case of an exciton - polariton condensate is a middle ground , in which electromagnetic and electronic degrees of freedom participate equally . as we will see below , the electronic degrees of freedom still play an important role .
a standard exciton - polariton microcavity structure used for condensation is different from a laser in having _ strong coupling _ between exciton and photon states .
strong coupling is defined as a rabi splitting between the upper and lower polariton states which is large compared to the line with of either line .
( for a review of the basic properties of microcavity polaritons , see ref .
the term `` rabi frequency '' for the splitting between the states is no accident it is equal to the standard rabi frequency @xmath89 in a two - level oscillator , where @xmath90 is the dipole matrix element coupling the two states , when @xmath91 is the electric field amplitude of a single photon of frequency @xmath92 in the cavity , and @xmath92 is the exciton resonance .
when the system is in strong coupling , the polariton is the proper eigenstate , with an equal , or near - equal , superposition of an exciton and a photon .
thus , if spontaneous coherence occurs in this strong coupling regime , we may say that the electronic polarization is playing an important role in driving the transition , as in an excitonic condensate . in many experiments , the system may be in strong coupling at low excitation density but revert to weak coupling at high density .
the reason is that the dipole matrix element @xmath90 is in general a function of carrier density , and can become greatly reduced at high density due to pauli phase space filling and carrier screening essentially , the excitons no longer are well - defined bound states at high density , so that the system becomes a gas of uncorrelated electrons and holes , i.e. a plasma , and the dipole coupling between photons and free electrons and holes is much less than the coupling of photons to excitons .
the transition from excitons to free carriers is sometimes called the `` mott transition , '' although the nature of this exciton - plasma transition in semiconductors is of quite different nature from that considered in cold atom gases and in doped solids @xcite . if the system reverts to weak coupling , this will result in closing of the gap between the upper and lower polaritons at the zero detuning point ( the point at which the excitons and photons have exactly the same energy ) .
it is not always easy to know when this has happened , however . if the bare exciton and photon states are not perfectly equal in energy
, then there will be a splitting between them just due to the energy difference of the bare states .
this energy gap will not close down when the system goes into weak coupling .
a better test of strong coupling is to measure the effective mass of the lower polaritons via angle - resolved far - field emission ( this has been done in several experiments ) . in strong coupling
, the effective mass of the polaritons will be twice the mass of the bare cavity photon mode @xcite .
one way to allow higher polariton density without reverting to weak coupling , which has been adopted by many groups , is to use multiple quantum wells in a cavity instead of just one .
this greatly reduces the amount of phase space filling , since the exciton component of any one polariton is shared among many wells .
typical samples now use 12 or 16 wells placed at antinodes of the confined cavity photon mode @xcite . in such samples it is possible to observe a clear distinction between polariton condensation and lasing
. figure [ fig.stress ] shows the emission energy from polaritons in a microcavity in which the exciton energy was shifted using stress to tune the exciton resonance frequency , while leaving the photon cavity frequency essentially unchanged .
as seen in the reflectivity data , the lower polariton mode follows the exciton shift with stress , since it has an excitonic component when there is strong coupling . at high density ,
when the polariton condensate appears , the coherent emission still shifts with stress to follow the exciton mode .
if the system were in a lasing mode with weak coupling , the emission would occur at the cavity photon mode , indicated by the dashed line in fig .
[ fig.stress ] .
( actually , at high density this cavity mode is slightly red shifted from its low - density value , due to a density - dependent change of the index of refraction @xcite , but this red shift is unaffected by the stress if the system is in weak coupling . )
as seen in this figure , at very large detuning , the coherent emission finally jumps up to near the bare cavity photon frequency , as the system reverts to weak coupling , and the system is at the point in standard vcsel lasing mode .
the excitation density required to get coherent emission is not the same in the polariton condensate state and the lasing state .
the polariton condensate in general requires much lower excitation power , because it does not require inversion , as discussed above . in the sample used for the data of fig .
[ fig.stress ] , the threshold for polariton condensation in the limit is about 6 times smaller than the threshold for lasing at large positive detuning at the same place on the same ( recall that via stress tuning , the detuning can be varied for the same spot of the sample . )
@xcite showed that polariton condensation could have a threshold as much as a factor of 20 lower than the lasing threshold in the same sample , when the detuning was changed by moving the laser spot to a different place in the sample with different cavity thickness .
since the lasing threshold is higher in power , it is possible to push the system from polariton condensation to standard lasing simply by turning up the excitation power .
figure [ fig.series ] shows a sequence of images which give the angle - resolved spectrum of the polariton emission from the same microcavity sample used for fig .
[ fig.stress ] when the stress is chosen to give nearly zero detuning , and the excitation power is increased .
the angle of emission of the externally emitted photons maps directly to the momentum of the polaritons in the plane of their motion . comparing the 23.8 mw data to the 8.4 mw data
, one sees that at first , as the power is increased , the polariton spectrum shifts to the blue , and broadens . both of these effects are expected for particles with repulsive interactions .
because these spectra are time averaged , there is also a contribution to the broadening if the blue shift fluctuates in time due to laser power fluctuations .
the 31.7 mw data shows the spectrum near the polariton condensation threshold .
the spectrum has narrowed in energy , and also has become narrower in momentum space .
the energy width here is limited by instrumental resolution ; ref .
@xcite has recently showed that extreme narrowing , corresponding to coherence times of more than 200 ps , can be measured in this type of system when a vert stable laser with low intensity fluctuations is used .
if the power is increased further , the spectrum broadens again , and shifts even more strongly to the blue .
both of these effects are related to phase space filling at high density . as discussed above
, phase space filling reduces the coupling , and therefore reduces the rabi splitting between the upper and lower polaritons , which shifts the lower polariton up .
the increased broadening is also likely to do strong dephasing due to free electrons and holes in the system .
when the system reaches 235 mw , however , a second spectral narrowing is seen .
this transition corresponds to standard lasing , and the energy of the photons corresponds to the bare cavity photon energy . at all high powers
, the coherent emission stays at this energy , pinned at the cavity photon energy as expected for a laser .
we thus see that the two different transitions , polariton condensation and standard lasing , are easily distinguishable experimentally , leading to two separate line narrowings at different excitation powers in the same place on the sample .
bloch and coworkers @xcite have also seen similar behavior by varying the laser power . in general
, in all these experiments , it is quite easy to see when the system has reverted to lasing , because the emission energy pops up to the bare cavity photon energy .
in the past few years , the number of polariton condensate experiments has exploded , with a number of demonstrations of effects analogous to those of atomic condensates , superconductors or liquid helium . in one sense
, it could be said that the main interest of these effects in atomic condensates , superconductors , and helium is to show wave - like behavior of matter ; i.e. , these effects all follow from being able to write down a gross - pitaevskii equation which treats the condensate as a classical wave . in the case of polaritons ,
the wave - like behavior is perhaps less surprising , because light is already seen as a wave .
but unless one disbelieves quantum mechanics , all matter is described as waves .
the significance of condensation , with atoms , cooper pairs , and polaritons , is that a system which normally has strong dephasing can spontaneously acquire a macroscopic coherent amplitude due to the collective effects .
polariton systems , like atomic systems , have a threshold of low temperature and high density at which the coherence due to interparticle interaction defeats the dephasing processes .
some of these effects with polaritons have analogs in lasing systems , but their behavior is different .
figure [ fig.jj ] shows an example of oscillations in the emission of two traps containing polariton condensates , connected by a thin tunneling barrier .
similar results have been seen by bloch and coworkers @xcite .
the period of the oscillations depends on the chemical potential difference between the two traps , as in a standard josephson junction experiment .
these oscillations look superficially like the beating of two laser modes .
but in the case of laser mode beating , the two states which interfere have discrete energies well separated in frequency from other laser modes .
the stability of the oscillations comes from the fact that there are no other nearby states which have gain . in the case of a polariton josephson junction ,
as discussed above in section [ sect.diff ] , the stability comes from the interaction of the particles in a continuum of @xmath4-states of the polaritons . at first
, it may seem that the josephson junction effect in a superconductor involves very different physics from a josephson junction of polaritons .
the superconductor case involves two fermi levels of the electrons , and the oscillation frequency depends on the difference between these fermi levels . but recall that a bcs superconductor is just the high - density limit of bec ( see , e.g. , ref .
@xcite , section 11.2.3 ) . the fermi level in a bcs superconductor
is determined by the pair wave function of the condensed pairs .
a condensate of excitons or polaritons also has a spread of @xmath4-states which is determined by the pair wave function , but the density is low enough that there is no phase space filling which produces a fermi level in this case . in both cases ,
once a condensate of pairs is formed , the oscillation frequency of a josephson junction depends on the chemical potential between the two sides , which depends on the sum of any externally applied potential plus the potential energy due to interactions of the particles . because the chemical potential difference between the two sides depends not just on the static frequency difference between the two traps , but also on the polariton - polariton interactions and on interactions of polaritons with excitons in higher energy states , the period of the oscillations changes in time for the polariton condensate as the density drops due to recombination
, the strength of these interactions drops , and the frequency difference between the traps is shifted .
again , one could argue that a laser system with a strong nonlinear shift of the index of refraction could also show this effect , but as discussed in section [ sect.gp ] , the exceedingly high value of the nonlinear term in the condensate gross - pitaevskii equation is what makes the polariton system special .
this high value of the nonlinear term comes from dressing the photons in the system with an excitonic part , which gives a strong interaction between the particles through the long - range coulomb interaction .
the cost of this is that the excitonic part also leads to much higher intrinsic dephasing due to scattering with phonons and impurities .
but the bosonic stimulated scattering of the polaritons can overcome this dephasing , leading to phase coherence which is like that of a laser but which shows nonlinear effects at much lower carrier densities .
there are many other experiments with polaritons that can be described by the formalism of condensation , including vortices @xcite , soliton propagation over long distances @xcite , and propagation without scattering in wires ( channels ) @xcite .
the bottom line is that these systems are easily described by the language of condensation , especially when taking into account the interplay between the condensate and excited states in the continuum @xcite , and insisting on the language of laser optics only adds difficulty .
the condensation paradigm has become dominant because it is successful in describing the experiments . when the polariton condensate is pumped resonantly by direct coupling to an external laser beam , the choice of language is more ambiguous .
as we have seen , once a condensate is formed , it can be described by a classical wave equation , the gross - pitaevskii equation .
one can also create such a state by direct pumping with a laser at the same frequency as a polariton state .
in this case the system can be far from equilibrium , and the contribution to the optical effects from particles in incoherent excited states can be unimportant . in general , many of the characteristics of an equilibrium condensate , such as the linear bogoliubov excitations , persist even as the system moves into a not - fully - equilibrated state @xcite .
recently , experiments on `` bose condensation of photons '' have been reported @xcite .
what the authors mean by this is that they created a system with approximate number conservation of photons in a cavity over a short time , similar to the case of a polariton condensate in a microcavity , but in the limit of weak coupling , so that the particles were purely photons . in this limit ,
the photons thermalized not by collisions with each other , since they are effectively non - interacting , but instead by being absorbed and then re - emitted by dye molecules in the cavity .
the dye was chosen such that one photon absorbed led to one photon emitted at a different energy , with very little nonradiative loss ; the energy of the emission depended on the temperature of the dye .
as in the case of strong coupling , the photons have an effective mass determined by the cavity properties . at high photon density ,
the energy distribution of the photons thermalized and then showed a peak at @xmath93 as in the case of a polariton condensate , as shown in fig .
[ fig.pbec ] .
this system is not the same as standard lasing , because the number of photons was approximately conserved .
however , it is not really the same as a standard condensate , either . as discussed above ,
one of the main properties of standard condensation is spontaneous phase coherence of the condensate , which comes about via interactions of the particles . in the case of the `` photon condensate , ''
there is no interaction of the particles which could lead to phase locking .
one thus has the situation that the number conservation at a given temperature make the particles want to occupy the lowest available state , but the interaction with the incoherent phonon bath may prevent them from actually becoming a phase coherent condensate . in fig .
[ fig.pbec ] there is a relatively broad width of the peak at @xmath93 , with a full width at half maximum of around 2 nm , or 7 mev , which corresponds to a coherence time of about 100 fs . by contrast , coherence times of hundreds of picoseconds have been observed in polariton condensates @xcite .
this broad width of the condensate peak in ref .
@xcite may be limited by the experimental spectral resolution ; more recent results @xcite using interferometry indicate a coherence time on the order of tens of picoseconds .
the experiments with photons in weak coupling are therefore an interesting intermediate case in which the system is prevented from acquiring optical coherence either by lasing or polariton condensation in strong coupling .
the system acts in many ways like an ideal bose gas , which , as is well known , can not undergo true bose - einstein condensation @xcite .
both standard lasing and polariton condensation are states in which there is a coherent polarization of the electrons in the matter , leading to spontaneous phase coherence of the electron polarization and the light emission . in the case of a condensate , the spontaneous coherence comes about due to a thermodynamic phase transition ( even when not fully equilibrated ) in which temperature is the controlling parameter , while in the case of a laser , the spontaneous coherence comes about far from equilibrium in a transition in which the pump power is the controlling parameter ; a laser requires population inversion , while polariton condensation is inhibited by population inversion .
experiments have been performed in which each of these transitions can be seen separately in the same physical system , with very different behavior . in many ways
the recently reported `` photon condensate '' acts like a polariton condensate , with an effective mass that gives a non - planckian energy distribution , but the nearly non - interacting nature of the photons in this case means that they can only thermalize via incoherent interactions with an dye medium , which may lead to stronger dephasing . | the similarities and differences between polariton condensation in microcavities and standard lasing in a semiconductor cavity structure are reviewed .
the recent experiments on `` photon condensation '' are also reviewed .
polariton condensation and lasing in a semiconductor vertical - cavity , surface - emitting laser ( vcsel ) have many properties in common .
both emit coherent light normal to the plane of the cavity , both have an excitation density threshold above which there is optical gain , and both can have in - plane coherence and spontaneous polarization .
is there , then , any real difference between them ?
should we drop the terminology of bose - einstein condensation ( bec ) altogether and only talk of polariton lasing ?
the best way to think of this is to view polariton condensation and standard lasing as two points on a continuum , just as bose - einstein pair condensation and bardeen - cooper - schreiffer ( bcs ) superconductivity are two points on a continuum . in the bec - bcs continuum ,
the parameter which is varied is the ratio of the pair correlation length to the average distance between the underlying fermions ; bec occurs when the pair correlation length , i.e. , the size of a bound pair , is small compared to the distance between particles , while bcs superconductivity ( which has the same mathematics as the excitonic insulator ( ei ) state , in the case of neutral electron - hole pairs ) occurs when the pair correlation length is large compared to the distance between the particles . in the condensation - lasing continuum ,
the parameter which is varied is the ratio of the recombination time of the electron - hole pairs to the interaction time between pairs ; in the exciton condensate limit , the recombination lifetime is essentially infinite compared to the time scale for interactions between the excitons , while in the case of lasing , the lifetime for recombination is short compared to the interaction time , and in fact , there may be no interaction between the pairs at all .
the case of the polariton condensate lies somewhere between these two limits , since the polaritons are mixed states of excitons and photons . in these different limits ,
there are various properties of the bec and standard lasing states that allow them to be distinguished . in this article
we will review the connections between the two states and also discuss the experiments that show the differences . |
some of the computations were performed on the supercomputers at the information technology center , nagoya university and at the research center for computational science , institute for molecular science .
this work was supported , in part , by grants - in - aid for scientific research on innovative areas ( fluctuations and biological functions " ) and for the next generation super computing project , nanoscience program from the ministry of education , culture , sports , science and technology ( mext ) , japan .
99 a. mitsutake , y. sugita , and y. okamoto : biopolymers ( peptide science ) * 60 * ( 2001 ) 96 .
b. a. berg and t. neuhaus : phys .
b * 267 * ( 1991 ) 249 .
b. a. berg and t. neuhaus : phys .
* 68 * ( 1992 ) 9 .
u. h. e. hansmann , y. okamoto , and f. eisenmenger : chem .
phys . lett . *
259 * ( 1996 ) 321 .
n. nakajima , h. nakamura , and a. kidera : j. phys .
b * 101 * ( 1997 ) 817 .
k. hukushima and k. nemoto : j. phys .
* 65 * ( 1996 ) 1604 .
e. marinari , g. parisi , and j. j. ruiz - lorenzo : in _ spin glasses and random fields _
young ( world scientific , singapore , 1998 ) p. 59 .
y. sugita and y. okamoto : chem .
phys . lett .
* 314 * ( 1999 ) 141 .
a. p. lyubartsev , a. a. martsinovski , s. v. shevkunov , and p. n. vorontsov - velyaminov : j. chem . phys .
* 96 * ( 1992 ) 1776 .
e. marinari and g. parisi : europhys . lett .
* 19 * ( 1992 ) 451 .
a. mitsutake and y. okamoto : phys .
e * 79 * ( 2009 ) 047701 .
a. mitsutake and y. okamoto : j. chem . phys .
* 130 * ( 2009 ) 214105 .
a. mitsutake : j. chem
* 131 * ( 2009 ) 094105 .
h. okumura and y. okamoto : chem .
* 383 * ( 2004 ) 391 .
h. okumura and y. okamoto : phys .
e * 70 * ( 2004 ) 026702 .
h. okumura and y. okamoto : chem .
phys . lett .
* 391 * ( 2004 ) 248 .
h. okumura and y. okamoto : j. comput .
* 27 * ( 2006 ) 379 .
t. nishikawa , h. ohtsuka , y. sugita , m. mikami , and y. okamoto : prog .
* 138 * ( 2000 ) 270 .
t. okabe , m. kawata , y. okamoto , and m. mikami : chem .
* 335 * ( 2001 ) 435 .
y. sugita and y. okamoto : in _ lecture notes in computational science and engineering _ , eds .
t. schlick and h. h. gan ( springer - verlag , berlin , 2002 ) p. 303 ; arxiv : cond - mat/0102296 [ cond-mat.stat-mech ] .
d. paschek and a. e. garca : phys .
* 93 * ( 2004 ) 238105 .
n. metropolis , a. w. rosenbluth , m. n. rosenbluth , a. h. teller , and e. teller : j. chem . phys . * 21 * ( 1953 ) 1087 .
i. r. mcdonald : mol . phys . * 23 * ( 1972 ) 41 .
g. j. martyna , d. j. tobias , and m. l. klein : j. chem . phys .
* 101 * ( 1994 ) 4177 .
s. nos : mol . phys .
* 52 * ( 1984 ) 255 .
s. nos : j. chem .
* 81 * ( 1984 ) 511 .
w. g. hoover : phys .
a * 31 * ( 1985 ) 1695 .
h. c. andersen : j. chem .
* 72 * ( 1980 ) 2384 .
y. mori and y. okamoto : submitted .
y. sugita and y. okamoto : chem
. phys . lett . *
329 * ( 2000 ) 261 .
a. mitsutake and y. okamoto : chem .
* 332 * ( 2000 ) 131 .
a. m. ferrenberg and r. h. swendsen : phys .
* 63 * ( 1989 ) 1195 .
s. kumar , d. bouzida , r. h. swendsen , p. a. kollman , and j. m. rosenberg : j. comput .
* 13 * ( 1992 ) 1011 .
j. w. ponder and f. m. richards : j. comput .
* 8 * ( 1987 ) 1016 .
j. wang , p. cieplak , and p. a. kollman : j. comput .
* 21 * ( 2000 ) 1049 .
w. l. jorgensen , j. chandrasekhar , j. d. madura , r. w. impey , and m. l. klein : j. chem .
* 79 * ( 1983 ) 926 .
t. darden , d. york , and l. pedersen : j. chem .
* 98 * ( 1993 ) 10089 .
u. essmann , l. perera , m. l. berkowitz , t. darden , h. lee , and l. g. pedersen : j. chem . phys .
* 103 * ( 1995 ) 8577 .
g. j. martyna , m. e. tuckerman , d. j. tobias , and m. l. klein : mol . phys .
* 87 * ( 1996 ) 1117 .
ryckaert , g. ciccotti , and h. j. c. berendsen : j. comput . phys .
* 23 * ( 1977 ) 327 .
h. c. andersen : j. comput .
* 52 * ( 1983 ) 24 .
a. m. ferrenberg and r. h. swendsen : phys .
* 61 * ( 1988 ) 2635 [ errata ; * 63 * ( 1989 ) 1658 ] .
h. okumura and y. okamoto : bull .
* 80 * ( 2007 ) 1114 .
h. okumura and y. okamoto : j. phys .
b * 112 * ( 2008 ) 12038 . | we present generalized - ensemble algorithms for isobaric - isothermal molecular simulations .
in addition to the multibaric - multithermal algorithm and replica - exchange method for the isobaric - isothermal ensemble , which have already been proposed , we propose a simulated tempering method for this ensemble .
we performed molecular dynamics simulations with these algorithms for an alanine dipeptide system in explicit water molecules to test the effectiveness of the algorithms .
we found that these generalized - ensemble algorithms are all useful for conformational sampling of biomolecular systems in the isobaric - isothermal ensemble .
monte carlo ( mc ) and molecular dynamics ( md ) simulations with generalized - ensemble algorithms have been widely used for studies of proteins and peptides in order to have efficient conformational sampling ( for a review , see , e.g. , ref . 1 ) .
three well - known generalized - ensemble algorithms are the multicanonical algorithm ( muca ) @xcite ( the md version was developed in refs . 4 and 5 ) ,
replica - exchange method ( rem ) @xcite ( the method is also referred to as parallel tempering @xcite and the md version , which is referred to as remd , was developed in ref .
8) , and simulated tempering ( st ) @xcite .
these methods can be used for obtaining accurate physical quantities in the canonical ensemble . among them , rem ( particularly remd ) is often used because the weight factor is _ a priori _ known ( i.e. , the boltzmann factor ) , while those for muca and st have to be determined before the simulations .
multidimensional ( or multivariable ) extensions of the original generalized - ensemble algorithms have been developed in many ways ( see the references in ref . 1 ) and recently the general formulations have been given in refs .
11 - 13 . in this letter , we present a multidimensional simulated tempering algorithm in the isobaric - isothermal ( @xmath0 ) ensemble . several generalized - ensemble algorithms for the @xmath0 ensemble have been proposed ( for example , the multibaric - multithermal ( mubath ) algorithm @xcite and rem in temperature and pressure @xcite ) and here , we present an st algorithm for the @xmath0 ensemble to complete generalized - ensemble algorithms for this ensemble .
we first briefly review the generalized - ensemble algorithms for isobaric - isothermal molecular simulations .
let us consider a physical system that consists of @xmath1 atoms and that is in a box of a finite volume @xmath2 .
the states of the system are specified by coordinates @xmath3 and momenta @xmath4 of the atoms and volume @xmath2 of the box .
the potential energy @xmath5 for the system is a function of @xmath6 and @xmath2 .
we first describe mc simulation algorithms for muca , rem , and st in the @xmath0 ensemble . in these cases ,
momenta of atoms do not have to be considered . to make a system an equilibrium state ,
the detailed balance condition is imposed and a transition probability @xmath7 from an old state @xmath8 to a new state @xmath9 can be given by the metropolis criterion .
@xcite in mubath simulations , we introduce a function @xmath10 and use a weight factor @xmath11 $ ] so that the distribution function @xmath12 of @xmath13 and @xmath2 may be uniform : @xmath14 where @xmath15 is an arbitrary inverse reference temperature defined as @xmath16 ( @xmath17 is the boltzmann constant ) and @xmath18 is the density of states . to perform mubath mc simulations ,
the trial moves are generated in the same way as in the usual constant @xmath0 mc simulations @xcite and the transition probability from @xmath19 to @xmath20 is given by @xcite @xmath21,\ ] ] where @xmath22 - \mathcal{h}\left [ e(s , v),v\right ] - nk_{\text{b}}t_0 \ln ( v'/v ) \right\ } , \ ] ] and @xmath23 is the scaled coordinates defined by @xmath24 .
here , we are assuming the box is a cube of side @xmath25 . in rem simulations , we prepare a system that consists of @xmath26 non - interacting replicas of the original system , where @xmath27 and @xmath28 are the number of temperature and pressure values used in the simulation , respectively .
the replicas are specified by labels @xmath29 @xmath30 , temperature by @xmath31 @xmath32 , and pressure by @xmath33 @xmath34 . to perform rem mc ( remc ) simulations ,
we carry out the following two steps alternately : ( 1 ) perform a usual constant @xmath0 simulation in each replica at assigned temperature and pressure and ( 2 ) try to exchange the replicas . if the temperature ( specified by @xmath31 and @xmath35 ) and pressure ( specified by @xmath33 and @xmath36 ) between the replicas are exchanged , the transition probability from @xmath37},v^{[i]};t_{m_t},p_{m_p}),\cdots,(s^{[j]},v^{[j]};t_{n_t},p_{n_p } ) , \cdots\}$ ] to @xmath38},v^{[i]};t_{n_t},p_{n_p}),\cdots,(s^{[j]},v^{[j]};t_{m_t},p_{m_p } ) , \cdots\}$ ] at the trial is given by @xcite @xmath39 , \label{rem_transition1}\ ] ] where @xmath40},v^{[j ] } ) - e(s^{[i]},v^{[i ] } ) \right ] + ( \beta_{m_t } p_{m_p}- \beta_{n_t } p_{n_p})\left ( v^{[j ] } - v^{[i ] } \right ) .
\label{rem_transition2 } \ ] ] in st simulations , we introduce a function @xmath41 and use a weight factor @xmath42 $ ] so that the distribution function @xmath43 of @xmath44 and @xmath45 may be uniform : @xmath46 = \text{constant}. \label{eq : dis : st}\ ] ] from eq .
( [ eq : dis : st ] ) , it is found that @xmath41 is formally given by @xmath47 \right\},\ ] ] and the function is the dimensionless gibbs free energy except for a constant . to perform st mc simulations ,
we carry out the following two steps alternately : ( 1 ) perform a usual constant @xmath0 simulation and ( 2 ) try to update the temperature and pressure .
the transition probability from @xmath48 to @xmath49 for this trial is given by @xmath50 , \label{st_transition1}\ ] ] where @xmath51 .
\label{st_transition2}\ ] ] for md simulations with muca , rem , and st in the @xmath0 ensemble , the actual formulations depend on constant temperature and pressure algorithms . here
, we employ the md methods with the martyna - tobias - klein ( mtk ) algorithm , @xcite whose equations of motion follow nos @xcite and hoover @xcite for the thermostat and andersen @xcite for the barostat . to perform mubath md simulations , we solve the usual equations of motion for the mtk algorithm except that it is necessary to modify the equations for @xmath52 and @xmath53 as follows : @xmath54 where @xmath55,\ ] ] @xmath53 is the momentum associated with the logarithm of @xmath2
, @xmath56 is the momentum of the thermostat , and @xmath57 and @xmath58 are the masses of barostat and thermostat , respectively .
when we perform md simulations with rem and st , the momenta should be rescaled if the replicas are exchanged for the temperature in rem @xcite and the temperature is updated in st @xcite
. in rem md ( remd ) simulations with the mtk algorithm , let an old state be @xmath59},v^{[i]},p^{[i]},p_{\varepsilon}^{[i]},p_{\xi}^{[i]};t_{m_t},p_{m_p}),\cdots,(r^{[j]},v^{[j]},p^{[j]},p_{\varepsilon}^{[j]},p_{\xi}^{[j]};t_{n_t},p_{n_p } ) , \cdots\}$ ] and a new state @xmath59},v^{[i]},p^{[i]\prime},p_{\varepsilon}^{[i]\prime},p_{\xi}^{[i]\prime};t_{n_t},p_{n_p}),\cdots,(r^{[j]},v^{[j]},p^{[j]\prime},p_{\varepsilon}^{[j]\prime},p_{\xi}^{[j]\prime};t_{m_t},p_{m_p } ) , \cdots\}$ ] .
when replica exchange is performed , the momenta should be rescaled as follows : @xcite @xmath60\prime } = \sqrt{\frac{t_{n_t}}{t_{m_t}}}p^{[i ] } , \quad p^{[j]\prime } = \sqrt{\frac{t_{m_t}}{t_{n_t}}}p^{[j ] } \\
p_{\varepsilon}^{[i]\prime } = \sqrt{\frac{t_{n_t}w_{n}}{t_{m_t}w_{m}}}p_{\varepsilon}^{[i ] } , \quad p_{\varepsilon}^{[j]\prime } = \sqrt{\frac{t_{m_t}w_{m}}{t_{n_t}w_{n } } } p_{\varepsilon}^{[j ] } \\
p_{\xi}^{[i]\prime } = \sqrt{\frac{t_{n_t}q_{n}}{t_{m_t}q_{m}}}p_{\xi}^{[i ] } , \quad p_{\xi}^{[j]\prime } = \sqrt{\frac{t_{m_t}q_{m}}{t_{n_t}q_{n } } } p_{\xi}^{[j ] } , \end{aligned}\ ] ] where @xmath61 and @xmath62 are the masses in the simulation at @xmath63 and @xmath64 , and @xmath65 and @xmath66 are the ones at @xmath67 and @xmath68
. then the transition probability ( [ rem_transition1 ] ) and ( [ rem_transition2 ] ) can also be used in the remd simulations
. in st md simulations with the mtk algorithm , let an old state be @xmath69 and a new state @xmath70 .
if @xmath71 and @xmath72 are written as @xmath73 where @xmath74 and @xmath75 are the masses in the simulation at @xmath76 and @xmath77 , then it can be shown that the transition probability is given by eqs .
( [ st_transition1 ] ) and ( [ st_transition2 ] ) .
while the weight factors in rem simulations are _ a priori _ known , those in mubath and st simulations have to be determined before the simulations .
we propose two - dimensional extensions of the replica - exchange multicanonical algorithm ( remuca ) @xcite and the replica - exchange simulated tempering ( rest ) @xcite for the isobaric - isothermal ensemble to overcome the difficulty of determining the weight factors .
first , we perform a rem simulation for the isobaric - isothermal ensemble and then calculate the density of states and the dimensionless gibbs free energy by the multiple - histogram reweighting techniques , @xcite or the weighted histogram analysis method ( wham ) , @xcite using the results of the rem simulation .
the wham equations for isobaric - isothermal simulations are given by @xmath78 } , \\
\label{wham2 } \exp \left ( -g_{ij}\right ) & = \sum_e \sum_v n(e , v ) \exp \left[-\beta_i ( e+p_jv ) \right ] , \end{aligned}\ ] ] where @xmath79 , @xmath80 is the histogram of @xmath13 and @xmath2 , @xmath81 is the number of the data , and @xmath82 is the integrated autocorrelation time in the simulation at @xmath83 and @xmath84 , respectively .
these equations need to be solved self - consistently .
we remark that for biomolecular systems , it was found that @xmath82 may be set to a constant value for all @xmath29 and @xmath85 in eq .
( [ wham1 ] ) .
@xcite in the isobaric - isothermal version of rest , the weight factor can also be determined using the dimensionless gibbs free energy obtained from eqs .
( [ wham1 ] ) and ( [ wham2 ] ) . in order to verify that the generalized - ensemble algorithms discussed above can be effective for conformational sampling and give the same results , we performed md simulations with the three generalized - ensemble algorithms .
we used a system of an alanine dipeptide in 73 surrounding water molecules and the system was placed in a cubic cell with periodic boundary conditions . both of the backbone
dihedral angles @xmath86 and @xmath87 of the peptide were initially set to 180@xmath88 .
the simulations were performed with the tinker program package .
@xcite several of the programs were modified and a few programs were added so that mubath , rem , and st simulations with the mtk algorithm can be performed .
we used the amber parm99 force field @xcite for the peptide and the tip3p water model .
@xcite the electrostatic interactions were calculated by the particle mesh ewald method .
@xcite in the van der waals interaction calculations , we used the spherical cutoff method and the cutoff distance was set to 12.0 .
the integration of the equations of motion was employed by the method proposed by martyna _
et al . _
@xcite and the shake / rattle / roll constraint method @xcite was used so that the water molecules are rigid body molecules .
the unit time step was set to 0.5 fs .
the mass parameters @xmath57 and @xmath58 were determined as in ref .
39 .
first , we performed the two - dimensional remd simulation .
the simulation time was set to 2.0 ns .
we used the following 6 temperature ( @xmath89 ) and 4 pressure ( @xmath90 ) values : 280 , 305 , 332 , 362 , 395 , and 430 k for temperature and 0.1 , 65 , 150 , and 250 mpa for pressure . at the replica - exchange trial , either exchanging temperature ( @xmath44-exchange ) or exchanging pressure ( @xmath45-exchange ) was chosen randomly and then the pairs @xmath91 or @xmath92 for @xmath44-exchange and @xmath93 or @xmath94 for @xmath45-exchange were also chosen randomly . and
@xmath45 in one of the replicas and ( b ) the time series of the replica label at 280 k and 250 mpa.,width=453 ] figure [ result : remd](a ) shows the time series of @xmath44 and @xmath45 in one of the replicas and fig .
[ result : remd](b ) shows the time series of the label of the replicas in the simulation at 280 k and 250 mpa . from these figures
, it is found that random walks in @xmath44-@xmath45 space were realized in each of the replicas .
the acceptance ratios in the rem simulation were in the range from 0.152 to 0.205 for @xmath44-exchange and from 0.296 to 0.446 for @xmath45-exchange , respectively , which also implies that sufficient number of replica exchange was realized .
we then performed 24 mubath simulations of 2.0 ns , where the total simulation time was 48 ns so that it is equal to that in the remd simulation . in each of the simulations ,
different initial velocities were given .
the data were sampled every 100 fs . the reference temperature was set to 430 k. @xmath95 and @xmath96 were interpolated by the third - order polynomial , following ref .
17 . of @xmath13 and @xmath2 ( a ) in the mubath simulations and ( b ) for the @xmath0 ensemble at 280 k and 250 mpa ( left ) and at 430 k and 0.1 mpa ( right ) .
the distributions corresponding to the @xmath0 ensemble were calculated by the single - histogram reweighting techniques @xcite using the results of the mubath simulations.,width=453 ] figure [ result : mubath ] shows the probability distributions of @xmath13 and @xmath2 from the mubath simulations . from fig . [
result : mubath](a ) , it is found that the mubath simulations gave a uniform distribution in the range where the density of states was obtained accurately in the remd simulation and that the distribution of the mubath simulations was much wider than the ones for the @xmath0 ensemble ( see fig .
[ result : mubath](b ) ) .
twenty - four st simulations of 2.0 ns were carried out , which gave us the same number of sampled data as in the remd and mubath simulations . in each simulation
, different initial velocities were given .
we used the same temperature and pressure values as in the remd simulation .
we tried to update the parameter ( @xmath44 or @xmath45 ) every 100 fs and the data were sampled just before the trials .
the choice between updating @xmath44 ( @xmath44-update ) and updating @xmath45 ( @xmath45-update ) was made randomly and then either @xmath97 or @xmath98 for @xmath44-update and @xmath99 or @xmath100 for @xmath45-update was also chosen randomly , where @xmath31 and @xmath33 stand for the labels corresponding to the temperature and the pressure before the trial , respectively . and
@xmath45 for 2.0 ns and ( b ) the logarithm of the probability distribution @xmath101 of @xmath44 and @xmath45 . ,
width=453 ] figure [ result : st](a ) shows the time series of @xmath44 and @xmath45 and fig .
[ result : st](b ) shows the probability distribution of @xmath44 and @xmath45 .
these figures indicate that random walks in @xmath102 space were successfully realized .
the acceptance ratios in the st simulations were in the range from 0.260 to 0.430 for @xmath44-update and from 0.426 to 0.642 for @xmath45-update , respectively . and
@xmath87 in the simulations with ( a ) remd , ( b ) mubath , and ( c ) st and ( d ) in the conventional isobaric - isothermal simulations . in these figures ,
the probability distributions at 298 k and 0.1 mpa are plotted in logarithmic scale and were calculated by the single - histogram reweighting techniques for mubath and by wham for remd and st.,width=302 ] figure [ fig : dihedral ] shows the probability distributions of the backbone dihedral angles at 298 k and 0.1 mpa in all the simulations .
compared with the simulations with remd , mubath , and st , we also performed 24 conventional isobaric - isothermal simulations of 2.0 ns at 298 k and 0.1 mpa with different initial velocities .
the dihedral angle distributions in the simulations with the generalized - ensemble algorithms had a small peak in @xmath103 and @xmath104 , although there was no peak in the range in the distribution of the conventional simulations .
all the simulations with the three generalized - ensemble algorithms were able to provide the same results and reproduce the distribution obtained previously in refs .
43 and 44 . in this letter
, we presented the extensions of muca , rem , and st for isobaric - isothermal molecular simulations .
these algorithms can be effective methods for conformational sampling and give accurate physical quantities in the isobaric - isothermal ensemble .
therefore , one can use the generalized - ensemble algorithms to study temperature and pressure effects on complex systems such as biomolecular systems .
we expect the simulated tempering to be a more useful method for simulations in large systems .
many replicas are needed in rem simulations of large systems , while only one replica is used in st simulations .
in addition , st can be easily implemented into existing program packages compared to muca ( or mubath ) , in which a force calculation part of programs has to be modified for md simulations .
therefore , molecular simulations with st will be more easily applicable to larger systems than with the other two generalized - ensemble algorithms . |
let @xmath1 be the projective plane over the galois field @xmath27 .
an @xmath15-arc is a set of @xmath15 points no three of which are collinear .
an @xmath15-arc is called complete if it is not contained in an @xmath28-arc of @xmath1 . for an introduction in projective geometries over finite fields ,
see @xcite . in @xcite the close relationship between the theory of @xmath15-arcs , coding theory and mathematical statistics
is presented . in particular , a complete arc in a plane @xmath29 points of which are treated as 3-dimensional @xmath30-ary columns , defines a parity check matrix of a @xmath30-ary linear code with codimension 3 , hamming distance 4 , and covering radius 2 .
arcs can be interpreted as linear maximum distance separable ( mds ) codes @xcite and they are related to optimal coverings arrays @xcite and to superregular matrices @xcite .
one of the main problems in the study of projective planes , which is also of interest in coding theory , is finding of the spectrum of possible sizes of complete arcs . a great part of this work is devoted to upper bounds on @xmath0 , the smallest size of a complete arc in @xmath1 .
also we propose new constructions of complete arcs in @xmath1 and consider the spectrum of their possible sizes .
surveys of results on the sizes of plane complete arcs , methods of their construction and comprehension of the relating properties can be found in @xcite . in particular , as it is noted in @xcite , the following idea of segre @xcite and lombardo - radice @xcite is fruitful : the points of the arc are chosen , with some exceptions , among the points of a conic or a cubic curve .
we use this idea for constructions of complete arcs and for finding the spectrum of arc sizes , see sections [ sec2_computsearch ] , [ sec5_constr ] , and [ sec6_spectrum ] . the maximum size @xmath31 of a complete arc in @xmath1 is well known .
it holds that @xmath32 on the other hand , finding an estimation of the minimum size @xmath0 is a hard open problem .
problems connected with small complete plane arcs are considered in @xcite , see also the references therein .
we denote the aggregates of @xmath30 values : @xmath33 the exact values of @xmath0 are known only for @xmath34 see @xcite and recent work @xcite where the equalities @xmath35 are proven . also , there are the following lower bounds ( see @xcite ) : @xmath36 let @xmath37 be the smallest _ known _ size of a complete arc in @xmath1 . for @xmath38 the values of @xmath37 ( up to june 2009 )
are collected in ( * ? ? ?
* tab.1 ) whence it follows that @xmath39 @xmath40 @xmath38 . in @xcite , see also @xcite , complete @xmath41 -arcs are obtained for @xmath42 odd , @xmath43 or @xmath44 . in @xcite
complete @xmath41-arcs are obtained for even @xmath45 by the results above - cited it holds that @xmath46 for even @xmath47 , @xmath48 , the smallest known sizes of complete @xmath49 -arcs in @xmath1 are obtained in @xcite , see also @xcite .
they are as follows : @xmath50 also , @xmath51-arcs in @xmath29 @xmath52 are constructed in @xcite ; for @xmath53 it is proved that they are complete .
it gives a complete 3066-arc in @xmath54 let @xmath55 be the size of the smallest complete arc in any ( not necessarily galois ) projective plane @xmath56 of order @xmath30 . in @xcite , for _ sufficiently large _
@xmath30 , the following result is proved ( we give it in the form of ( * ? ? ? * tab.2.6 ) ) : @xmath57 where @xmath58 and @xmath59 are constants independent of @xmath30 ( i.e. universal constants ) .
the logarithm basis is not noted as the estimate is asymptotic . in this work , by computer search using randomized greedy algorithms ( see section [ sec2_computsearch ] ) , new small complete arcs in @xmath1 are obtained for all @xmath60 and for @xmath61 for @xmath62 the new complete arcs arose from a theoretical study on orbits of subgroups , helped by computer @xcite . from the sizes of the new arcs , with the use of ( [ eq1_<4sqroot(q ) ] ) and , @xcite , the following theorems result ( see also theorems [ th3_4.5sqroot(q ) ] , [ th3_5sqroot(q ) ] , and [ th4_ln0.75 ] for more details ) .
[ th1_<4.5_4.62 ] in @xmath29 the following holds .
@xmath63 [ th1_ln0.75 ] in @xmath29 @xmath64 moreover , the study of the values of @xmath37 allows us to conjecture that the estimate ( [ eq1_ln0.75 ] ) holds for all @xmath65 and the last estimate of theorem [ th1_<4.5_4.62 ] is right for all @xmath66 .
[ conj1_ln0.75 ] in @xmath29 @xmath67 regarding the spectrum of complete arc sizes , we note ( going after @xcite ) that in literature , complete arcs have constructed with sizes approximately @xmath21 ( see @xcite ) , @xmath18 ( see @xcite ) , @xmath19 ( see @xcite ) , @xmath68 ( see @xcite where such the arcs are constructed for @xmath69 )
. in particular , for even @xmath70 there is a complete @xmath71-arc @xcite .
important results on the spectrum of complete arc sizes are collected in ( * ? ? ?
* th.2.6 ) where it is noted , for example , that in @xmath1 with @xmath30 square there exists a complete @xmath72-arc . in @xcite ,
_ large complete arcs _ in @xmath73 are defined and new infinite families of the such arcs are constructed .
much attention is given to @xmath74-arcs and @xmath75-arcs sharing @xmath76 points with a conic for @xmath30 odd @xcite , , @xcite .
it is proved that for all odd @xmath30 there is a complete @xmath74-arc @xcite , see also @xcite .
also , a complete @xmath75-arc exists at least for the following _ odd _
@xmath77 @xmath78{dfmp - jg2005},\cite [ sec.\thinspace 2 , tab.\thinspace 2]{dfmp - jg2009},\cite[tab.\thinspace 2.4 ] { fp},\cite[introduction]{korchson2010 } ; } \notag \\ q & \equiv & 2\text { } ( \bmod~3),\text { } q\leq 4523\text { \cite{dfmp - jg2009,dmp - acct2010 } ; } \notag \\ q & \equiv & 1\text { } ( \bmod~4),\text { } q\leq 337\text { \cite{giordano } ; } \label{eq1+(q+7)/2 } \\ q & = & 2bt-1,\text { } t\text { odd prime , } b=1,2\text { \cite[introduction ] { korchson2010 } ; } \notag \\ q & \equiv & 3\text { } ( \bmod~4)\text { and condition of \cite[cor.\thinspace 4.16(i)]{korchson2010 } holds , see \cite[introduction]{korchson2010 } ; } \notag \\
q^{2 } & \equiv & 1\text { } ( \bmod~16)\text { and condition of \cite [ cor.\thinspace 4.17]{korchson2010 } holds , see \cite[introduction ] { korchson2010}. } \notag\end{aligned}\ ] ] for @xmath79 the known sizes of complete arcs in @xmath1 are collected in ( * ? ? ?
* tab.2 ) , ( * ? ? ?
* tab.2 ) , ( * ? ? ?
* tab.2.4 ) . in this work new constructions a , b ,
c of complete arcs are proposed , see section [ sec5_constr ] .
these constructions form families of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in a region @xmath16 where @xmath17 is of order @xmath18 or @xmath19 while @xmath20 has order @xmath21 .
the completeness of the arcs obtained by the new constructions is proved for @xmath80 and @xmath23 . moreover , there is reason to suppose that the arcs are complete for all @xmath24 . from theorems [ th5.1_result ] , [ th5.2_result ] , [ th5.3_result ] the following theorem results : [ th1_construc ] constructions a , b , and c of section _ [ sec5_constr ] _ form families of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the following regions : @xmath81 for the given @xmath30 , in order to show that arcs obtained by a construction are complete
we should calculate by computer some special value , say @xmath82 ( see definitions [ def5.1_overlineh ] , [ def5.2_gq ] , [ def5.3_jq ] ) and check if @xmath83 where @xmath84 for constructions a , b and @xmath85 for construction c. the calculations are relatively simple .
moreover , for @xmath86 and @xmath23 it holds that @xmath87 and the difference @xmath88 has a tendency to increasing when @xmath30 grows , see theorems [ th5.1_bytable1](iii),(iv ) , [ th5.2_bytable6](iii),(iv ) , [ th5.3_bycomput](iii)(iv ) .
it allows us to conjecture the following , cf .
conjecture [ conj5_constr ] .
[ conj1_construc ] the assertions of theorem _ [ th1_construc ]
_ hold also for all @xmath24 .
a @xmath15-arc of constructions a and b contains @xmath89 points in common with a conic and two points lying on a tangent to the conic ( construction a ) or on a bisecant of the conic ( construction b ) .
a @xmath15-arc of construction c contains @xmath90 points in common with a conic , two points lying on a bisecant and one point on a tangent .
note that in @xcite , @xmath15-arcs containing @xmath89 points of a hyperoval ( the nucleus among them ) and two points on its bisecant are constructed .
also , in the space @xmath91 , @xmath15-caps with @xmath89 points in common with a quadric are considered in @xcite . in @xcite , the cap contains two points on a tangent to the quadric while in @xcite two points lie on an external line .
the complete arcs of constructions a , b , c can be used as starting objects for inductive constructions of @xcite and @xcite , see remark [ rem5_starting - objects ] . in that way , using results of @xcite together with constructions a , b , c , one can generate infinite sets of families of complete caps in projective spaces @xmath92 of growing dimensions @xmath93 . also , infinite families of large complete arcs in @xmath73 with growing @xmath94 can be obtained by constructions of @xcite using arcs of constructions a , b , c. in this work , using constructions a , b , c and randomized greedy algorithms , new complete arcs in @xmath1 are obtained for @xmath95 and @xmath26 in section [ sec2_computsearch ] we describe the greedy algorithms used for obtaining new arcs . in section [ sec3_tables ] we collect the known and new upper bounds on @xmath0 for @xmath96 and @xmath97 .
the bounds are represented by tables , where values of @xmath37 are written , and by the corresponding relations . in section [ sec4_observations ]
we give the upper bounds on @xmath98 in the form of ( [ eq1_ln0.75 ] ) and substantiate conjecture [ conj1_ln0.75 ] . in section [ sec5_constr ] new constructions a , b , c of complete arcs are described .
finally , in section [ sec6_spectrum ] we present new sizes of complete arcs in @xmath1 with @xmath99 and @xmath100 .
some of the results of this work were briefly presented without proofs in @xcite , see also @xcite .
in this paper for computer search we use the randomized greedy algorithms ( * ? ? ? * sec.2 ) , ( * ? ? ? *
sec.2 ) that are convenient for relatively large @xmath30 and for obtaining examples of different sizes of complete arcs . at every step an algorithm minimizes or maximizes an objective function @xmath101 but
some steps are executed in a random manner .
the number of these steps and their ordinal numbers have been taken intuitively . also , if the same extremum of @xmath101 can be obtained in distinct ways , one way is chosen randomly .
we begin to construct a complete arc by using a starting set of points @xmath102 . at the @xmath103-th step one point
is added to the set and we obtain a point set @xmath104 . as the value of the objective function @xmath101 we consider the number of points in @xmath1 that lie on bisecants of the set obtained . for small arcs we look for the maximum of the objective function @xmath105 for the spectrum of arc sizes we use both the maximum and the minimum of @xmath101 . on every of random steps
we take @xmath106 of randomly chosen uncovered points of @xmath1 and compute the objective function @xmath101 adding each of these @xmath106 points to @xmath104 . the point providing the extremum is included into @xmath107 the value of @xmath106
is given intuitively depending upon @xmath108 upon the number of chosen points ( i.e. @xmath109 ) , and upon the current task ( small arcs or the spectrum of arc sizes ) .
for example , one can put @xmath110 for finding of the spectrum and @xmath111 with @xmath112 for small arcs .
as @xmath113 we can use a subset of points of an arc obtained in previous stages of the search . also , for finding the spectrum of arc sizes it is fruitful to take as @xmath113 a part of points of a conic .
a generator of random numbers is used for a random choice . to get arcs with distinct sizes , starting conditions of the generator
are changed for the same set @xmath113 . in this way
the algorithm works in a convenient limited region of the search space to obtain examples improving the size of the arc from which the fixed points have been taken . in order to obtain arcs with new sizes
one should make sufficiently many attempts with the randomized greedy algorithms . for small arcs ,
the so called predicted sizes considered in section [ sec4_observations ] are useful for understanding if a good result have been obtained .
if the result is not close to the predicted size , the attempts should be continued .
note also that arcs with sizes close to @xmath37 usually are obtained as a byproduct when we execute the computer search for the smallest arcs using a few attempts .
throughout the paper , in all tables we denote @xmath116 where @xmath117 also , in all tables , @xmath118 is a superior approximation of @xmath119 for @xmath38 , the values of @xmath37 ( up to june 2009 ) are collected in ( * ? ? ?
* tab.1 ) . in this work we obtained small arcs with new sizes for @xmath120 the new arcs
are obtained by computer search , based on the randomized greedy algorithms .
complete 90-arcs for @xmath121 came from a theoretical study on orbits of subgroups , helped by computer , see @xcite .
these arcs are announced also in ( * ? ? ?
* tab.1 ) .
a complete 104-arc for @xmath122 is obtained by the greedy algorithm with the starting point set taking from @xcite .
the current values of @xmath37 for @xmath38 are given in table 1 .
the data for @xmath123 improving results of ( * ? ? ?
* tab.1 ) are written in table 1 in bold font .
the exact values @xmath124 are marked by the dot @xmath125 . in particular , due to the recent result @xcite we noted the values @xmath126 * table 1 * the smallest known sizes @xmath127 of complete arcs in planes @xmath128 @xmath129 @xmath130 , @xmath131 @xmath132 from table 1 and the results of @xcite , on complete @xmath41-arcs for @xmath133 ( see introduction ) we obtain theorem [ th3_4sqroot(q ) ] improving and extending the results of ( * ? ? ?
* th.1 ) .
[ th3_4sqroot(q ) ] in @xmath29 the following holds .
@xmath134 also , @xmath135
@xmath136 in table 2 , the current values of @xmath37 for @xmath137 are given .
the data for @xmath133 with @xmath138 @xcite are written in bold font .
* table 2 * the smallest known sizes @xmath139 of complete arcs in planes @xmath29 @xmath140 @xmath141 , @xmath142 @xmath143 * table 2 * ( continue ) the smallest known sizes @xmath139 of complete arcs in planes @xmath29 @xmath140 @xmath144 , @xmath142 @xmath145 from table 2 , we obtain theorem [ th3_4.5sqroot(q ) ] . [ th3_4.5sqroot(q ) ] in @xmath29 the following holds .
@xmath146 also , @xmath147 in table 3 , the current values of @xmath37 for @xmath148 are given .
the data with @xmath149 are written in bold font .
values of @xmath37 for relatively great @xmath150 are given in table 4 .
the notation @xmath151 is explained in the next section . in table 2 for @xmath152 we use the results of @xcite , see also @xcite . in table 4 , for @xmath153 we use the results of @xcite , see also @xcite and introduction . the rest of sizes @xmath15 for small complete @xmath15-arcs in tables 2 , 3 and 4
is obtained in this work by computer search with the help of the randomized greedy algorithms .
note that a complete @xmath154-arc in @xmath155 of table 2 , a complete @xmath156-arc in @xmath157 of table 3 , and a complete @xmath158-arc in @xmath159 of table 4 improve the results of @xcite for @xmath160 , see introduction . * table 3 * the smallest known sizes @xmath161 of complete arcs in planes @xmath29 @xmath148 , @xmath162 , @xmath142 @xmath163 * table 3 * ( continue ) the smallest known sizes @xmath161 of complete arcs in planes @xmath29
@xmath148 , @xmath162 , @xmath142 @xmath164 from tables 3 and 4 , we obtain theorem [ th3_5sqroot(q ) ] . [ th3_5sqroot(q ) ] in @xmath29 the following holds .
@xmath165 also , @xmath166 @xmath167 * table 4 * the smallest known sizes @xmath168 of complete arcs in planes @xmath1 with @xmath97
@xmath169 @xmath170 @xmath171
we look for upper estimates of the collection of @xmath37 values from tables 1 - 4 in the form ( [ eq1_kimvu_c=300 ] ) , see @xcite and ( * ? ? ?
* tab.2.6 ) . for definiteness
, we use the natural logarithms .
let @xmath58 be a constant independent of @xmath30 .
we introduce @xmath172 and @xmath173 as follows : @xmath174 let @xmath175 be the average value of @xmath176 calculated in the region @xmath177 and @xmath115 under condition @xmath178 from tables 1 - 4 , we obtain observation 1 . * observation 1 .
* _ let _ @xmath179 or @xmath115 , _ under condition _
@xmath180__. then _ _ * ( i ) * _ when @xmath30 grows , _ @xmath181 _
_ has a tendency to decreasing . _
_ ( * ii ) * _ when @xmath30 grows , _ @xmath182 _
_ has a tendency to increasing . _
_ * ( iii ) * _ when @xmath30 grows _ , _ the values of _ @xmath183 _
_ oscillate about the average value _
_ + @xmath184 ( see fig . 1 ) .
_ also , _ @xmath185 data for relatively big @xmath30 , collected in table 4 , in large confirm observation 1 .
by observation 1 it seems that the values of @xmath186 and @xmath187 are sufficiently convenient for estimates of @xmath188 and @xmath189 from tables 1 - 4 , we obtain theorem [ th4_ln0.75 ] .
[ th4_ln0.75 ] in @xmath29 @xmath190 in theorem [ th1_ln0.75 ] we slightly rounded the estimate ( [ eq4_9987 ] ) .
the graphs of values of @xmath191 , @xmath192 , @xmath193 , and @xmath194 are shown on fig . 2 where @xmath191 is the top curve and @xmath194 is the bottom one .
one can see on fig .
2 that always @xmath195 and , moreover , when @xmath30 grows , the graphs @xmath196 and @xmath37 diverge so that positive difference @xmath197 increases .
we denote @xmath198 one can treat @xmath199 as a _ predicted _ value of @xmath188 .
then @xmath200 is the difference between the smallest known size @xmath37 of complete arcs and the predicted value . finally , @xmath201 is this difference in percentage terms of the smallest known size .
* observation 2 . *
@xmath179 _ _ or _ _ @xmath202 , @xmath203__. then _ _
@xmath204 @xmath205 by ( [ eq4_delta_q ] ) and ( [ eq4_percent ] ) , see also fig . 3 and 4 , the upper bounds of @xmath200 and @xmath201 are relatively small .
moreover , the upper bound of @xmath201 decreases when @xmath30 grows
. therefore the values of @xmath200 and @xmath206 are useful for computer search of small arcs . the relations ( [ eq4_dq74])([eq4_percent ] ) , theorems [ th1_ln0.75 ] and [ th4_ln0.75 ] , and figures 14 are the foundation for conjecture [ conj1_ln0.75 ] . [ rem4 ] by above
, @xmath196 seems a reasonable upper bound on the current collection of @xmath37 values .
it gives some reference points for computer search and foundations for conjecture [ conj1_ln0.75 ] on the upper bound for @xmath0 . in principle , the constant @xmath207 can be sightly reduced to move the curve @xmath208 near to the curve of @xmath37 , see fig . 2 .
for example , from tables 1 - 4 , theorem [ th4_ln0.741 ] holds .
[ th4_ln0.741 ] in @xmath29 @xmath209
in the homogenous coordinates of a point @xmath210 we put @xmath211 @xmath212 let @xmath213 let @xmath214 be a primitive element of @xmath215 remind that _ indexes of powers of
_ @xmath214 _
_ are calculated modulo _ _
@xmath216 throughout this section we use the conic @xmath217 of equation @xmath218 we denote points of @xmath217 as follows : @xmath219 through this subsection , @xmath220 is an _ odd _ _ prime_. let @xmath221 be an integer in the region@xmath222we denote by @xmath223 the following @xmath224-subset of the conic @xmath225 @xmath226we denote the points of @xmath227@xmath228let @xmath229 be the line of equation @xmath230 it is the _ tangent _ to @xmath217 at @xmath231 it holds that @xmath232,@xmath233 . * construction a. * let @xmath30 be an _ odd
_ _ prime_. let @xmath234 @xmath235 @xmath236 and @xmath237 be given by ( [ eq5.1_h])([eq5.1_pt_m ] ) .
we construct a point @xmath238-set @xmath239 in the plane @xmath1 as follows:@xmath240 the following lemma can be proved by elementary calculations .
[ lem5.1_01b_i+j]__(i ) _ _ * * * * let @xmath241 a point @xmath242 is collinear with points @xmath243 @xmath244 if and only if @xmath245 _ ( ii ) _ let @xmath246 @xmath247 @xmath248 then a point @xmath249 is collinear with @xmath243 @xmath244 if and only if @xmath250 _ ( iii ) _ let @xmath251 @xmath252 then a point @xmath253 is collinear with @xmath254 the @xmath238-set @xmath239 of construction a is an arc in @xmath1 . by ( [ eq5.1_h]),([eq5.1_vh ] ) ,
the sum @xmath255 in ( [ eq5.1_b = i+j ] ) is running on @xmath256 where @xmath257 if @xmath258 and @xmath259 if @xmath260 so , @xmath261 see ( [ eq5.1_pt_m ] ) . therefore @xmath236 and @xmath237 do not lie on bisecants of@xmath262 in other side , any point of @xmath223 does not lie on the line @xmath263 as @xmath263 is a tangent to @xmath217 in @xmath231 @xmath264 [ th5.1_l0||c - vm_onbiseckm]let @xmath221 be given by _ ( [ eq5.1_h])_. then all points of @xmath265 lie on bisecants of @xmath266 all points of @xmath229 are covered as two points @xmath236 and @xmath237 of this line belong to @xmath266 let @xmath267 and @xmath268 be sets of integers modulo @xmath108
i.e. @xmath269 . let @xmath270 by lemma [ lem5.1_01b_i+j](i ) , points @xmath271 are collinear .
therefore , a point @xmath244 of @xmath272 with @xmath273 lies on the bisecant of @xmath239 through @xmath236 and @xmath274 where @xmath275 @xmath276 let @xmath277 by lemma [ lem5.1_01b_i+j](i ) , a point @xmath278of @xmath272 with @xmath279 lies on the bisecant @xmath280 where @xmath281 @xmath282 let @xmath283 then @xmath284 also , by ( [ eq5.1_pt_m ] ) , @xmath285 whence @xmath286 where @xmath287 and @xmath288 .
hence , @xmath289 , @xmath290 let @xmath291 then @xmath292 also , by ( [ eq5.1_pt_m ] ) , @xmath293 whence @xmath294 where @xmath287 is as above .
therefore @xmath295 , @xmath296 , @xmath297 we proved that @xmath298 also we showed that the points @xmath278of @xmath272 with @xmath299 are covered by bisecants of @xmath239 through @xmath236 ( if @xmath273 ) or through @xmath237 ( if @xmath300 in the other side , @xmath301 where @xmath302 .
so , all points of @xmath303 are covered@xmath304 @xmath305
[ def5.1_overlineh]let @xmath30 be an _
_ prime_. let @xmath306 be an integer and let @xmath307 we call _ critical value _ _ of _ @xmath308 and denote by @xmath309 the _ smallest _ value of @xmath308 such that all points of the form @xmath310 @xmath247 @xmath311 lie on bisecants of @xmath312 let @xmath220 be an _
_ prime_. let @xmath313 and let @xmath314 then the arc @xmath239 of construction a is complete .
we use theorem [ th5.1_l0||c - vm_onbiseckm ] and definition [ def5.1_overlineh].@xmath264 in this subsection , we put @xmath220 as we checked by computer that @xmath315 if @xmath316 [ cor5.1_familyfq]let @xmath220 be an _ odd _ _ prime_. let @xmath317 then construction a forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath318 if @xmath319 then cardinality of this family is equal to @xmath320 and size of the smallest complete arc of the family is @xmath321 by computer search using lemma [ lem5.1_01b_i+j](ii),(iii ) we obtained the following theorem .
[ th5.1_bytable1]let @xmath220 be an _
_ prime_. let @xmath322 be given by definition _ [ def5.1_overlineh]_. we introduce @xmath323 and @xmath324 as follows : @xmath325 @xmath326 then the following holds .
@xmath327 for situations @xmath328 the values of @xmath322 are given in table 5 . * table 5 * the values @xmath329 for cases @xmath330
@xmath331 [ th5.1_result ] let @xmath30 be an _
_ prime _ with @xmath332 , @xmath333 , or @xmath334 then construction a forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath335 we use corollary [ cor5.1_familyfq ] and theorem [ th5.1_bytable1](ii).@xmath336 throughout this subsection , @xmath337 is a _
prime power_. let @xmath338 be an integer in the region @xmath339 we denote by @xmath340 the following @xmath341-subset of the conic @xmath342 @xmath343 clearly , @xmath344 let @xmath345 we denote the points of @xmath227 @xmath346 let @xmath347 be the line of equation @xmath348 it is the _ bisecant _ @xmath349 of @xmath350 we have @xmath351 . using ( [ eq5.2_g]),([eq5.2_gamma_-1_1 ] ) , by elementary calculations we obtained the following lemma .
[ lem5.2_(0,1,b)]_(i ) let _
@xmath352 a point @xmath242 is collinear with points @xmath353 @xmath354 if and only if @xmath355 _ ( ii ) _ a point @xmath242 is collinear with points @xmath356 and @xmath357 if and only if @xmath358 [ cor5.2_(0,1,0)]_(i ) _ for all @xmath108 the point @xmath359 does not lie on any bisecant of @xmath340 . _ ( ii ) _ let @xmath30 be even .
then the points @xmath360 are collinear if and only if @xmath361 also , the points @xmath362 are collinear if and only if @xmath363 . _ ( iii )
_ let @xmath364 @xmath365 then the points @xmath366 are collinear if and only if @xmath367 also , the points @xmath362 are collinear if and only if @xmath368 @xmath369 . _
( iv ) _ let @xmath370 @xmath365 then the points @xmath366 and @xmath362 are not collinear for any@xmath371 \(i ) in ( [ eq5.2_b = xi^i+xi^j])@xmath372 the case @xmath373 implies @xmath374 for even @xmath108 it is impossible@xmath304 for odd @xmath108 we obtain @xmath375 whence , by ( [ eq5.2_gamma_-1_1 ] ) , @xmath376 by ( [ eq5.2_g ] ) , it is impossible .
( ii)-(iv ) in ( [ eq5.2_b = xi^2i+u])@xmath372 the case @xmath373 implies @xmath377 whence @xmath378 if @xmath379 and @xmath380 if @xmath381 remind that @xmath382 and @xmath383 \(ii ) here @xmath30 is even but @xmath384 is odd . for @xmath385
we have @xmath386 whence @xmath361 for @xmath387 it holds that @xmath388 whence @xmath389 @xmath390 so , @xmath391 \(iii ) here both @xmath384 and @xmath392 are even . if @xmath393then @xmath394 whence @xmath395 @xmath396 it is possible if @xmath397 or @xmath398 . if @xmath399then , by ( [ eq5.2_gamma_-1_1 ] ) , @xmath400 whence @xmath401 @xmath390
so , @xmath402 or @xmath403 .
\(iv ) here @xmath384 is even whereas @xmath392 is odd . if @xmath379 then @xmath394 whence @xmath395 @xmath404 it is impossible .
for @xmath387 it holds that @xmath405 that is impossible .
@xmath406 * construction b. * let @xmath30 be a _ prime power .
_ assume that @xmath407 @xmath365 _ _ _ _ let @xmath408 @xmath409 @xmath410 and @xmath411 be given by ( [ eq5.2_g])([eq5.2_z_bg ] ) .
we construct a point @xmath412-set @xmath413 in @xmath1 as follows : @xmath414 from lemma [ lem5.1_01b_i+j ] it follows .
[ lem5.2_01b_i+j]let @xmath352 a point @xmath415 is collinear with @xmath356 , @xmath354 if and only if @xmath416 [ th5.2_is - arc]the @xmath412-set @xmath413 of construction b is an arc . by ( [ eq5.2_g]),([eq5.2_dg ] ) ,
the sum @xmath417 in ( [ eq5.2_beta = i+j ] ) is running on @xmath418 where @xmath419 so , @xmath420 see ( [ eq5.2_z_bg ] ) . therefore @xmath421 and @xmath411 do not lie on bisecants of@xmath422 in other side , any point of @xmath340 does not lie on the line @xmath423 as @xmath423 is the bisecant of @xmath217 through @xmath424 and @xmath425 where @xmath426 .
@xmath336 [ th5.2_l0||c - vm_onbiseckm]let @xmath30 be a _ prime power . _
assume that @xmath407 @xmath365 _ _ _ _ let @xmath338 be given by _
( [ eq5.2_g])_. then all points of @xmath427 lie on bisecants of the arc @xmath413 of construction@xmath428b . by ( [ eq5.2_g]),([eq5.2_dg ] ) , @xmath429 so , the point @xmath236 is covered by corollary [ cor5.2_(0,1,0)](ii),(iii ) .
all points of @xmath347 are covered as two points @xmath421 and @xmath411 of this line belong to @xmath430 throughout this proof , @xmath267 and @xmath268 are sets of integers modulo @xmath216 it can be said that @xmath267 and @xmath268 are sets of indexes of powers of @xmath431 let @xmath432 bylemma [ lem5.2_01b_i+j ] , points @xmath433 are collinear .
therefore , a point @xmath354 of @xmath434 with @xmath435 lies on the bisecant of @xmath413 through @xmath436 and @xmath421 where @xmath437 @xmath438 let @xmath439 by lemma [ lem5.2_01b_i+j ] , a point @xmath440of @xmath441 with @xmath442 lies on the bisecant @xmath443 where @xmath444 , @xmath445 as @xmath446 we have @xmath447 also , @xmath448 if @xmath449 then @xmath450 whence @xmath451 if @xmath452 then @xmath453 whence @xmath454 we proved that @xmath455 also we showed that the points @xmath440of @xmath456 with @xmath457 are covered by bisecants of @xmath413 either through @xmath421 ( if @xmath458 ) or through @xmath411 ( if @xmath459 ) . in the other side ,
@xmath460 where @xmath461 .
so , all points of @xmath462 are covered@xmath304 @xmath264 [ def5.2_gq]let @xmath30 be a _ prime power . _
let @xmath407 @xmath463 _ _ _ _ for integer @xmath464 let @xmath465 we call _ critical value _ _ of _ @xmath466 and denote by @xmath467 the _ smallest _ value of @xmath466 such that all points @xmath249 with @xmath468 and all points @xmath242 with @xmath469 lie on bisecants of @xmath470 let @xmath337 be a _ prime power .
_ let @xmath407 @xmath365 if @xmath471 and @xmath472 then the arc @xmath413 of construction b is complete .
we use theorem [ th5.2_l0||c - vm_onbiseckm ] and definition [ def5.2_gq ] .
@xmath264 in this subsection , we put @xmath337 as we checked by computer that @xmath473 if @xmath474 [ cor5.2_family]let @xmath407 @xmath475 be a _ prime power . _ if @xmath471 then construction b forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath476 if @xmath477 then size of the smallest complete arc of the family is @xmath478 by computer search using lemmas [ lem5.1_01b_i+j ] , [ lem5.2_(0,1,b ) ] , [ lem5.2_01b_i+j ] we obtained the following theorem .
[ th5.2_bytable6]let @xmath407 @xmath475 be a _ prime power .
_ let @xmath479 be given by definition _
[ def5.2_gq]_. _ _ _ _ we introduce @xmath106 and @xmath480 as follows : @xmath481 @xmath482 then the following holds .
@xmath483 for situations @xmath484 the values of @xmath485 are given in table 5 .
[ th5.2_result ] let @xmath407 @xmath475 be a _ prime power .
_ let @xmath486 , @xmath487 , or @xmath488 then construction b forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath489 we use corollary [ cor5.2_family ] and theorem [ th5.2_bytable6](ii).@xmath336 throughout this subsection , @xmath490 is a _ prime power _ and also @xmath491
@xmath365 let @xmath492 be an integer in the region @xmath493 notations @xmath494 @xmath495 and @xmath496 are taken from ( [ eq5.2_dg ] ) and ( [ eq5.2_z_bg ] ) with substitution @xmath338 by @xmath497 using ( [ eq5.3_j ] ) , it is easy to see that corollary [ cor5.2_(0,1,0)](i),(iv ) , theorem [ th5.2_is - arc ] and their proofs hold for @xmath494 @xmath498 and @xmath496 as well as for @xmath499 and @xmath500 * construction c. * let @xmath370 @xmath475 be a _ _ prime power.__@xmath304 _ _ _ _ let @xmath501 @xmath494 @xmath410 and @xmath502 be given by ( [ eq5.1_pt_m]),([eq5.3_j]),([eq5.2_dg ] ) , and ( [ eq5.2_z_bg ] ) .
we construct a point @xmath503-set @xmath504 in @xmath505 as follows : @xmath506 the @xmath503-set @xmath504 of construction c is an arc .
the set @xmath507 is an arc due to theorem [ th5.2_is - arc ] . by corollary [ cor5.2_(0,1,0)](i),(iv ) , the point @xmath236 does not lie on bisecants of @xmath508 and @xmath509 .
finally , @xmath510 are not collinear.@xmath305 [ th5.3_l0||c - dj_onbisecej]let @xmath370 @xmath475 be a _ prime power .
_ let @xmath492 be given by _
( [ eq5.3_j])_. then all points of @xmath511 lie on bisecants of the arc @xmath504 of construction c. all points of @xmath347 are covered as two points @xmath421 and @xmath502 of this line belong to @xmath512 throughout this proof , @xmath513 and @xmath514 are sets of integers modulo @xmath216 it can be said that @xmath513 and @xmath514 are sets of indexes of powers of @xmath431 we act similarly to the proof of theorem [ th5.2_l0||c - vm_onbiseckm ] .
let @xmath515 bylemma [ lem5.2_01b_i+j ] , a point @xmath354 of @xmath516 with @xmath435 lies on the bisecant of @xmath504 through @xmath436 and @xmath421 where @xmath517 @xmath518 let @xmath519 by lemma [ lem5.2_01b_i+j ] , a point @xmath440of @xmath520 with @xmath442 lies on the bisecant @xmath521 where @xmath522 , @xmath523 let @xmath524 by([eq5.2_gamma_-1_1]),([eq5.2_b = xi^i+xi^j ] ) , points @xmath525 and @xmath526 are collinear .
therefore , a point @xmath354 of @xmath527 with @xmath528 lies on the bisecant @xmath529 where @xmath530 @xmath531 as @xmath532 we have @xmath533 where @xmath534 see ( [ eq5.3_j ] ) .
also , by ( [ eq5.3_j ] ) , @xmath535 while @xmath536 we proved that @xmath537 also we showed that the points @xmath440of @xmath538 with @xmath539 are covered by bisecants of @xmath540 either through @xmath421 ( if @xmath458 ) or through @xmath502 ( if @xmath442 ) or , finally , through @xmath236 ( if @xmath528 ) . in the other side , @xmath541 where @xmath461 .
so , all points of @xmath542 are covered@xmath304 @xmath264 [ def5.3_jq]let @xmath370 @xmath475 be a _ prime power . _ for integer @xmath543 let @xmath544 we call _ critical value _ _ of _ @xmath545 and denote by @xmath546 the _ smallest _ value of @xmath545 such that all points @xmath249 with @xmath468 and all points @xmath242 with @xmath469 lie on bisecants of @xmath547 let @xmath490 be a _ prime power .
_ let @xmath370 @xmath365 if @xmath548 and @xmath549 then the arc @xmath504 of construction c is complete .
we use theorem [ th5.3_l0||c - dj_onbisecej ] and definition [ def5.3_jq ] .
@xmath264 in this subsection , we put @xmath490 as we checked by computer that @xmath550 if @xmath551 [ cor5.3_family]let @xmath370 @xmath475 be a _ prime power . _
if @xmath548 then construction c forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath552 if @xmath553 then cardinality of this family is equal to @xmath554 and size of the smallest complete arc of the family is @xmath555 by computer search using lemmas [ lem5.1_01b_i+j ] , [ lem5.2_(0,1,b ) ] , [ lem5.2_01b_i+j ] we obtained the following theorem .
[ th5.3_bycomput]let @xmath370 @xmath475 be a _ prime power .
_ let @xmath556 be given by definition _ [ def5.3_jq]_. _ _ _ _ we introduce @xmath557 and @xmath558 as follows : @xmath559 @xmath560 then it holds that @xmath561 for situations @xmath562 the values of @xmath563 are given in table 5 .
[ th5.3_result ] let @xmath370 @xmath475 be a _ prime power .
_ let @xmath564 or @xmath565 @xmath566 then construction c forms a family of complete @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in the region @xmath567 we use corollary [ cor5.3_family ] and theorem [ th5.3_bycomput](ii).@xmath336 basing on theorems [ th5.1_bytable1 ] , [ th5.2_bytable6 ] , [ th5.3_bycomput ] and taking into account that @xmath568 , @xmath569 , we conjecture the following , cf . conjecture [ conj1_construc ] . [ conj5_constr ] let @xmath570 @xmath485 , @xmath571 be given by definitions _
[ def5.1_overlineh ] _ , _ [ def5.2_gq ] _ , _ [ def5.3_jq]_. let for @xmath570 @xmath30 be prime while for @xmath485 and @xmath571 it holds that @xmath30 is a prime power . finally , let @xmath407 @xmath475 for @xmath485 and @xmath370 @xmath475 for @xmath571 .
then the following holds .
@xmath572 @xmath573 it is interesting to compare the relations ( [ eq5.1_ln^0.9]),([eq5.2_ln^0.9]),([eq5.3_098ln^095]),([eq5.3_conject ] ) with theorems [ th1_ln0.75 ] , [ th4_ln0.75 ] , [ th4_ln0.741 ] and conjecture [ conj1_ln0.75 ] and to compare also computer results providing theorems [ th5.1_bytable1 ] , [ th5.2_bytable6 ] , [ th5.3_bycomput ] with tables 1 and 2 .
one can see that the upper estimates of @xmath574 @xmath575 @xmath576 and @xmath571 have the same structure and the values of @xmath577 @xmath570 @xmath576 and @xmath546 have a close order .
this seems to be natural as _ almost all _ points of @xmath1 lie on bisecants of @xmath578 @xmath579 and @xmath580 see definitions [ def5.1_overlineh ] , [ def5.2_gq ] and [ def5.3_jq ] . [ rem5_starting - objects ] the complete arcs of constructions a , b , c can be used as starting objects in inductive constructions .
for example , for even @xmath30 , arcs of construction b can be used in constructions of ( * ? ? ?
* ths 1.1,3.14 - 3.17,4.6 - 4.8 ) . in that way
, one can generate infinite sets of families of complete caps in projective spaces @xmath92 of growing dimensions @xmath93 .
for every @xmath93 , constructions of @xcite can obtain a complete cap from every complete arc of construction b. also , it can be shown that in constructions a , b , c all points not on conic are _
external_. so , the arcs of constructions a and c for @xmath581 can be used as starting objects in constructions of @xcite , see ( * ? ? ? * th.23 ) .
thereby , infinite families of large complete arcs in @xmath73 with growing @xmath94 can be obtained .
the main known results on the spectrum of possible sizes of complete arcs in @xmath1 are given in introduction with the corresponding references . taking into account the results cited in introduction , we denote @xmath582
we suppose that the smallest known sizes @xmath37 are given in tables 1 - 4 of this paper .
for @xmath587 the assertion of the theorem follows from ( * ? ? ?
* tab.2 ) and ( * ? ? ?
* tab.2 ) . for @xmath588 and @xmath100
, we used constructions a , b , c of section [ sec5_constr ] .
the sizes not following from the constructions are obtained in this work by the randomized greedy algorithms . an experience obtained in computer search for the proof of theorem [ th5_spectrum ]
allows us to do conjecture [ conj5 ] .
here we took into account sizes that can be got by constructions a , b , c and the remark on sizes close to @xmath37 in the end of section [ sec2_computsearch ] . note also that the rest of sizes for all @xmath588 and @xmath100 was relatively easy obtained by the greedy algorithms with point subset of a conic taken as the starting set @xmath113 .
for this we used consequently subsets of cardinality approximately @xmath589 of the conic cardinality @xmath590 .
[ conj5 ] let @xmath591 , @xmath592 , be a prime power .
then in @xmath1 there are complete @xmath15-arcs of * all * the sizes in the region @xmath593 . moreover ,
complete @xmath15-arcs with @xmath594 can be obtained either by constructions a , b , c or by the randomized greedy algorithms .
davydov , g. faina , s. marcugini , f. pambianco , on the spectrum of sizes of complete caps in projective spaces @xmath599 of small dimension , in proc .
workshop on algebraic and combin .
coding theory , acct2008 , pamporovo , bulgaria , 2008 , 5762 .
math.bas.bg/acct2008/b10.pdf a.a .
davydov , g. faina , s. marcugini , f. pambianco , new sizes of complete arcs in @xmath1 , in proc .
xii int .
workshop on algebraic and combin .
coding theory , acct2010 , novosibirsk , russia , 2010 , 103108 .
davydov , s. marcugini , f. pambianco , complete @xmath602-caps in the spaces @xmath603 @xmath604 @xmath605 @xmath606 an odd prime , and a complete 20-cap in @xmath607 , des
. codes cryptogr .
50 ( 2009 ) 359372 .
davydov , s. marcugini , f. pambianco , a geometric construction of complete arcs sharing @xmath608 points with a conic , in proc .
xii int .
workshop on algebraic and combin .
coding theory , acct2010 , novosibirsk , russia , 2010 , 109115 .
hirschfeld , maximum sets in finite projective spaces . in : lloyd ,
( ed . ) : surveys in combinatorics , london math .
lecture note ser .
82 , cambridge university press , cambridge , 1983 , pp .
5576 .
hirschfeld , l. storme , the packing problem in statistics , coding theory and finite geometry : update 2001 , in : _ _
a. blokhuis , j.w.p .
hirschfeld , d. jungnickel , j.a .
thas ( eds . ) , finite geometries , developments of mathematics 3 ( proc . of the fourth isle of thorns conf . , chelwood gate , 2000 ) kluwer , 2001 , pp . 201246 . | new upper bounds on the smallest size @xmath0 of a complete arc in the projective plane @xmath1 are obtained for @xmath2 and @xmath3 where @xmath4 , @xmath5 . from these new bounds
it follows that for @xmath6 and @xmath7 , the relation @xmath8 holds .
also , for @xmath9 we have @xmath10 it is showed that for @xmath11 and @xmath12 , the inequality @xmath13 is true . moreover ,
the results obtained allow us to conjecture that this estimate holds for all @xmath14 the new upper bounds are obtained by finding new small complete arcs with the help of a computer search using randomized greedy algorithms .
also new constructions of complete arcs are proposed .
these constructions form families of @xmath15-arcs in @xmath1 containing arcs of all sizes @xmath15 in a region @xmath16 where @xmath17 is of order @xmath18 or @xmath19 while @xmath20 has order @xmath21 .
the completeness of the arcs obtained by the new constructions is proved for @xmath22 and @xmath23 .
there is reason to suppose that the arcs are complete for all @xmath24 .
new sizes of complete arcs in @xmath1 are presented for @xmath25 and @xmath26 projective plane , complete arcs , small complete arcs , spectrum of complete arc sizes |
in the last few years the fkm @xcite has been extensively studied in connection with the exciting idea of electronic ferroelectricity @xcite .
it is generally supposed that the ferroelectricity in mixed - valent compounds is of purely electronic origin , i.e. it results from an electronic phase transition , in contrast to the conventional displacive ferroelectricity due to a lattice distortion . since the fkm is probably the simplest model of electronic phase transitions in rare - earth and transition - metal compounds it was natural to test the idea of electronic ferroelectricity just on this model .
the fkm is based on the coexistence of two different types of electronic states in a given material : localized , highly correlated ionic - like states and extended , uncorrelated , bloch - like states .
it is accepted that insulator - metal transitions result from a change in the occupation numbers of these electronic states , which remain themselves basically unchanged in their character .
taking into account only the intra - atomic coulomb interaction between the two types of states , the hamiltonian of the spinless fkm with hybridization can be written as the sum of four terms : where @xmath3 , @xmath4 are the creation and annihilation operators for an electron in the localized state at lattice site @xmath5 with binding energy @xmath6 and @xmath7 , @xmath8 are the creation and annihilation operators of the itinerant spinless electrons in the @xmath9-band wannier state at site @xmath5 .
the first term of ( 1 ) is the kinetic energy corresponding to quantum - mechanical hopping of the itinerant @xmath9 electrons between sites @xmath5 and @xmath10 .
these intersite hopping transitions are described by the matrix elements @xmath11 , which are @xmath12 if @xmath5 and @xmath10 are the nearest neighbors and zero otherwise ( in the following all parameters are measured in units of @xmath13 ) .
the second term represents the on - site coulomb interaction between the @xmath9-band electrons with density @xmath14 and the localized @xmath15 electrons with density @xmath16 , where @xmath17 is the number of lattice sites .
the third term stands for the localized @xmath15 electrons whose sharp energy level is @xmath6 .
the last term represents the hybridization between the itinerant and localized states . in spite of fact that many of ground - state properties of the spinless fkm ( the nature of the ground state @xcite , the picture of valence and metal - insulator transitions @xcite , etc . )
are well understood at present , the problem of electronic ferroelectricity remains still an open question .
very recently portengen at al .
@xcite studied the fkm with a k - dependent hybridization in hartree - fock approximation and found , in particular , that a non - vanishing excitonic @xmath0-expectation value exists even in the limit of vanishing hybridization @xmath18 .
as an applied ( optical ) electrical field provides for excitations between d- and f - states and thus for a polarization expectation value @xmath19 , the finding of a spontaneous @xmath20 ( without hybridization or electric field ) has been interpreted as evidence for electronic ferroelectricity .
however , analytical calculations within well controlled approximation ( for @xmath21 small ) performed by czycholl @xcite in infinite dimensions do not confirm this conclusion .
in contrast to results obtained by portengen et al .
@xcite he found that the symmetric ( @xmath22 ) fkm does not allow for a ferroelectric ground state with a spontaneous polarization , i.e. there is no nonvanishing @xmath0-expectation value in the limit of vanishing hybridization . in order to shed some light on this controversy
we have decide to study the problem of electronic ferroelectricity in the fkm using the small - cluster exact - diagonalization method that was so successful in describing of valence and metal - insulator transitions in this model @xcite .
it should be noted that for given @xmath17 the full hilbert space of the spinless fkm consists of @xmath23 quantum states , thereby strongly limiting numerical computations .
although the number of states can be reduced considerably by the use of symmetries of @xmath24 , there is still a limit @xmath25 on the size of clusters that can be studied using small - cluster exact - diagonalization calculations .
however , we will show later that due to small sensitivity of the fkm on @xmath17 ( for a wide range of parameters ) , already such small clusters can describe satisfactory the behavior of the @xmath26-expectation value that lies in the center of our interest . to compare our numerical results with czycholls ones @xcite obtained for small coulomb interactions
we have started our investigation in the weak - coupling limit and half - filled band case ( @xmath27 ) .
the weak - coupling numerical results for @xmath20 obtained using the modified lanczos method @xcite are displayed in fig . 1 .
to reveal the finite - size effects on @xmath20 numerical calculations have been performed for three finite clusters of @xmath28 and 10 sites .
it is seen that there are nonzero finite - size effects on the @xmath26-expectation value in the weak - coupling limit , however they do not change qualitatively the behavior of @xmath20 in the limit @xmath29 that is crucial for the verification of spontaneous polarization . in all cases
the @xmath26-expectation value vanishes in the limit @xmath29 , so there is no spontaneous polarization in the spinless fkm .
thus , in accordance with czycholls weak - coupling results we can conclude that the spinless fkm does not allow for a ferroelectric ground state with a spontaneous polarization at least for @xmath21 small . unlike the method used by czycholl that is restricted to small interactions we can proceed in the numerical study of the fkm at arbitrary @xmath21 .
the strong - coupling numerical results for @xmath20 are displayed in fig . 1 for @xmath30 and @xmath31 .
obviously the one - dimensional fkm does not exhibit a ferroelectric ground state with a spontaneous polarization in the strong - coupling limit .
for both values of @xmath21 the @xmath26-expectation value vanishes for @xmath29 , and it is demonstrated that this result is independent of @xmath17 .
thus , the strong - coupling results can be satisfactory extended to large systems and they should be considered as definite . of course , the absence of ferroelectric ground state in the spinless fkm does not exclude that some other models could exhibit such a ground state .
the spinless fkm is not too realistic model of a rare - earth compound , because any real fermi system has at least a spin degeneracy .
therefore , it is natural to ask if the spin - one - half fkm would not allow for the ferroelectric ground state with a spontaneous polarization . numerically the problem can be easily solved since the numerical method used for the spinless fkm can be straightforwardly generalized also for the spin - one - half fkm .
unfortunately , including spins will result in further reduction of the size of clusters that can be analyzed using the exact - diagonalization method . in order to compensate partially for the small size of clusters
we next examine the model only for strong @xmath9-@xmath15 interactions that ( as was shown for the spinless fkm ) minimize considerably the finite - size effects .
the hamiltonian of the spin - one - half fkm can be obtained directly from the spinless model by including the spins for both @xmath9 and @xmath15 electrons and by adding the on - site coulomb interaction @xmath32 that acts between two @xmath15 electrons of opposite spins ( the last term ) : the ground - state properties of this model , for @xmath34 , have been investigated in our preceding paper @xcite .
we have found that numerical results depend strongly on @xmath15-@xmath15 interaction strength @xmath32 , but they are relatively insensitive to @xmath9-@xmath15 interaction strength @xmath21 ( for @xmath35 ) . therefore to represent the typical behavior of the model at nonzero @xmath1 , and to minimize finite - size effects we choose in the next study the value @xmath36 that is sufficiently large to stabilize the system .
another advantage of this selection is that the ground - state phase diagram of the spin - one - half fkm without hybridization is well understood @xcite for large values of @xmath21 . particularly , in the strong - interaction limit @xmath37 the ground state is insulating for @xmath38 and metallic for @xmath39 . at @xmath40
the model exhibits a discontinuous insulator - metal transition that is accompanied by a integer - valence transition from @xmath41 ( @xmath42 ) to @xmath43 ( @xmath44 ) .
the same behavior exhibits the model also in the opposite limit @xmath45 : the ground state is insulating for @xmath46 and metallic for @xmath47 .
however , a discontinuous insulator - metal transition that takes place at @xmath48 realizes now between an integer - valence state @xmath41 and an inhomogeneous intermediate - valence state @xmath49 .
these results show that there are only three physically different ground states in the spin - one - half fkm without hybridization , and namely , an insulating integer - valence ground state with @xmath41 , a metallic integer - valence ground state with @xmath43 , and a metallic intermediate - valence ground state with @xmath50 . here
we examine whether these ground states are stable against a small , finite hybridization or whether some new ground states are obtained if one starts from a finite hybridization and studies the @xmath29 limit of the model . again
the special attention is devoted to the question if the model can exhibit a ferroelectric ground state with a spontaneous , nonvanishing polarization @xmath51 .
the strong - coupling ( @xmath52 ) numerical results for @xmath20 obtained on finite clusters of 4 and 6 sites are displayed in fig .
it is seen that for both @xmath38 ( for @xmath34 an insulating integer - valence state ) and @xmath39 ( for @xmath34 a metallic integer - valence state ) the @xmath26-expectation value vanishes in the limit @xmath29 indicating that there is no spontaneous polarization in the spin - one - half fkm .
it should be noted that this result is expected since approximate solutions @xcite lead also to @xmath53 for the integer - valence states with @xmath41 and @xmath43 .
a less trivial situation is expected in the intermediate - valence state with @xmath50 . as was discussed above this state exists in the spin - one - half fkm without hybridization for @xmath45 and @xmath47 .
particularly , for @xmath54 @xmath36 and @xmath55 we have found that @xmath41 for @xmath56 , @xmath57 for @xmath58 , and @xmath43 for @xmath59 . to examine the stability of intermediate - valence state against a small , finite hybridization we chose the value @xmath60 for numerical calculations .
the results obtained for @xmath20 are shown in fig .
again the @xmath26-expectation value vanishes in the limit @xmath29 indicating the absence of spontaneous polarization .
we have verified this important result also for @xmath61 ( see inset in fig . 3 ) .
unfortunately , due to the memory limitations we were not able to continue with numerical computations on larger lattices ( the memory requirement of the lanczos method for clusters larger than @xmath61 is beyond the reach of present day computers ) to exclude definite the existence of spontaneous polarization in the spin - one - half fkm .
however , in accordance with results obtained for the spinless fkm we do not expect that the behavior of @xmath20 for @xmath29 could be qualitatively changed on larger lattices . in summary
, we have used the extrapolation of small - cluster exact - diagonalization calculations to study the possibility of electronic ferroelectricity in the one dimensional spinless and spin - one - half fkm .
it was found that neither spinless nor spin - one - half version of the fkm does not allow for a ferroelectric ground state with a spontaneous polarization , i.e. there is no nonvanishing @xmath0-expectation value for vanishing hybridization @xmath1 .
fig . 2 .
hybridization dependence of the @xmath9-@xmath15-polarization @xmath51 in the spin - one - half fkm calculated for @xmath64 ( for @xmath34 an insulating integer - valence state ) and @xmath63 ( for @xmath34 a metallic integer - valence state ) .
inset : hybridization dependence of @xmath20 in the limit of vanishing @xmath1 for @xmath65 and @xmath66 .
hybridization dependence of the @xmath9-@xmath15-polarization @xmath51 in the spin - one - half fkm calculated for @xmath60 ( for @xmath34 an intermediate - valence state ) .
inset : hybridization dependence of @xmath20 in the limit of vanishing @xmath1 for @xmath65 and @xmath61 . | the extrapolation of small - cluster exact - diagonalization calculations is used to examine the possibility of electronic ferroelectricity in the one dimensional spinless and spin - one - half falicov - kimball model ( fkm ) .
it is found that neither spinless nor spin - one - half version of the fkm does not allow for a ferroelectric ground state with a spontaneous polarization , i.e. there is no nonvanishing @xmath0-expectation value for vanishing hybridization @xmath1 .
= -1pt |
in this note , we would like to discuss in detail the choice of the radiation gauge as it is frequently applied in holographic approaches to qcd .
our specific interest arose from the soft wall model , but the arguments given in the next section do apply to any gauge field in an ads or ads black hole background . + in the soft wall model , one studies excitations of a gauge field in ads with action @xmath2,\ ] ] for left and right gauge fields @xmath3 and @xmath4 .
we will denote 5-dimensional indices as @xmath5 and 4-dimensional ( boundary ) indices with @xmath6 .
the corresponding equations of motion of either gauge field are given by : @xmath7 where the background geometry is either ads : @xmath8 or the ads black hole : @xmath9 with @xmath10 and @xmath11 the horizon location .
the background includes a dilaton field @xmath12 whose backreaction on the geometry is assumed to be minor .
this model gained popularity to holographically capture important qcd physics , e.g. because of its correct scaling behavior of the meson spectrum and the reader is referred elsewhere for more motivation and details of this model @xcite .
as a particular example where the soft model was used to study strongly coupled qcd physics , let us refer to @xcite where heavy quarkonia were studied .
the authors of @xcite suggested choosing a flavor - dependent soft - wall parameter @xmath13 , where the light quarks ( @xmath14 , @xmath15 , @xmath16 ) are combined into a @xmath17 soft wall model and the heavy quark of interest ( charm in our case ) is treated on its own in a @xmath18 abelian model : @xmath19 our goal in this work is to compute the momentum - dependence of the @xmath20 spectral function in this model and to demonstrate that one of the conclusions of @xcite , namely that isotropy ( rotational invariance ) is present in the spectral function , is actually a consequence of a forbidden choice of gauge .
the field @xmath21 has of course a large gauge redundancy and we will investigate here whether the radiation gauge can always be imposed to begin with .
the authors of @xcite and @xcite impose the gauge choice @xmath22 where @xmath23 , the transverse indices .
we first note that , for a diagonal metric that is only warped in the @xmath24-direction , one readily finds that @xmath25 and hence there is no difference between covariant derivatives and ordinary partial derivatives when computing the divergence of @xmath21 .
+ the main discussion we want to start here is : can this gauge be chosen in the first place ? + without loss of generality , we may consider a two - step process to get there .
first let @xmath26 .
we choose @xmath27 such that @xmath28 which eliminates @xmath29 .
in general this @xmath27 depends on all 5 coordinates .
next we perform a second gauge transformation @xmath30 .
since we want to maintain @xmath31 , we should have @xmath32 .
additionally , we want to impose the lorenz ( landau ) gauge in the transverse ( 4d ) dimensions . this can be done by solving the following wave equation for @xmath33 : @xmath34 in terms of ordinary coordinate partial derivatives we have @xmath35 the tricky part is that this metric still depends on @xmath24 .
but we just established that @xmath33 does not ! in general
this equation is hence unsolvable . +
before delving into a more detailed exposition , let us look at an analogous problem in ( 3 + 1)d classical electrodynamics : we are inclined to choose the temporal and coulomb gauge simultaneously ( thereby defining the radiation gauge ) : @xmath36 and @xmath37 .
the latter leads to the equation for @xmath33 : @xmath38 for time - independent @xmath33 .
this equation is nevertheless solvable : in spite of the time dependence of @xmath39 itself , it holds that @xmath40 by virtue of gauss law .
the above gauge is hence possible by imposing the ( sourceless ) maxwell equations .
though , in the presence of sources , the radiation gauge is inadmissible .
+ in our case , we are interested in the analogous problem of a sourcefree maxwell field , but this time in a curved background .
+ maxwell s equations take the form @xmath41 in the gauge @xmath0 , we consider the @xmath42 component of this equation . for a diagonal metric
whose components are independent of @xmath43 , we obtain @xmath44 taking then the @xmath24-derivative of ( [ tosolve ] ) , we obtain @xmath45 in the particular case of a metric for which all components have the same @xmath24-dependence , we can factor out all @xmath24-dependence , and rewrite the above as @xmath46 where @xmath47 is independent of @xmath24 . even though @xmath48 depends on @xmath24 ,
the constraint ( [ constraint ] ) implies the left hand side is @xmath24-independent , just as the right hand side .
no problems occur and a solution of the above equation in terms of @xmath33 is possible .
this is the case for ads spacetimes ( [ ads ] ) .
+ for the ads black hole ( [ adsbh ] ) , the metric has a different @xmath24-dependence for the @xmath49 component compared to the other components .
for such a space , we can write equation ( [ tosolve ] ) as @xmath50 dividing by @xmath51 and differentiating w.r.t .
@xmath24 we obtain @xmath52 the constraint ( [ constraint ] ) for this case reduces to @xmath53 which simplifies equation ( [ alm ] ) and yields @xmath54 which is impossible to satisfy since @xmath55 is @xmath24-dependent in general , unlike the right hand side .
we are thus forced to conclude that the radiation gauge choice is impossible to implement for the ads black hole . for a non - exhaustive list of instances where such has been done ,
we refer to @xcite .
the gauge choice issue that we highlighted in the previous section , is relevant for the momentum dependence of the spectral functions . from the latter quantity
, we can infer information on the melting behavior of the quarkonium in the plasma .
meson melting in a holographic context was also considered in e.g. @xcite , and holographic quarkonia in @xcite . in @xcite
it was argued that , even though a spatial momentum breaks isotropy , the spectral functions still are , quite miraculously , isotropic .
sym can not support jets . ]
we will demonstrate that this actually arises due to the faulty choice of gauge : if one does not make the additional gauge choice @xmath56 , then isotropy is broken as it would be expected when momentum is inserted . +
our main interest here is on the charmonium bound state where we focus on the vector modes ( @xmath57 ) .
to start , one makes the ansatz @xmath58 for the dependence on the boundary coordinates .
we will choose @xmath59 along the 3-axis .
+ if one includes spatial momentum into the spectral functions , the transverse polarization differential equation for @xmath60 becomes @xmath61 where the metric of the black hole ( [ adsbh ] ) is to be used .
this ode can be readily solved numerically using the same methods as in @xcite .
we will present the results further on .
+ the @xmath62 polarization on the other hand becomes much more intricate : it couples directly to @xmath63 and one hence needs to solve a coupled system of differential equations .
these are given by @xmath64 and @xmath65 one readily finds that the combined equation to be solved equals @xmath66 this is a second order differential equation for @xmath67 .
a frobenius analysis yields the following asymptotic behavior of @xmath68 . for @xmath69 ,
one finds for the two independent solutions @xmath70 and @xmath71 : @xmath72 for @xmath73 , one finds @xmath74 with these boundary values , one can solve the differential equation for @xmath68 from the boundary @xmath75 to the horizon @xmath11 .
+ following the real - time dictionary @xcite , see also @xcite , a linear combination of @xmath70 and @xmath71 has to be taken to satisfy ingoing boundary conditions at the horizon .
we parameterize this @xmath68 as @xmath76 in terms of two complex numbers @xmath77 and @xmath78 . with this solution for @xmath68 ,
the spectral function can be distilled as follows . evaluating ( [ ode1 ] ) near the boundary , one finds @xmath79 or with the above form of @xmath68 : @xmath80 yielding @xmath81 the coupled differential equation leads to three correlators : @xmath82 , @xmath83 and @xmath84 where the currents are the charm vector currents .
for instance , in evaluating the spectral function for @xmath63 , one needs to evaluate the combination @xcite @xmath85 to obtain the retarded green function , from which follows the spectral function as @xmath86 we therefore look at @xmath87 and functionally differentiate it w.r.t .
the boundary value of @xmath63 twice .
this equals @xmath88 the second term can be renormalized by introduction of a local counterterm ( cf .
holographic renormalization ) .
+ one can also look at the @xmath89 correlator , though we will have no interest in this mixed correlator in this work .
+ finally , we can also look at the @xmath90 correlator .
the correlator can in this case be distilled as ( using @xmath91 ) @xmath92 so for the spectral function , it turns out we should look at @xmath93 .
it is this correlator that we will be interested in in the remainder of this work .
below are the resulting figures of the spectral functions as @xmath94 is varied for the two different polarizations .
we denote everything in dimensionless quantities ( by rescaling @xmath24 ) where in particular @xmath95 for different values of @xmath96 at @xmath97 .
blue : @xmath98 , purple : @xmath99 , yellow : @xmath100 , green : @xmath101 . ] for different values of @xmath96 at @xmath102 .
blue : @xmath98 , purple : @xmath99 , yellow : @xmath100 , green : @xmath101 . ] for different values of @xmath96 at @xmath97 .
blue : @xmath98 , purple : @xmath99 , yellow : @xmath100 , green : @xmath101 . ]
figures [ momdep1 ] and [ momdep1 t ] show the melting of the spectral peaks as the temperature is increased .
this effect remains the same regardless of the value of @xmath94 .
comparing figures [ momdep1 ] and [ bv1parmom ] , it is clear that the spectral functions are not isotropic : if the polarization is tangential to the momentum , the peaks hardly decline ; whereas if the polarization is perpendicular to the momentum , the peaks decrease quickly as the momentum is increased .
note that for @xmath103 , the spectral functions indeed are the same : this is of course expected as in this case @xmath104 and hence full isotropy should be restored indeed .
+ figure [ momdep1 ] has the peculiar property that the spectral peaks do not damp as @xmath94 increases .
if one increases @xmath94 even further , one finds the result of figure [ momdepf ] . for different values of @xmath96 at @xmath97 .
blue : @xmath105 , purple : @xmath106 , yellow : @xmath107 , green : @xmath108 . ]
in this section , we will demonstrate how the spectral functions are modified if a background magnetic field @xmath109 is turned on . in order to analyze the influence of the magnetic field
, we will follow our earlier model @xcite where we proposed a dbi - modification of the soft - wall model to allow the magnetic field to couple to the charged constituents of the mesons . if one compares the differential equations to those described in the previous section , only a few things change . the background @xmath110-field , which we orient along the 3-axis , corresponds to @xmath111 since the charm quark charge is @xmath112 .
we obtain for @xmath113 : @xmath114,\end{aligned}\ ] ] with its determinant @xmath115 we will denote by @xmath116 only the symmetric part of the metric tensor @xmath117 : @xmath118,\end{aligned}\ ] ] where @xmath119 . + then substituting these values of @xmath120 and @xmath121 in the earlier differential equations ( [ v1par ] ) , ( [ ode1 ] ) and ( [ ode2 ] ) , one obtains the correct equations for the dbi - modification of the model . was fixed in @xcite in terms of the ads length @xmath122 .
the latter only figures as an overall prefactor of the action and we ignore it further on .
] moreover , the asymptotic behavior of the solutions both at the horizon and at the boundary is unaffected by including the magnetic field . + for the sake of brevity , we will only look at the special case where the spatial momentum @xmath59 is tangential to the magnetic field @xmath110 . the other case can be dealt with analogously but will not be discussed here .
the set - up of the different vector quantities is demonstrated in figure [ configur ] . turning on the magnetic field ,
one finds the results of figures [ momdep2 ] and [ bv1parmomb ] . with @xmath123 gev@xmath124 for different values of @xmath96 at @xmath97 .
blue : @xmath98 , purple : @xmath99 , yellow : @xmath100 , green : @xmath101 . ] with @xmath123 gev@xmath124 for different values of @xmath96 at @xmath97 .
blue : @xmath98 , purple : @xmath99 , yellow : @xmath100 , green : @xmath101 . ]
first of all , we note that the spectral functions are not the same , even for @xmath103 : this corresponds to the breaking of the isotropy already by the magnetic field alone . of course
, we have only discussed the case where @xmath59 and @xmath110 are aligned . the other case where for instance @xmath59
is directed along the 1-axis can also be studied analogously , though we will not pursue this here .
+ both the results of this and the previous section demonstrate that turning on momentum causes the peaks to shift to higher @xmath125 as expected , where the peak location is at fixed @xmath126 .
the reason for this location is that these peaks are identified with the delta - peaks in the thermal ads case ( where @xmath127 ) . in that case ,
the time direction and the spatial directions along the boundary are fully equivalent ( no warping in the @xmath24-direction ) and full lorentz invariance should be manifestly present . as a conclusion , all spectral functions widen as @xmath94 is increased ;
this implies the excitations melt at a lower temperature in correspondence with @xcite :
the meson melts under the hot wind of the quark - gluon plasma .
we do want to remark that the height of the spectral peak on the other hand does not systematically decrease , but this depends on the polarization and the magnitude of the magnetic field .
we have scrutinized the quite common application of the radiation gauge in holography and demonstrated that for the ads black hole , one can not impose this gauge .
we then demonstrated that this issue is of physical relevance .
this was achieved by showing that the previously observed emergent isotropy for the momentum - dependent ( quarkonium ) spectral function is actually the result of a faulty choice of gauge .
as the momentum @xmath94 increases , the spectral peaks widen and melt at a lower temperature than before .
we furthermore provided some results on the momentum dependence of the spectral functions when including a background magnetic field , following our earlier proposed model @xcite of a dbi - extension of the soft wall model .
we thank d. r. granado for collaboration at an early stage of this work .
t. mertens gratefully acknowledges financial support from the ugent special research fund , princeton university , the fulbright program and a fellowship of the belgian american educational foundation .
m. fujita , t. kikuchi , k. fukushima , t. misumi and m. murata , `` melting spectral functions of the scalar and vector mesons in a holographic qcd model , '' phys .
d * 81 * ( 2010 ) 065024 [ arxiv:0911.2298 [ hep - ph ] ] . | we point out a subtlety in choosing the radiation gauge ( @xmath0 combined with the lorenz gauge ) for gauge fields in ads / qcd for black hole backgrounds .
we then demonstrate the effect of this on the momentum - dependence of the spectral functions of the @xmath1 vector meson , showing a spreading with momentum and a breaking of isotropy , in contrast to previous results in the literature .
we also discuss the dependence on a background magnetic field , following our earlier proposed model . |
low - mass ( @xmath111 m@xmath7 ) , pre - main sequence stars are fully convective as they gravitationally contract and descend their vertical tracks in the color - magnitude diagram ( hayashi 1961 ) .
when core temperatures reach 2.5 to 3.0 mk , lithium depletion begins as its dominant isotope ( @xmath12li ) fuses with hydrogen nuclei to form @xmath13he ( bodenheimer 1965 ) .
convective processes circulate material through the stellar core , depleting all available lithium over timescales of tens of myr ( baraffe et al . 1998 ; song et al .
2002 , and references therein ) . in spite of its low abundance , @xmath12li
is readily observed in stellar atmospheres in an electronic transition that is a strong doublet at @xmath106707.78 , 6707.93 .
given the short timescale of lithium depletion , the presence of @xmath86708 absorption in the spectra of low - mass stars is recognized as an indicator of youth ( e.g. bonsack & greenstein 1960 ; herbig 1965 ) . stars presumably begin their lives with an amount of lithium consistent with interstellar abundance levels .
anders & grevesse ( 1989 ) derive an abundance of lithium extracted from carbonaceous chondrites of @xmath9(li)[email protected] dex , which is adopted here as the interstellar @xmath12li abundance level .
hbc 425 ( st 34 ) was first identified as an h@xmath14 emission star in the objective prism survey of stephenson ( 1986 ) .
the star was classified as an early to mid m - type star exhibiting balmer line and @xmath85876 emission by downes & keyes ( 1988 ) . the spectral type was further refined by white & hillenbrand ( 2005 ) to m3 , who determined that hbc 425 is a double - line spectroscopic binary with components of identical effective temperature and luminosity .
hbc 425 lies @xmath11.5@xmath15 from the molecular cloud core l1558 near the southern extremity of the taurus - auriga molecular cloud complex and was subsequently assumed to be a member of the star forming region given its strong t tauri - like attributes . surprisingly , white & hillenbrand ( 2005 ) discovered that neither component of the double - line spectroscopic binary exhibits @xmath86708 absorption .
the isochronal age of hbc 425 was determined to be 8@xmath63 myr , substantially younger than the predicted lithium depletion timescale of @xmath125 myr using the pre - main sequence models of baraffe et al .
white & hillenbrand ( 2005 ) suggested that a distance of @xmath190 pc would reconcile the discrepancy between isochronal and lithium depletion ages at @xmath121 - 25 myr .
hartmann et al .
( 2005 ) favor this interpretation making hbc 425 the oldest known accreting pre - main sequence star .
given such strong evidence for membership in taurus - auriga , however , white & hillenbrand ( 2005 ) conclude that hbc 425 was likely much younger than implied by lithium depletion .
hartmann et al .
( 2005 ) obtained a mid - infrared spectrum of hbc 425 using the infrared spectrograph ( irs ) on _ spitzer space telescope _
( werner et al . 2004 ) .
the spectrum revealed substantial excess emission at wavelengths @xmath165.4 @xmath2 m , but the lack of excess emission at shorter wavelengths ( i.e. 2.2 @xmath2 m ) implied the presence of an evacuated inner disk region .
the depleted inner disk would likely result from dynamical clearing by the two components of the spectroscopic binary , similar to the coku tau 4 system that was previously recognized as a transition disk source ( ireland & kraus 2008 ) .
andrews & williams ( 2005 ) place an upper limit on the mass of the circumbinary disk of @xmath175@xmath1810@xmath19 m@xmath7 .
the kinematic survey of taurus - auriga by bertout & genova ( 2006 ) found insignificant proper motions for hbc 425 as well as for 10 other assumed taurus - auriga members including coku tau 4 , dq tau , and v836 tau .
while purely circumstantial in nature , the lack of significant proper motion may argue in favor of a larger distance estimate for hbc 425 , i.e. more consistent with the distance of taurus - auriga . in this work we present near - infrared , adaptive optics imaging and spectroscopy of hbc 425c , an apparent low - mass companion positioned @xmath0 southeast of the spectroscopic binary .
this component was first noted by white & hillenbrand ( 2005 ) who estimated it to be @xmath12.5 magnitudes fainter than the primary in @xmath5band .
we use the pre - main sequence models of baraffe et al .
( 1998 ) to re - evaluate the mass and age of the hbc 425 system .
we also present high dispersion optical spectra of hbc 425 and hbc 425c and use these spectra to examine the kinematics of the system and to place limits on the abundance of lithium present in their photospheres .
the near infrared spectrograph ( nirspec ; mclean et al . 1998 ) installed behind the adaptive optics system of keck ii ( nirspao ) was used on 2010 december 12 to obtain moderate - dispersion ( @xmath3 ) @xmath20 and @xmath5 band spectra of hbc 425 and hbc 425c .
the spectrograph was configured with the 3-pixel slit and the nirspec-3 ( @xmath4band ) and nirspec-7 ( @xmath5band ) filters . using the rotator of the adaptive optics system ,
the slit was aligned with the axis of the visual binary ( position angle @[email protected] ) .
the stars were nodded along the slit and observed in 3 positions to ensure proper sky subtraction and reasonable signal to noise levels in the extracted spectra .
integration times were 180 and 120 s in the nirspec-3 and nirspec-7 filters , respectively .
flat lamp and argon arcline spectra were obtained for each instrument configuration to remove pixel to pixel variations and for wavelength calibration . a nearby , bright a0 v star ( hd 35036 )
was observed immediately after hbc 425 in each filter for telluric correction .
basic image reduction , spectral extraction , and wavelength calibration were achieved using redspec , an analysis package written by l. prato , s. kim , and i. mclean that is publically available through the nirspec webpage .
the slit viewing camera ( scam ) was used to image hbc 425 and hbc 425c in the nirspec-3 , 5 , and 7 filters ( @xmath21band , respectively ) . behind the adaptive optics system ,
the platescale of scam is 0.0168 arcsec per pixel , yielding a nominal field of view of 4@xmath223 square .
integration times per coadd were 1 s or less to maintain linearity , requiring the use of correlated double sampling ( cds ) readout mode .
detector read noise in cds mode is @xmath110 e@xmath23 .
the field was dithered with hbc 425 and hbc 425c off the slit in several positions to allow for sky and dark current subtraction .
aperture photometry was performed on hbc 425 and hbc 425c using the _ phot _ task in the daophot package of iraf on sky - subtracted images .
an aperture radius of 10 pixels ( 0@xmath22168 ) was adopted for the analysis .
photometric uncertainties are represented by the standard deviation of measurements on several sky - subtracted frames in each filter .
absolute photometric calibration was achieved by using the 2mass @xmath20 , @xmath24 , and @xmath25band magnitudes for hbc 425 .
photometry for these sources is presented in table 1 .
the high resolution echelle spectrometer ( hires ) is a grating cross - dispersed spectrograph permanently mounted on the nasmyth platform of keck i ( vogt et al . 1994 ) .
high dispersion spectra were obtained using hires with the red cross - disperser and collimator in beam on the nights of 2010 december 02 , 2011 october 12 , and 2011 october 19 ( ut ) , abbreviated henceforth by j2010 , j2011a , and j2011b , respectively . all nights were photometric with seeing conditions of @xmath10@xmath2285 for j2010 and j2011b and @xmath10@xmath2260 for j2011a .
the b2 decker ( 0@xmath22574@xmath187@xmath220 ) was used during j2010 to provide a spectral resolution of @xmath166,000 ( @xmath14.5 km s@xmath23 ) . near complete wavelength coverage from @xmath136507900
was achieved , a region that includes many gravity and temperature sensitive photospheric features as well as permitted and forbidden transitions generally associated with accretion or chromospheric activity : h & k , h@xmath26 , h@xmath27 , @xmath85876 , [ o i ] @xmath86300 , and h@xmath14 . for the j2011 observations ,
the cross - disperser angle was altered to provide expanded wavelength coverage in the red , from @xmath143008680 .
the c1 and c5 deckers ( 0@xmath22861@xmath187@xmath220 and 1@xmath22148@xmath187@xmath220 ) were used to provide spectral resolutions of @xmath150,000 ( 6.0 km s@xmath23 ) and 37,500 ( 8.0 km s@xmath23 ) , respectively . the excellent seeing conditions on 2011 october 12 allowed the placement of the slit directly on hbc 425c , perpendicular to the axis joining hbc425c to hbc 425 .
the one hour integration yielded a signal to noise level of @xmath120 near @xmath86700 .
the 3-chip mosaic of mit - ll ccds with 15 @xmath2 m pixels was used in low gain mode resulting in readout noise levels of @xmath12.8 , 3.1 , and 3.1 e@xmath23 , for the red , green , and blue detectors , respectively .
internal quartz lamps were used for flat fielding and thar lamp spectra were obtained for wavelength calibration .
several radial and rotational velocity standard stars having m0m3.5 spectral types from nidever et al .
( 2002 ) and browning et al .
( 2010 ) were observed during the nights .
the hires observations were reduced using the mauna kea echelle extraction , _ makee _ , reduction script written by tom barlow , which is publically available through links on the hires webpage .
the heliocentric radial velocities of the individual components of hbc 425 were determined by cross - correlation analysis using the m0.5 type radial velocity standard hd 28343 as a template . to distinguish primary from secondary
, we assume that the depths of absorption lines are greater for the former than for the latter , implying a slight difference in spectral type .
the j2010 hires spectrum of hbc 425 was obtained when the velocities of the two components were nearly reversed from those reported by white & hillenbrand ( 2005 ) , @xmath2810.67 and + 46.32 km s@xmath23 for the primary and secondary , respectively .
the j2011a hires spectrum reveals single component absorption line profiles , implying equal velocities of @xmath1 + 16.8 km s@xmath23 .
the j2011b radial velocities were measured to be + 41.42 and @xmath285.53 km s@xmath23 for the primary and secondary , respectively .
lacking sufficient observations for a complete orbit determination , we use the method of wilson ( 1941 ) to derive the mass ratio ( @xmath29 ) and systemic velocity ( @xmath26 ) of the system defined by : @xmath30 and @xmath31 where @xmath32 and @xmath33 are the j2010 and j2011b velocities of the primary and @xmath34 and @xmath35 are those of the secondary .
we find @xmath36 and @xmath26=17.89 km s@xmath23 .
this systemic velocity is consistent with that derived by white & hillenbrand ( 2005 ) and with the radial velocities of known taurus - auriga members . adopting the masses for the primary and secondary derived by white & hillenbrand ( 2005 ) , 0.37 m@xmath7 , and assuming circular orbits with an inclination angle of 90@xmath15 ,
an upper limit for the semi - major axis of the system is @xmath10.2 au .
an upper limit for the period of the spectroscopic binary follows , @xmath10.15 years or @xmath154 days .
it is likely , however , that these observations and that of white & hillenbrand ( 2005 ) were not made at maximum velocity separation and that the system is not observed edge - on .
we provide a summary of observed properties of the spectroscopic binary in table 2 and inferred or derived properties in table 3 .
the hires spectra of hbc 425 reveal strong , broad h@xmath14 emission with w(h@xmath14)[email protected] (j2010 ) , @xmath2840.0 ( j2011a ) , and @xmath2836.8 ( j2011b ) .
taken together with the equivalent widths reported by white & hillenbrand ( 2005 ) , w(h@xmath14)[email protected] and downes & keyes ( 1988 ) , @xmath2878 , these measurements imply significant variability .
the j2010 and j2011 h@xmath14 emission profiles are double - peaked and similar in appearance with velocity widths of @xmath1366 km s@xmath23 ( j2010 ) , 462 km s@xmath23 ( j2011a ) , and 468 km s@xmath23 ( j2011b ) , indicative of accretion using the h@xmath14 10% width of peak emission criterion of white & basri ( 2003 ) .
higher order balmer transitions ( h@xmath27 , h@xmath26 , h@xmath37 , h@xmath38 ) are also double - peaked , but with smaller velocity separations .
broad , blue - shifted wings are evident in these emission lines suggestive of accretion driven winds .
[ o i ] @xmath86300 is also found in emission with radial velocities of + 10.6 km s@xmath23 ( j2010 ) , + 18.1 km s@xmath23 ( j2011a ) , and + 20.2 km s@xmath23 ( j2011b ) .
the uncertainties for these velocities are @xmath62 km s@xmath23 .
such forbidden emission may arise from a circumbinary disk wind given its near systemic velocity ( e.g. hartigan et al .
there is no evidence of [ s ii ] @xmath106717 , 6730 emission in the hires spectra .
the h & k lines ( j2010 ) are double - peaked , but while strongly in emission the features are narrow and lack the broadened wings evident in the balmer lines .
these emission peaks are consistent with the velocities of the primary and secondary , implying that they arise from chromospheric activity associated with each component .
the d lines ( j2010 ) exhibit central absorption flanked by emission peaks that also correlate with the velocities of the primary and secondary .
the strength of these chromospheric emission lines ( i.e. , ) is greater in the secondary than in the primary .
the near - infrared triplet ( @xmath8@xmath88498 , 8542 , 8662 ) was included in the j2011 spectra .
these features exhibit emission reversal that nearly fills the wings of the underlying absorption profiles .
the radial velocities of the double emission lines present in the near - infrared triplet in the j2011b spectrum again correlate with the velocities of the primary and secondary . shown in figure 1 are sections of the j2010 , j2011a , and j2011b hires spectra including k , @xmath88542 , @xmath86102.72 ( exemplifying the double - lined features observed in the j2010 and j2011b spectra ) , and h@xmath14 .
lithium is clearly depleted in both components of the spectroscopic binary . shown in figure 2
is an expanded region of one order of the j2010 spectrum centered near @xmath86708 .
the spectrum of the m3-type main sequence star gl 806 is shown for comparison .
although weak absorption features appear at the wavelengths expected for the primary and secondary , the strengths of these features are consistent with the level of variation observed in the pseudo - continuum of the star .
the measured equivalent widths , @xmath130 m serve as upper limits that are a factor of two lower than that determined by white & hillenbrand ( 2005 ) . using the curves of growth for the @xmath86708 transition derived by pavlenko & magazzu ( 1996 ) and adopting an effective temperature of 3500 k with log @xmath39=4.5 , we determine an upper limit for the surface abundance in hbc 425 to be between log @xmath9(li)[email protected] and @xmath281.0 dex . from this
we infer that lithium is depleted in the photospheres of hbc 425a+b by at least @xmath13 orders of magnitude relative to interstellar abundance levels .
hires guide camera images ( figure 3a ) revealed the presence of a tertiary component of hbc 425 .
white & hillenbrand ( 2005 ) noted this companion and estimated it to be @xmath12.5 magnitudes fainter in @xmath5band than the primary .
the companion is separated by @xmath11@xmath40 , corresponding to a projected separation of @xmath1172 au assuming the traditionally accepted distance of taurus - auriga , ( @xmath1140@xmath610 pc : kenyon et al .
1994 ) . to estimate the probability of a chance alignment ,
we follow the example of metchev & hillenbrand ( 2009 ) and sum the number of sources within a 5 radius of hbc 425 in the 2mass point source catalog . multiplying this value by the ratio of solid angles , ( 1@xmath2223)@xmath41 / ( 5@xmath42 , we derive a probability for a purely geometric alignment of @xmath10.2% . the radial velocity of hbc 425c
is also measured to be + 15.9 km s@xmath23 , consistent with the systemic velocity of hbc 425 . shown in figure 3b
is a background - subtracted nirspao @xmath43band image of hbc 425 and hbc 425c obtained using scam . the sources are clearly separated allowing the slit to be aligned along the axis of the binary without concern for overlapping point spread functions . shown in figures 4 and 5
are the extracted @xmath4 and @xmath5band spectra respectively for hbc 425 and hbc 425c with critical atomic and molecular features identified .
superposed and plotted with additive constants are spectra for standard stars ( m3m8 types ) obtained from the irtf spectral library ( rayner et al .
weak pa@xmath27 and br@xmath26 emission are observed in the spectra of hbc 425 , consistent with accretion activity .
the strength of atomic lines and molecular features in the spectra of hbc 425 support the m3 classification of white & hillenbrand ( 2005 ) .
the @xmath4 and @xmath5band spectra of hbc 425c and its near - infrared colors are consistent with a [email protected] spectral type .
the lack of strong absorption near 2.26 @xmath2 m is suggestive of being later than m5 .
although the @xmath4band spectrum is affected by the lower throughput of the n3 filter at shorter wavelengths , the doublet near 1.17 @xmath2 m is clearly stronger in hbc 425c than in hbc 425 and is suggestive of @xmath44m5 spectral type .
there is no evidence for pa@xmath27 or br@xmath26 emission in the spectrum of hbc 425c , implying that this source is either not accreting or accreting at levels below the detection threshold for these diagnostics .
the near - infrared colors of hbc 425c are also consistent with purely photospheric emission .
it is evident from weak @xmath45 excesses that both hbc 425 and hbc 425c suffer minimal foreground extinction .
the hires spectrum of hbc 425c reveals strong @xmath86708 absorption .
shown in figure 2 are the spectra of hbc 425c and ux tau c , an m5-type taurus - auriga member , centered near @xmath86708 .
the measured equivalent width of @xmath86708 , w(li ) , for hbc 425c is @[email protected] , substantially lower than that of ux tau c , w(li)=0.6 .
this is in part due to scattered light from hbc 425 that artificially elevates the observed continuum level .
to estimate the level of scattered light contamination , we use the slit guide camera images of the hbc 425 system taken through the rg610 photometric filter . using the nearby field star j04542362@xmath461709434 to construct a model point spread function ( psf )
, we remove the psf of hbc 425 from the image and measure the flux in a circular aperture equivalent in size to the width of the c5 decker centered on hbc 425c . performing aperture photometry for hbc 425c in the orginal image , we find that @xmath12040% of the incident light likely originates from hbc 425 . removing this scattered light contribution from the continuum near @xmath86700 in the hires spectrum of hbc 425c , we measure w(li)=0.47@xmath47 .
using the lte curves of growth for the @xmath86708 transition from zapatero osorio et al .
( 2002 ) and assuming an effective temperature consistent with hbc 425c ( 29003100 k ) , we find that the measured equivalent width , w(li)@xmath10.32 , implies a lithium abundance of log @xmath9(li)@xmath11.6 dex , suggestive of substantial depletion relative to intersteller levels . correcting w(li ) for the scattered light contribution from hbc 425 , we estimate the lithium abundance level to be between log @xmath9(li)@xmath11.9 and 3.1 dex , where the latter is assumed by zapatero osorio et al .
( 2002 ) to be representative of the cosmic or primordial abundance level .
while it is possible that the abundance of lithium in hbc 425c is depleted relative to interstellar levels , we can not conclusively demonstrate this with the present seeing limited observations .
binaries play a crucial role in assessing the accuracy of pre - main sequence evolutionary models and isochrones ( e.g. hillenbrand & white 2004 , mathieu et al .
2007 ) . using the effective temperature scale of luhman et al .
( 2003 ) , extinction corrected @xmath4band magnitudes , and the main sequence bolometric corrections of kenyon & hartmann ( 1995 ) , we place hbc 425a+b and hbc 425c on the hertzsprung - rusell ( hr ) diagram shown in figure 6a ) . superimposed
are the 1 , 3 , 10 , 20 myr , and 1 gyr isochrones , and the 0.35 , 0.1 , and 0.08 m@xmath7 evolutionary tracks of baraffe et al .
( 1998 ) , placed at the nominal distance of taurus - auriga @xmath1140 pc .
the solar metallicity baraffe et al .
( 1998 ) models incorporate non - grey atmospheres and assume a general mixing length parameter given by @xmath14=1 .
each component of hbc 425 falls near the 0.35 m@xmath7 evolutionary track at an age of @xmath110 myr .
hbc 425c lies near the 0.09 m@xmath7 evolutionary track with a coeval age of @xmath1810 myr .
also depicted in figure 6a ) are the theoretical lithium depletion curves of chabrier & baraffe ( 1997 ) .
the onset of depletion is represented by the enhanced dashed curve and the cross - hatched region represents significant depletion of @xmath13 orders of magnitude or greater from initial abundance levels .
the models predict that hbc 425 should have experienced some degree of lithium depletion at its assumed age of @xmath110 myr while hbc 425c should still retain its primordial lithium abundance .
as originally noted by white & hillenbrand ( 2005 ) , if a member of taurus - auriga , the level of lithium depletion observed in hbc 425 is clearly inconsistent with the models of chabrier & baraffe ( 1997 ) .
white & hillenbrand ( 2005 ) summarize 3 possible explanations : 1 ) the effective temperature of hbc 425 is @xmath1340 k hotter than anticipated for its m3 spectral type ; 2 ) hbc 425 lies in the foreground of taurus - auriga by some @xmath150 pc and is consequently much older than previously assumed ; or 3 ) a problem exists in models of lithium depletion for low - mass , pre - main sequence stars .
we can only address the second of these hypotheses with the observations presented here .
shown in figure 6b ) is the hr diagram of the hbc 425 system placed at a distance of 90 pc .
the ages of hbc 425 and hbc 425c are coeval near @xmath125 myr . from their placement relative to the region of significant lithium depletion , both components of the spectroscopic binary should have experienced near complete lithium destruction , as observed .
the substantially lower mass hbc 425c , however , lies on or near the boundary representing the onset of lithium depletion .
if shown to be lithium depleted relative to interstellar abundance levels , even minimally , this would greatly strengthen the argument for the more advanced age ( @xmath125 myr ) and the more proximal distance ( @xmath190 pc ) to the system .
the presence of mid - infrared excess indicative of a primordial disk around hbc 425 ( hartmann et al .
2005 ) as well as strong h@xmath14 emission consistent with accretion ( white & hillenbrand 2005 and this investigation ) complicates this issue considerably .
the timescale of disk dissipation within the terrestrial region has been reasonably well - established by ground - based ( e.g. haisch et al .
2001 ; mamajek et al .
2004 ) and _ spitzer space telescope _
( e.g. uchida et al .
2004 ; silverstone et al .
2006 ) observations to be @xmath4810 myr .
the existence of an accretion disk having an age @xmath12.5 times this limit , while certainly not beyond reason , is exceptional nonetheless .
the question of lithium depletion in hbc 425 remains unresolved .
it is , however , not without precedent : e.g. hip 112312a in the @xmath112 myr old @xmath27 pictoris moving group ( song et al .
while the lithium depletion age of hip 112312a is @xmath135 myr , its isochronal age is @xmath16 myr .
high precision proper motions of hbc 425 that could unambiguously determine its membership status in taurus - auriga are needed .
clearly a dedicated effort must be made to determine the distance to this enigmatic system and thereby resolve the lithium depletion problem .
the digitized sky surveys , which were produced at the space telescope science institute under u.s .
government grant nag w-2166 , were used as were the the simbad database operated at cds , strasbourg , france , and the two micron all sky survey ( 2mass ) , a joint project of the university of massachusetts and the infrared processing and analysis center ( ipac)/california institute of technology , funded by nasa and the national science foundation .
sed is grateful to russel white , lynne hillenbrand , and g. h. herbig for insightful discussions regarding the nature of hbc 425 and to an anonymous referee whose comments and suggestions greatly improved this manuscript .
sed is also grateful to the director of w. m. keck observatory for the use of director s time in carrying out this program and to heather hershley and terry stickel for their dedicated efforts in supporting these observations . _
facilities : _ .
anders , e. , & grevesse , n. 1989 , , 53 , 197 andrews , s. m. , & williams , j. p. 2005
, , 631 , 1134 baraffe , i. , chabrier , g. , allard , f. , & hauschildt , p. h. 1998 , , 337 , 403 bertout , c. , & genova , f. 2006 , , 460 , 499 bodenheimer , p. 1965
, , 142 , 451 bonsack , w. k. , & greenstein , j. l. 1960 , , 131 , 83 chabrier , g. , & baraffe , i. 1997 , , 327 , 1039 downes , r. a. , & keyes , c. d. 1988 , , 96 , 777 haisch , k. e. , jr . , lada , e. a. , & lada , c. j. 2001 , , 553 , l153 hartigan , p. , edwards , s. , & ghandour , l. 1995 , , 452 , 736 hartmann , l. , et al .
2005 , , 628 , l147 hayashi , c. 1961 , , 13 , 450 herbig , g. h. 1965 , , 141 , 588 hillenbrand , l. a. , & white , r. j. 2004 , , 604 , 741 ireland , m. j. , & kraus , a. l. 2008 , , 678 , l59 kenyon , s. j. , dobrzycka , d. , & hartmann , l. 1994 , , 108 , 1872 kenyon , s. j. , & hartmann , l. 1995 , , 101 , 117 luhman , k. l. , stauffer , j. r. , muench , a. a. , et al . 2003 , , 593 , 1093 mamajek , e. e. , meyer , m. r. , hinz , p. m. , et al .
2004 , , 612 , 496 mathieu , r. d. , baraffe , i. , simon , m. , stassun , k. g. , & white , r. 2007 , protostars and planets v , 411 mclean , i. s. , becklin , e. e. , bendiksen , o. , et al .
1998 , , 3354 , 566 metchev , s. a. , & hillenbrand , l. a. 2009 , , 181 , 62 nidever , d. l. , marcy , g. w. , butler , r. p. , fischer , d. a. , & vogt , s. s. 2002 , , 141 , 503 pavlenko , y. v. , & magazzu , a. 1996 , , 311 , 961 rayner , j. t. , cushing , m. c. , & vacca , w. d. 2009 , , 185 , 289 silverstone , m. d. , meyer , m. r. , mamajek , e. e. , et al .
2006 , , 639 , 1138 song , i. , bessell , m. s. , & zuckerman , b. 2002 , , 581 , l43 stephenson , c. b. 1986 , , 300 , 779 uchida , k. i. , calvet , n. , hartmann , l. , et al .
2004 , , 154 , 439 vogt , s. s. , et al .
1994 , , 2198 , 362 werner , m. w. , et al .
2004 , , 154 , 1 white , r. j. , & basri , g. 2003 , , 582 , 1109 white , r. j. , & hillenbrand , l. a. 2005 , , 621 , l65 wilson , o. c. 1941 , , 93 , 29 zapatero osorio , m. r. , bjar , v. j. s. , pavlenko , y. , et al .
2002 , , 384 , 937 ccccc julian date & 2,455,532.925 & & & + w(ca ii k ) ( ) & ... & @xmath281.72 & @xmath282.37 & ... + w(h@xmath37 ) ( ) & @xmath286.43 & ... & ... & ... + w(h@xmath26 ) ( ) & @xmath288.96 & ... & ... & ... + w(h@xmath14 ) ( ) & @xmath2827.35 & ... & ... & ... + h@xmath14 10% width ( km s@xmath23 ) & 366 & .. & ... & ... + w(li i ) ( ) & ... & @xmath480.03 & @xmath480.03 & ... + radial velocity ( km s@xmath23 ) & + 17.89 & @xmath2810.67 & + 46.32 & ... + + + parameter & system & primary & secondary & hbc 425c + + julian date & 2,455,846.555 & & & + w(h@xmath26 ) ( ) & @xmath2814.48 & ... & ... & ... + w(h@xmath27 ) ( ) & @xmath2814.36 & ... & ... & @xmath285.8 + w(h@xmath14 ) ( ) & @xmath2840.0 & ... & ... & @xmath2817.2 + h@xmath14 10% width ( km s@xmath23 ) & 462 & ... & ... & 200 : + w(li i ) ( ) & ... & @xmath480.05 & @xmath480.05 & [email protected] + w(ca ii @xmath88542 ) ( ) & @xmath280.58 & ... & ... & @xmath280.2 + radial velocity ( km s@xmath23 ) & ... & + 16.77 & + 16.77 & + 15.9 + + + parameter & system & primary & secondary & hbc 425c + + julian date & 2,455,853.454 & & & + w(h@xmath26 ) ( ) & @xmath2811.52 & ... & ... & ... + w(h@xmath27 ) ( ) & @xmath2814.08 & ... & ... & ... + w(h@xmath14 ) ( ) & @xmath2836.83 & ... & ... & ... + h@xmath14 10% width ( km s@xmath23 ) & 468 & ... & ... & ... + w(li i ) ( ) & ... & @xmath480.05 & @xmath480.05 & ... + w(ca ii @xmath88542 ) ( ) & ... & @xmath280.18 & @xmath280.32 & ... + radial velocity ( km s@xmath23 ) & + 17.89 & + 41.42 & @xmath285.53 & ... + lcc mass ( m@xmath7 ) & 0.35 & 0.09 + t@xmath49 ( k ) & 3415 & 3058 + spectral type & m3 & m5.5 + log l / l@xmath7 & @xmath281.07 & @xmath281.82 + mass ratio ( q ) & 1.00 & ... + semi - major axis ( au ) & @xmath170.2 & 172 + period ( days ) & @xmath1754 & ... + | we present high angular resolution , near - infrared imaging and spectroscopy of a low - mass companion to the lithium - depleted , double - line spectroscopic binary hbc 425 ( st 34 ) obtained using the near infrared spectrograph ( nirspec ) and the keck ii adaptive optics system .
positioned @xmath0 southeast of the primary pair , the companion , hbc 425c , is @xmath12.4 magnitudes fainter at 2.2 @xmath2 m .
moderate - resolution ( @xmath3 ) @xmath4 and @xmath5band spectroscopy reveal hbc 425c to have an m5.5 ( @xmath60.5 ) spectral type .
comparisons with pre - main sequence evolutionary models imply a mass of @xmath10.09 m@xmath7 and ages of 810 myr , assuming the nominal distance of taurus - auriga ( @xmath1140 pc ) , or @xmath125 myr if placed at @xmath190 pc .
we also present high dispersion , optical spectra of hbc 425 and hbc 425c obtained using the high resolution echelle spectrometer ( hires ) on keck i. we detect strong @xmath86708 absorption in the spectrum of hbc 425c . using curves of growth for the @xmath86708 doublet
, we estimate its abundance level to lie between log @xmath9(li)=1.9 and 3.1 dex .
the spectrum of hbc 425 exhibits h & k , @xmath105876 , 6678 , and strong balmer line emission , consistent with accretion .
we place more restrictive upper limits on the surface abundance of lithium and find that hbc 425 retains less than @xmath10.1% of its primordial abundance .
the presence of lithium in the photosphere of hbc 425c does not resolve the discrepancy between isochronal and lithium depletion ages for the primary pair .
however , if lithium were depleted relative to interstellar abundance levels , even minimally , considerable support would be gained for the more advanced age of this hierarchical triple system . |
to demonstrate our method of selecting autonomous nodes we consider two er graphs with average degree @xmath11 and @xmath7 of autonomous nodes ( @xmath33 ) .
first we consider a method based on the degree of the node and later we compare with the method based on the betweenness . under a sequence of random failures ,
the networks are catastrophically fragmented when close to @xmath34 of the nodes fail , as seen in fig .
[ fig::orderparam ] . for a single er , with the same average degree
, the global connectivity is only lost after the failure of @xmath35 of the nodes .
figure [ fig::orderparam ] also shows ( line ) the results for choosing as autonomous nodes in both networks the fraction @xmath36 of the nodes with the highest degree and coupling the remaining ones at random . with this strategy ,
the robustness @xmath10 can be improved and the corresponding increase of @xmath37 is about @xmath38 , from close to @xmath39 to close to @xmath40 .
also the order of the transition changes from first to second order .
further improvement can be achieved if additionally the coupled nodes are paired according to their position in the ranking of degree , since interconnecting similar nodes increases the global robustness @xcite . in the inset of fig .
[ fig::orderparam ] we see the dependence on @xmath9 of the relative robustness for the degree strategy compared to the random case @xmath41 . for the entire range of @xmath9 the proposed strategy is more efficient and a relative improvement of more than @xmath42 is observed when still @xmath43 of the nodes are coupled .
two types of technological challenges are at hand : either a system has to be designed robust from scratch or it already exists , constrained to a certain topology , but requires improved robustness . in the former case , the best procedure is to choose as autonomous the nodes with highest degree in each network and couple the others based on their rank of degree . for the latter , rewiring is usually a time - consuming and expensive process , and the creation of new autonomous nodes may be economically more feasible .
the simplest procedure consists in choosing as autonomous both nodes connected by the same inter - network link .
however , in general , the degree of a pair of coupled nodes is not , necessarily , correlated . in fig .
[ fig::orderparam ] we compare between choosing the autonomous pairs based on the degree of the node in network @xmath1 or in network @xmath8 . when pairs of nodes are picked based on their rank in the network under the initial failure ( network @xmath1 ) , the robustness almost does not improve compared to choosing randomly . if , on the other hand , network @xmath8 is considered , the robustness is significantly improved , revealing that this scheme is more efficient .
this asymmetry between @xmath1 and @xmath8 network is due to the fact that we attack only nodes in network @xmath1 , triggering the cascade , that initially shuts down the corresponding @xmath8-node .
the degree of this @xmath8-node is related to the number of nodes which become disconnected from the main cluster and consequently affect back the network @xmath1 . therefore , the control of the degree of vulnerable @xmath8-nodes is a key mechanism to downsize the cascade . on the other hand , when a hub is protected in network @xmath1 it can still be attacked since the initial attack does not distinguish between autonomous and non - autonomous nodes . in fig .
[ fig::robustness](a ) we compare four different criteria to select the autonomous nodes : betweenness , degree , k - shell , and random choice , for two coupled er networks . in the betweenness strategy , the selected autonomous are the ones with highest betweenness .
the betweenness is defined as the number of shortest paths between all pairs of nodes passing through the node @xcite . in the k - shell strategy , the autonomous are the ones with highest k - shell in the k - shell decomposition @xcite . the remaining nodes , for all cases , have been randomly inter - linked . since er
networks are characterized by a small number of k - shells , this strategy is even less efficient than the random strategy for some values of @xmath9 , while the improved robustness for degree and betweenness strategies is evident compared with the random selection . while in the random case , for @xmath44 , a significant decrease of the robustness with @xmath9 is observed , in the degree and betweenness cases , the change is smoother and only significantly drops for higher values of @xmath9 .
a maximum in the ratio @xmath41 occurs for @xmath45 , where the relative improvement is above @xmath46 . since in random networks ,
the degree of a node is strongly correlated with its betweenness @xcite , their results are similar .
many real - world systems are characterized by a degree distribution which is scale free with a degree exponent @xmath47 @xcite . in fig .
[ fig::robustness](b ) we plot @xmath10 as a function of @xmath9 for two coupled scale - free networks ( sf ) with @xmath14 nodes each and @xmath12 . similar to the two coupled er , this system is also significantly more resilient when the autonomous nodes are selected according to the highest degree or betweenness . for values of @xmath48 the robustness is similar to that of a single network ( @xmath49 ) since the most relevant nodes are decoupled .
a peak in the relative robustness , @xmath41 ( see inset of fig . [ fig::robustness]b ) , occurs for @xmath50 where the improvement , compared to the random case , is almost @xmath51 .
betweenness , degree , and @xmath52-shell , have similar impact on the robustness since these three properties are strongly correlated for sf . from fig .
[ fig::robustness ] , we see that , for both sf and er , the robustness is significantly improved by decoupling , based on the betweenness , less than @xmath42 of the nodes . studying the dependence of the robustness on the average degree of the nodes we conclude that for average degree larger than five , even @xmath53 autonomous nodes are enough to achieve more than @xmath54 of the maximum possible improvement . for the cases discussed in fig .
[ fig::robustness ] , results obtained by selecting autonomous nodes based on the highest degree do not significantly differ from the ones based on the highest betweenness .
this is due to the well known finding that for erds - rnyi and scale - free networks , the degree of a node is strongly correlated with its betweenness @xcite . however ,
many real networks are modular , i.e. , composed of several different modules interconnected by less links , and then nodes with higher betweenness are not , necessarily , the ones with the largest degree @xcite .
modularity can be found , for example , in metabolic systems , neural networks , social networks , or infrastructures @xcite . in fig .
[ fig::fourer ] we plot the robustness for two coupled modular networks .
each modular network was generated from a set of four erds - rnyi networks , of @xmath16 nodes each and average degree five , where an additional link was randomly included between each pair of modules . for a modular network , the nodes with higher betweenness are not necessarily the high - degree nodes but the ones bridging the different modules .
figure [ fig::fourer ] shows that the strategy based on the betweenness emerges as better compared to the high degree method .
another example that shows that betweenness is superior to degree is when we study coupled random regular graphs . in random regular graphs all nodes have the same degree and are connected randomly .
figure [ fig::rr ] shows the dependence of the robustness on the degree of coupling , for two interconnected random regular graphs with degree @xmath55 .
the autonomous nodes are selected randomly ( since all degrees are the same ) or following the betweenness strategy . though all nodes have the same degree and the betweenness distribution is narrow ,
selecting autonomous nodes based on the betweenness is always more efficient than the random selection .
thus , the above two examples suggest that betweenness is a superior method to chose the autonomous nodes compared to degree .
the vulnerability is strongly related to the degree of coupling @xmath9 .
_ @xcite have analytically and numerically shown that , for random coupling , at a critical coupling @xmath56 , the transition changes from continuous ( for @xmath57 ) to discontinuous ( for @xmath58 ) . in fig .
[ fig::diagram ] we see the two - parameter diagram ( @xmath37 vs @xmath9 ) with the tricritical point and the transition lines ( continuous and discontinuous ) for the random ( inset ) and the degree ( main ) strategies . as seen in fig .
[ fig::diagram ] , when autonomous nodes are randomly selected , about @xmath38 autonomous nodes are required to soften the transition and avoid catastrophic cascades , while following the strategy proposed here only a relatively small amount ( @xmath59 ) of autonomous nodes are needed to avoid a discontinuous collapse . above the tricritical point , the jump increases with the degree of coupling , lending arguments to the paramount necessity of an efficient strategy for autonomous selection , given that the fraction of nodes which can be decoupled is typically limited .
the dependence of @xmath60 on the average degree @xmath19 is shown in fig .
[ fig::qtk ] .
the ratio between the tricritical coupling for degree and random strategies increases with decreasing @xmath19 .
for example , for the fraction of autonomous nodes needed to soften the transition with the random selection is six times the one for the degree strategy . as in ref .
@xcite , following the theory of riedel and wegner @xcite , we can characterize the tricritical point .
two relevant scaling fields are defined : one tangent ( @xmath20 ) and the other perpendicular ( @xmath61 ) to the critical curve at the tricritical point . in these
coordinate axes the continuous line is described by @xmath62 , where the tricritical crossover exponent @xmath63 for degree and random strategies .
the tricritical order parameter exponent , @xmath21 , can be evaluated from , @xmath64 giving @xmath65 for both strategies . since these two exponents are strategy independent ( see fig .
[ fig::crossover ] ) , we conjecture that the tricritical point for degree and random selection are in the same universality class .
here , we propose a method to chose the autonomous nodes in order to optimize the robustness of coupled networks to failures .
we find the betweenness and the degree to be the key parameters for the selection of such nodes and we disclose the former as the most effective for modular networks .
considering the real case of the italian communication network coupled with the power grid , we show in fig .
[ fig::italy ] that protecting only the four communication servers with highest betweenness reduces the chances of catastrophic failures like that witnessed during the blackout in 2003 . when this strategy is implemented the resilience to random failures or attacks
is significantly improved and the fraction of autonomous nodes necessary to change the nature of the percolation transition , from discontinuous to continuous , is significantly reduced .
we also show that , even for networks with a narrow diversity of nodes like erds - rnyi graphs , the robustness can be significantly improved by properly choosing a small fraction of nodes to be autonomous . as a follow - up
it would be interesting to understand how correlation between nodes , as well as dynamic processes on the network , can influence the selection of autonomous nodes . besides , the cascade phenomena and the mitigation of vulnerabilities on regular lattices and geographically embedded networks are still open questions .
we consider two coupled networks , @xmath1 and @xmath8 , where a fraction of @xmath3 @xmath1-nodes fails . the cascade of failures can be described by the iterative equations @xcite @xmath66 , \\
\alpha_n & = & p \left ( 1 - q_{\beta , n } \left [ 1 - s_{\beta_{n-1}b}(\beta_{n-1 } ) \right ] \right ) , \nonumber\end{aligned}\ ] ] where @xmath27 and @xmath28 are , respectively , the fraction of @xmath1 and @xmath8 surviving nodes at iteration step @xmath29 ( not necessarily in the largest component ) , and @xmath67 ( @xmath68 , @xmath69 ) is the fraction of nodes in the largest component in network @xmath70 given that @xmath71 nodes have failed .
this can be calculated using generating functions . @xmath72
( @xmath73 ) is the fraction of nodes in network @xmath1 ( @xmath8 ) which have fragmented from the giant cluster ; and changes as the iterative procedure advances . as proposed by parshani _
@xcite , when autonomous nodes are randomly selected and the degree of coupling is the same in @xmath1 and @xmath8 , the set of eqs .
[ iteration ] simplifies to @xmath74 , \\
\alpha_n & = & p \left ( 1 - q \left [ 1 - s_b(\beta_{n-1 } ) \right ] \right ) , \nonumber\end{aligned}\ ] ] where @xmath9 is the degree of coupling .
the degree distribution of the networks does not change in the case of random failures and , therefore , @xmath67 is simplified to @xmath75 , which can be calculated as @xcite @xmath76 where @xmath77 is the generating function of the degree distribution of network @xmath70 , @xmath78 and @xmath79 satisfies the transcendental equation @xmath80 when autonomous nodes are selected following the degree strategy , the fraction of dependent nodes @xmath85 changes with the iteration step @xmath29 and the set of eqs . [ iteration ] no longer simplifies .
we divide the discussion below into three different parts : the degree distribution , the largest component , and the coupling ( fraction of dependent nodes ) .
the networks a and b are characterized by their degree distributions , @xmath86 and @xmath87 , which are not necessarily the same .
the developed formalism applies to any arbitrary degree distribution .
we start by first splitting the degree distribution into two parts , the component corresponding to the low - degree dependent nodes , @xmath88 , and the component corresponding to the high - degree autonomous ones , @xmath89 . to accomplish this ,
one must determine two parameters , @xmath90 and @xmath91 , from the relations @xmath92 and @xmath93 where @xmath94 is the initial degree of coupling and @xmath91 is the fraction of nodes with degree @xmath90 that are dependent , i.e. , coupled with nodes in the other network .
one can then write @xmath95 and @xmath96 in the model , a fraction of @xmath3 @xmath1-nodes are randomly removed . if , at iteration step @xmath29 , @xmath27 nodes survive , @xmath97 nodes are necessarily autonomous and the remaining ones , @xmath98 , are dependent nodes .
one can then show that the degree distribution of network @xmath1 , under the failure of @xmath99 nodes , @xmath100 , is given by @xmath101 while the fraction of surviving links is @xmath102 with the degree distribution @xmath106 and the fraction of surviving links @xmath107 one can calculate the size of the largest component as @xmath108 where @xmath109 satisfies the self consistent equation @xmath110 to calculate the fraction @xmath72 ( and @xmath73 ) one must first calculate the degree distribution of the nodes in the largest component .
this is given by @xmath111 the fraction of nodes in the largest component that are autonomous is then given by @xmath112 so that the fraction of autonomous nodes from the original network remaining in the largest component is @xmath113 while the total fraction of autonomous nodes is given by @xmath114 the fraction of nodes disconnected from the largest component that are autonomous is then given by @xmath115 so that the fraction of dependent nodes which have fragmented from the largest component is @xmath116 for simplicity , here we consider that @xmath117 and @xmath91 are constant and do not change during the iterative process .
in fact , this is an approximation as the degrees of the autonomous nodes are expected to change when their neighbors fail .
however , in spite of shifting the transition point , this consideration does not change the global picture described here . where @xmath119 is the number of node failures , @xmath120 the size of the largest connected cluster in a network after @xmath119 failures , and @xmath121 is the total number of nodes in the network @xcite .
this definition corresponds to the area under the curve of the fraction of nodes , in the largest connected cluster , as a function of the fraction of failed nodes ( shown in fig .
we extend this definition to coupled systems by performing the same measurement , given by eq .
( [ eq::def.robustness ] ) , only on the network where the random failures occur , namely , network @xmath1 .
we acknowledge financial support from the eth risk center , from the swiss national science foundation ( grant no .
200021 - 126853 ) , and the grant number fp7 - 319968 of the european research council .
we thank the brazilian agencies cnpq , capes and funcap , and the grant cnpq / funcap .
sh acknowledges the european epiwork project , the israel science foundation , onr , dfg , and dtra . | natural and technological interdependent systems are highly vulnerable due to dependencies which pose a huge integrative risk due to the recently discovered abrupt collapse under failure . mitigating the risk by partial disconnection endangers their functionality . here
we propose a systematic strategy of establishing a minimum number of autonomous nodes that guarantee a smooth transition in the robustness .
our method which is based on betweenness is tested on various examples including the famous @xmath0 electrical blackout of italy .
we show that we can reduce the necessary number of autonomous nodes by a factor of five compared to the random choice which for practical purposes means the recovery of functionality .
we also find that the transition to abrupt collapse follows tricritical scaling characterized by a set of exponents which is independent on the protection strategy .
interconnected complex networks are ubiquitous in today s world .
they control infrastructures of modern society ( energy - communication - transportation ) , the financial system or even the human body @xcite .
unfortunately , they are much more fragile than uncoupled networks as recently recognized through the finding that the robustness changes from a second order transition in uncoupled systems to first order in interdependent systems @xcite .
the obvious mitigation strategy consists in partially decoupling the networks by the creation of autonomous nodes @xcite .
too much disconnection however risks endangering the functionality of the system .
the question which we will address here is how to reduce fragility without losing functionality and we will in fact answer this question by developing an explicit algorithm based on betweenness that enables to avoid the abrupt collapse with a minimum number of autonomous nodes .
communication servers . in a )
all communication servers are coupled while in b ) four servers have been decoupled following the strategy proposed here .
the coupling between the networks was established based on the geographical location of the nodes , such that each communication server is coupled with the closest power station @xcite .
[ fig::italy ] ] fraction of @xmath1-nodes in the largest connected cluster , @xmath2 , as a function of the fraction of randomly removed nodes @xmath3 from network a , for two coupled er ( average degree @xmath4 ) with @xmath5 of the nodes connected by inter - network links ( @xmath6 ) .
it is seen that robustness can significantly be improved by properly selecting the autonomous nodes .
we start with two fully interconnected er and decouple @xmath7 of their nodes according to three strategies : randomly ( line ) , the ones with highest degree in network @xmath1 ( line ) and in network @xmath8 ( line ) .
we also include the case where @xmath7 autonomous nodes in both networks are chosen as the ones with highest degree and all the others are interconnected randomly ( line ) .
the inset shows the dependence of the relative robustness of the degree strategy on the degree of coupling @xmath9 compared with the random case .
results for the degree have been obtained with the formalism of generation functions ( see _ methods _ ) . ]
dependence of the robustness , @xmath10 , on the degree of coupling , @xmath9 , for two , interconnected , ( a ) er ( average degree @xmath11 ) and ( b ) sf with degree exponent @xmath12 . applying our proposed strategy is applied , the optimal fraction of autonomous nodes is relatively very small .
autonomous nodes are chosen in four different ways : randomly ( ) , high degree ( ) , high betweenness ( ) , and high k - shell ( ) . the insets show the relative improvement of the robustness , for the different strategies of autonomous selection compared with the random case .
results have been averaged over @xmath13 configurations of two networks with @xmath14 nodes each . for each configuration we averaged over @xmath14 sequences of random attacks .
, title="fig : " ] + dependence of the robustness , @xmath10 , on the degree of coupling , @xmath9 , for two , randomly interconnected modular networks with @xmath15 nodes each .
the modular networks were obtained from four erds - rnyi networks , with @xmath16 nodes each and average degree five , by randomly connecting each pair of modules with an additional link .
autonomous nodes are selected in three different ways : randomly ( blue triangles ) , higher degree ( black dots ) , and higher betweenness ( red stars ) . in the inset
we see the relative enhancement of the robustness , for the second and third schemes of autonomous selection compared with the random case .
results have been averaged over @xmath13 configurations and @xmath14 sequences of random attacks to each one . ]
dependence of the robustness , @xmath10 , on the degree of coupling , @xmath9 , for two , randomly interconnected random regular graphs with @xmath17 nodes each , all with degree four .
autonomous nodes are selected in two different ways : randomly ( blue triangles ) and higher betweenness ( red stars ) . in the inset
the relative enhancement of the robustness is shown for the betweenness compared to the random case .
results have been averaged over @xmath13 configurations and @xmath14 sequences of random attacks to each one . ] ) under random attack .
the horizontal axis is the degree of coupling @xmath9 and the vertical one is @xmath18 so that @xmath3 is the fraction of initially removed nodes .
the size of the jump in the fraction of @xmath1-nodes in the largest connected cluster is also included ( curve ) . the dashed curve stands for a discontinuous transition while the solid one is a critical line ( continuous transition ) .
the two lines meet at a tricritical point ( tp ) .
autonomous nodes are selected based on the degree ( main plot ) and randomly ( inset ) .
results have been obtained with the formalism of generating functions .
[ fig::diagram ] ] dependence on the average degree @xmath19 for two coupled er , showing that the fraction of autonomous nodes to smoothen out the transition is significantly reduced with the proposed strategy when compared with the random case .
autonomous nodes are selected following two different strategies : randomly ( red squares ) and high degree ( black circles ) .
[ fig::qtk ] ] -nodes in the largest connected cluster on the scaling field @xmath20 along the direction perpendicular to the transition line at the tricritical point .
the slope is the tricritical exponent @xmath21 related with the order parameter .
autonomous nodes in the two coupled er have been selected randomly ( red line ) and following the ranking of degree ( black line ) .
[ fig::crossover ] ] buldyrev _ et al . _
@xcite proposed a percolation framework to study two coupled networks , @xmath1 and @xmath8 , where each @xmath1-node is coupled to a @xmath8-node , via bi - directional links , such that when one fails the other can not function either .
the removal of a fraction of @xmath1-nodes may trigger a domino effect where , not only their counterparts in @xmath8 fail , but all nodes that become disconnected from the giant cluster of both networks also fail .
this causes further cascading of failures , yielding an abrupt collapse of connectivity , characterized by a discontinuous ( first order ) percolation transition .
parshani _
et al . _
@xcite showed that damage can be mitigated by decreasing the degree of coupling , but only if a significant fraction ( @xmath22 ) of nodes is decoupled , the transition changes from discontinuous to continuous .
the coupling is reduced by randomly selecting a fraction of nodes to become autonomous and , therefore , independent on the other network . for the coupling between power stations and communication servers , for example , autonomous power stations have alternative communication systems which are used when the server fails and an autonomous server has its own energy power supply .
we propose a method , based on degree and centrality , to identify these autonomous nodes that maximize the system robustness .
we show that , with this scheme , the critical coupling increases , i.e. , the fraction of nodes that needs to be decoupled to smoothen out the transition is much smaller ( close to @xmath23 compared to @xmath24 ) .
significant improvement is observed for different coupled networks including for erds - rnyi graphs ( er ) where such improvement in the robustness was unexpected given their narrow degree distribution . to demonstrate the strength of our approach , in fig .
[ fig::italy ] we apply the proposed strategy to the real coupled system in italy @xcite and show that by only protecting four servers the robustness is significantly improved ( details in the figure caption ) .
we consider a pair of networks , @xmath1 and @xmath8 , where a fraction @xmath9 ( degree of coupling ) of @xmath1-nodes are coupled with @xmath8-nodes . to be functional , nodes need to be connected to the giant cluster of their network .
when an @xmath1-node fails , the corresponding one in @xmath8 can not function either .
consequently , all nodes bridged to the largest cluster through these nodes , together with their counterpart in the other network , become also deactivated .
a cascade of failures occurs with drastic effects on the global connectivity @xcite .
this process can also be treated as an epidemic spreading @xcite . as explained in the section _ methods _ , to study the resilience to failures , we follow the size of the largest connected cluster of active @xmath1-nodes , under a sequence of random irreversible attacks to network @xmath1 . extending the definition proposed for single networks
@xcite , we quantify the robustness @xmath10 as the area under the graph of the largest cluster as function of the fraction of node failures ( see _ methods _ ) . to follow the cascade triggered by the failure of a fraction @xmath3 of @xmath1-nodes , similar to @xcite , we solve the iterative equations , @xmath25 , \\
\alpha_n & = & p\left(1-q_{\alpha , n}\left[1-s_{\beta_{n-1}b}\left(\beta_{n-1}\right)\right]\right ) , \end{aligned}\ ] ] with the initial condition @xmath26 , where @xmath27 and @xmath28 are the fraction of @xmath1 and @xmath8 surviving nodes at iteration step @xmath29 and @xmath30 is the fraction of such nodes in the giant cluster .
@xmath31 is the fraction of dependent nodes in network @xmath32 fragmented from the largest cluster ( see _ methods _ for further details ) . |
long slit , spectroscopic blind surveys targeting the hi ly@xmath0 emission line have the potential to deliver detailed and otherwise unavailable insights into the gas dynamics and , in conjunction with deep , space - based imaging , the star - gas interactions in proto - galactic halos and the intergalactic medium .
several surveys of this kind ( rauch et al 2008 , rauch et al 2011 , paper i ; rauch et al 2013a , paper ii ; and rauch et al 2013b , paper iii ) have discovered a distinct subpopulation of of extended , asymmetric ly@xmath0 emitters at @xmath1 , with a comoving space density on the order of @xmath2 mpc@xmath3 and typical observed line fluxes of a few times @xmath4erg @xmath5 s@xmath6 . with a large number of processes capable of producing ly@xmath0 radiation
, one may expect the emitters to be drawn from a highly inhomogeneous group of objects .
however , the selection by ly@xmath0 emission is likely to favor galaxies in certain phases of their formation , when the stellar populations and gas dynamics are particularly conducive to the production and escape of ly@xmath0 photons .
the peculiar spatial distribution and clustering behaviour of ly@xmath0 emitters suggests that environmental effects and interactions may play an important role in determining whether a galaxy appears as a ly@xmath0 emitter ( e.g. , hamana et al 2004 ; hayashino et al 2004 ; kovac et al 2007 ; cooke et al 2010 , 2013 ; zheng et al 2011 ; matsuda et al 2012 ) .
indeed , all four extended emitters described so far in the papers in this series exhibit signs of interactions .
the duration of the processes leading to the production of ionizing radiation ( e.g. , the lifetimes of massive stars , agn activity ) , as well as the astrophysical timescales relevant for the emission of ly@xmath0 in high redshift gaseous halos ( recombination- and resonance - line radiative - transfer time scales ) tend to be short compared to the dynamical timescales and life times of the general stellar population .
thus the spectroscopic detection of such a halo amounts to a `` snapshot '' of a particularly interactive phase in their formation , illuminated by a `` flash '' of ly@xmath0 emission . among those extended ly@xmath0 emitters published to date
, the first one showed diffuse stellar features , in addition to a clear detection of the infall of cold gas into an ordinary high redshift galaxy ( paper i ) .
paper ii described what may be a milky - way - sized halo with multiple galaxies hosting disturbed , partly young stellar populations .
a thin filament apparently dominated by high equivalent width ly@xmath0 emission may reflect recent intra - halo star formation in a tidal tail or in the wake of a satellite galaxy .
the third object , revealing the only case in this sample clearly related to non - stellar processes , is an agn illuminating a satellite galaxy , possibly triggering the formation of very young stars in its halo ( paper iii ) .
the object examined in the present paper is a large halo surrounded by ly@xmath0 filaments illuminated by a group of distorted , mostly blue galaxies .
as we shall argue below , the interaction in this case appears to be between the galaxies and a gaseous medium through which they move , and which appears to strip off part of their gas and induce star formation in their wake .
the observations are described in the next section , followed by a description of the galaxies coinciding with the gaseous halo .
the presence of spatially extended metal emission , the nature of the ly@xmath0 filaments and the energetics of the emission are then discussed .
the paper concludes with a discussion of the likely nature of the phenomenon and its significance for the metal enrichment of the intergalactic medium and for the escape of photons responsible for its ionization .
[ cols=">,^,^,^,^,^,^,^,^",options="header " , ] + comments : a ) giavalisco et al 2004 ; b ) adopting @xmath7 flux error for b ; c ) adopting @xmath7 flux error for v the observations consist of a long slit , spectroscopic blind survey , with the slit positioned in precise n - s orientation on the object j123647.05 + 621237.2 in the hubble deep field north ( hdfn ) .
data were taken with the lris ( oke et al 1995 ; mccarthy et al 1998 , steidel et al 2004 ) b and r arms and the d560 dichroic , using the 600/4000 grism in 2x2 binning ( blue side ) and the 600/7500 grating in 1x1 binning ( red side ) , through a custom long slit built from two slit segments with a combined size of 2 `` @xmath8 430 '' .
total exposure times of 35.8 hours in the blue and 35 hours in the red arm of lris were obtained in march 2008 , may 2008 , and april 2009 .
the resulting 1-@xmath9 surface brightness detection limit , measured for a @xmath10 wide aperture , is approximately @xmath11 erg @xmath5 s@xmath12 .
the spectrum shows a strong ly@xmath0 emission complex near 5125.0 ( fig .
[ cuts ] ) , with the peak of the line corresponding to a redshift 3.2158 .
the identification as ly@xmath0 is supported by the large spatial extent of the emission , and the drop in the continuum blueward of the emission line , caused by the ly@xmath0 forest ( fig.[onspec ] ) .
the ly@xmath0 line is broad ( fwhm @xmath13 kms@xmath6 ) and shows a red shoulder , with part of the width coming from superposition of multiple sources ( see below ) .
after subtraction of the continuum trace , the total ly@xmath0 flux that passed through the slit is ( @xmath14 erg @xmath5 s@xmath6 .
inspection of archival goods - n v2.0 images ( giavalisco et al 2004 ; fig .
[ specplusim ] , top right panel ) shows the continuum source closest to the inferred slit position to be goods j123646.84 + 621608.1 .
a photometric redshift of 3.09 ( xue et al 2011 ) for that galaxy , referred to as object 1 in the table and the figures , agrees well with our spectroscopic redshift .
there is no prior redshift information for any of the other faint galaxies , except for a few foreground objects in the n - w corner of the field .
the ly@xmath0 emission line is remarkable in that several filaments extrude to distances of at least 27 proper kpc ( in projection ) away from the continuum position .
at least one of these structures corresponds to multiple features detected in the hst acs f435w ( `` b band '' ) ( rhs panels of fig . [ specplusim ] ) . in an area extending to about 5 arcsecs on either side of the slit ,
two highly distorted , tadpole - shaped galaxies ( 1 and 2 ; with object 1 producing the continuum in the spectrum ) can be seen , both have substructure with multiple cores of blue light in their heads ( figs.[tads ] , [ bowshock ] ) .
several more extended low surface brightness features can be seen further afield ( 3,4,5,6,7 ) partly also with ( less distinctive ) head and tail structures . among these
are blue sources with ( objects 5 and 8) and without ( object 10 ) tails , and at least one object ( number 3 ) with an apparent linear size of more than 40 kpc . of the features in the ly@xmath0 line , the central emission `` @xmath0 '' corresponds to tadpole 1 in the image , `` @xmath15 '' to the smudge near number 9 , `` @xmath16 '' to 7 , and `` @xmath17 '' to 10 .
the latter two also have very faint continua recognizable in some regions of the spectrum that line up very well with the ly@xmath0 emission , showing that the @xmath18 filament is subtended by at least two galaxies .
the other significant ly@xmath0 smudges with greek letters , @xmath19 and @xmath20 , plus the vaguely visible elongated features on either side of @xmath19 may also have counterparts in the broad band , with @xmath20 possibly related to object 4 ( see below ) .
the various white contours in the bottom right panel of fig .
[ specplusim ] , are marking extraction windows used to determine the fluxes listed in table 1 . even though the low surface brightness features are statistically significant detections ( objects 3 , 8 , and 7 are 5.0 , 4.4 , and 8.1 @xmath9 excursions in the b - band ) some of them could be observational artifacts .
moreover , even if they are real objects , they may not all be at the same redshift as the ly@xmath0 .
however , the spectroscopic identifications of objects 1 , 7 , and 10 , and the simultaneous presence of several highly distorted and/or blue ( @xmath21 ) objects suggest that most of the features are likely to be real and related to the ly@xmath0 halo
. further distorted objects in the b band occur to the n - w of the present region ( not shown ) .
if these belong to the same large scale structure ( which , due to a lack of redshifts is uncertain ) , the total maximum extent could be as large as 210 kpc proper . with the field being located near the edge of the goods - n imaging coverage we can not follow the structure further to the north .
the various objects shown in the broad band figure are listed with their identifying numbers in column 1 of table 1 , together with their name in the goods - n catalog ( where detected ) in column 2 .
the goods - n broad band b and b - v colors ( columns 3 and 4 ) are followed in columns 5 - 8 by the same quantities determined directly in extraction windows placed on the head and tail regions of the various objects shown in fig . [ specplusim ] . note that no corrections for extinction or intergalactic absorption have been applied . from the work by meiksin ( 2006 )
, we expect that a color correction of about @xmath22 ( for young star forming galaxies at z=3.2 ) needs to be added to the magnitudes in the table .
the magnitudes of the `` heads '' show good correlation with the goods - n `` magbest '' magnitudes where available for the same object . because of the irregular apertures used , the absolute magnitudes are not precisely comparable among objects .
the useful information mainly lies in the b - v colors and their gradients between heads and tails .
objects 2 , 4 , 5 , 6 , 7 , 9 , and 10 are generally blue ( overall b - [email protected] in the goods - n catalog , or , where uncatalogued , in our apertures ) .
we caution not to read too much into the formally very blue colors as the sometimes very large formal errors speak for themselves
. however , there is a clear trend in that all the cases where there is a distinct head - tail shape ( objects 1 , 2 , 3 , 5 , and 8) , with the exception of object 4 , have bluer tails than heads .
the tendency for the tails to be bluer than the heads is different from the prevailing pattern for tadpole galaxies ( e.g. , elmegreen & elmegreen 2010 ) .
the object 4 is an interesting case in that the tail is only visible in the v band ( fig .
[ tad4 ] ) , with a total flux of @xmath24erg @xmath5 s@xmath6 , corresponding to a 4.8 @xmath9 detection
. the situation here may be similar to the one presented in paper ii , a narrow filament of emission visible in only one broad band , that may be glowing in ly@xmath0 , except that in the current situation , the ly@xmath0 line , due to its higher redshift , would be situated in the v band and not in b. because of the low flux , other broad band images do not provide useful constraints ( e.g. , v - i = @xmath25 ) .
we can , however , put limits on the equivalent width from the b and v band fluxes , assuming that the b band measures the continuum , and the v band the continuum plus line flux , with the same assumptions as used in paper ii .
as the equivalent width is directly dependent on the ratio of v - band flux to b band flux , we estimate an equivalent width lower limit by setting the v - band flux to its @xmath26 excursion , and the b - band flux ( which is formally measured to be zero ) to its positive , @xmath27 value . the resulting rest frame equivalent width lower limit for n=1 is @xmath28 . for n=2 ,
the value is already a rather moderate 57 , and the `` 3@xmath9 '' result is compatible with both the v and b band just containing pure continuum emission of a flat spectrum source .
however , if the observed v band flux were mostly ly@xmath0 emission , the @xmath24erg @xmath5 s@xmath6 should be easily seen in a spectrum as it is more than two times the flux of in the main halo . in this case
, the fact that the western edge of the slit is about 3.9 " away from the tail of 4 is likely preventing us from detecting a strong ly@xmath0 emission line .
it is intriguing that the ly@xmath0 feature @xmath20 , though stronger than either @xmath16 or @xmath17 , which both correspond to galaxies clearly detected , does not have a corresponding broad band object near enough to the slit to explain the relatively strong signal .
thus @xmath20 may just be a small fraction of the ly@xmath0 emission from the tail of object 4 , spilling over into the slit . aside from the ly@xmath0 emission , several other extended emission features can be seen in the spectrum .
he ii 1640.4 is noisy but clearly present near an observed wavelength of 6914 . in the spatial direction it ranges over several arcsecs ( fig .
[ he_hi_portraits ] ) .
the presence of multiple residuals from the subtraction of sky lines and a strong spatial gradient in the sky background make a flux measurement difficult , but we estimate that the flux is about @xmath29 erg @xmath5s@xmath6 ( statistical errors only ) , within a distance of 4.3 `` on either side of the continuum trace . formally , this is approximately 19% of the hi ly@xmath0 flux .
this comparison with hi should not be taken too serious as heii 1640 is likely to be less optically thick than hi ly@xmath0 , so its spatial width may largely reflect the spatial extent of the emitting gas , whereas a substantial fraction of hi ly@xmath0 may have been scattered further outside of the slit .
thus , the observed heii 1640/ hi ly@xmath0 flux ratio may be a considerable overestimate . if we approximate the spatial profile of the ly@xmath0 halo by a circular gaussian profile with the radius set to the spatial extent _ along _ the slit ( fwhm=1.9 '' ) , only about 26% of the ly@xmath0 emission would have passed through the slit .
this could reduce the heii / ly@xmath0 to 0.05 , a value consistent with several different ionization mechanisms ( see the discussions in yang et al 2006 ; prescott , dey & jannuzi 2009 , and scarlata et al 2009 ) . several features belonging to metal line transitions can be seen in the one - dimensional spectrum ( fig .
[ onspec ] ) , including an absorption / emission complex , with the peak emission occuring near near 5234.6 ( fwhm @xmath30 kms@xmath6 ) , probably related to the nv 1238/1242 doublet , and absorption troughs at 5302.5 and 5418.8 .
we identify the lower redshift one with the siii 1260 resonance line .
a further , broader trough near 5480 almost certainly corresponds to the oi / siii 1302 - 1304 complex often seen in lyman break galaxies ( e.g. , shapley et al 2003 )
. the low signal - to - noise ratio does not permit a very meaningful analysis , but single component , gaussian fits to the absorption troughs at 5302.5 and 5418.8 show them to be blue - shifted with respect to the nominal redshift from the peak of the ly@xmath0 line , z=3.2158 , by about -630 and -500 km s@xmath6 , with rest frame equivalent widths on the order of 3 . in several wavelength regions below and above ly@xmath0
, the spectral trace seems to be broadened , suggesting spatially extended emission line regions .
we show some of these regions in ( fig .
[ juxtaspec ] ) , indicating the spatial correspondence between spectrum and image by dashed lines . to give an idea of the flux levels and significance of the excess emission along the slit , fig .
[ spatial_cuts ] shows some of the spatial profiles after collapsing the regions in the previous figure along the dispersion direction .
note that an average continuum spectrum has been subtracted in fig.[spatial_cuts ] , and the fluxes represent just the excess emission . because of the low signal - to - noise ratio , the presence of ly@xmath0 forest absorption , and the unknown mechanisms populating the atomic levels we found it difficult to assign the extended emission to definite transitions .
the few identifications suggested here are tentative and may change with better data : the first region ( top panel in fig .
[ juxtaspec ] ) comprises relatively prominent emission extended to both the north and south of tadpole 1 .
the spectral traces of two other objects , 10 and 7 in fig.[specplusim ] , are visible as well just below the trace of the main galaxy for a wavelength stretch of about 50 in the rest frame , of which we show the central part in this panel . having enhanced emission at the same wavelength as the main tadpole strengthens our proposed identification of these objects with galaxies sharing the same redshift , responsible for the ly@xmath0 emission in the filaments @xmath17 and @xmath16 . among the transitions in this wavelength range that could produce the band - like spectral character , possible candidates are the numerous feii and feii ] lines in the vicinity of 1104 , with several transitions from the ground - state and accompanying low - lying excited states .
the second panel from the top , shows an asymmetric , extended emission region ( 1157.3 - 1161.8 ) redward of an absorption trough .
the nature of this feature is unclear . in the third panel of fig.[juxtaspec ]
we find deep absorption troughs just blueward of extended emission near 1288.7 - 1294.0 and 1302.9 - 1304.1 .
the identification of the lower redshift system is uncertain , but the spectrally narrow , spatially very extended emission region 1302.9 - 1304.1 most likely corresponds to the oi 1302.17 , 1304.86 , 1306.02 triplet and the siii 1304.37 , 1309.28 doublet ( with possible contributions by other ions ) which would also account for the absorption trough immediately blueward .
siii may be expected to be the dominant contributor under a wider range of possible physical scenarios , but is perhaps somewhat too far to the red of the optimal position .
the spatial extent of the emission appears to exceed that of the other emission regions discussed earlier in both directions along the slit .
if the emission is due to oi , it could be enhanced by bowen fluorescence ( bowen 1947 ) , i.e. , pumping of oi by hi ly@xmath19 .
a situation like this could arise in the contact zone between partly neutral gas ( oi ) embedded in more highly ionized gas where recombinations produce ly series photons .
the fourth panel shows the nv 1238,1242 doublet region .
nv appears to be present in the form of an absorption trough , as commonly seen in high redshift galaxies , plus nv emission ( marked as `` 1241.3 '' ) , which is extended to the south . at a flux of @xmath31 erg @xmath5s@xmath6 , a comparison with other metals commonly seen in ionized gas ,
is difficult .
the flux at the civ 1548,1549 doublet position ( not shown ) , which is commonly stronger than nv , at @xmath32 erg @xmath5s@xmath6 is not significant . in the bottom panel , a wavelength region from the red side of the spectrum
contains a comb - shaped , multiple emission line pattern extending about 2.7 " ( 21 kpc proper ) to the north , unfortunately marred by two background residuals .
we tentatively identify the lines between 1658.8 - 1665.0 and 1666.6 - 1674.7 with oiii ] 1660 , 1666 ( which most likely would be collisionally excited ) and alii 1671 ( which could be pumped by continuum radiation ) .
the uncertain spatial extent of the ly@xmath0 emission beyond the slit prevents a measurement of the escape fraction of ly@xmath0 , but we can establish whether there is a sufficient source of ly@xmath0 that can explain the observed flux . if due to photoionization , the observed flux of ly@xmath0 photons requires an ionization rate of about @xmath33 erg @xmath5 s@xmath6hz@xmath6 . under the assumptions made in paper i and ii , this can be achieved by a star - forming galaxy with rest frame luminosity @xmath34 = @xmath35 erg s@xmath6hz@xmath6 or v band ( ab ) magnitude = 26.4 . with the brightest galaxy , number 1 , having a magnitude of 24.9 , there nominally are about four times as many ionizing photons available as required to explain the observed ly@xmath0 , not counting any of the other sources in the field . even if the slit losses approached a factor @xmath36 , as discussed above , stellar photoionization would remain a viable explanation .
based on the existing data we can not rule out the presence of an agn , but there currently is no positive evidence for such an object either .
the star formation rate associated with galaxy 1 , if estimated by the usual relation ( madau , pozetti & dickinson 1998 ) , is found to be @xmath37 for some of the more clearly circumscribed emission regions in the individual filaments the photon budget can be studied independently . for example , the tail of region 7 , with a v band magnitude 31.12 , coincides in projection with the ly@xmath0 spot @xmath16 in one of the southern filaments .
the observed ly@xmath0 flux of that spot requires v=30.9 , which is very close to the required value for a situation where virtually all ionizing radiation is being trapped and converted into ly@xmath0 photons , with all photons escaping .
the ly@xmath0 filamentary pattern is at least partly related to point sources that appear connected to the central emission peak by bridges of ly@xmath0 emitting gas . in principle , these bridges could consist of ionized gas , in which case the velocity position of the corresponding ly@xmath0 emission should be close to the systemic velocity .
we can not test the ionization state of the gas for the individual filaments , but for the main galaxy we have additional evidence from the absorption troughs present , which are all blue - shifted by on the order of @xmath38 kms@xmath6 with respect to the ly@xmath0 emission peak . in the case of ly@xmath0 emitters associated with lyman break galaxies ,
the redshifted velocity of the ly@xmath0 emission line with respect to the systemic redshift of the galaxy , is about 2 - 3 times the absolute value of the ( blue - shifted ) outflow velocity @xmath39 .
this puts the redshift of the ly@xmath0 emission relative to the systemic redshift at between 250 - 600 kms@xmath6 , which is well consistent with the range observed for lyman break galaxies ( e.g. , rakic et al 2011 ) , and would mean that at least the central ly@xmath0 peak is highly optically thick .
taken together with the presence of an absorption trough in the 1-d spectra ( fig.[onspec ] ) near 5100 , the evidence suggests that the ly@xmath0 emission is indeed the red peak of a double humped profile , with the blue peak strongly suppressed .
as for the nature of the galaxies in the field , most of the objects with a head - tail structure have their heads to the south of the tails ( 1,2,3,4 and 5 ) , making it less likely that the features have been produced mainly by tidal interactions .
rather , the evidence appears consistent with large scale ram - pressure stripping ( gunn & gott 1972 ; nulsen 1982 ) of gas , and recent star - formation in the down - stream ablated tails .
ram pressure stripping has been invoked to explain features seen in several astrophysical environments , including galaxy clusters ( e.g. , cayatte et al 1990 ; vollmer & huchtmeier 2003 ; chung et al 2007 ; cortese et al 2007 ; hester et al 2010 ; smith et al 2010 ; yagi et al 2010 ; fumagalli et al 2011 ; fossati et al 2012 ; arrigoni - battaia et al 2012 ) , the milky way halo ( e.g. , lin & faber 1983 ) , the local group ( e.g. , mcconnachie et al 2007 ) and in galaxy groups ( e.g. , rasmussen , ponman & mulchaey 2006 ; marcolini , brighenti , & dercole 2003 ) .
so far , there is relatively scant observational evidence for this process at high redshift , with the exception of the tadpole galaxies , a population of galaxies increasingly common with redshift ( e.g. , straughn et al 2006 ) , that may partly have been shaped by ram - pressure ( elmegreen & elmegreen 2010 ) .
groups of galaxies falling into clusters seem to produce @xmath40 morphologies similar to the galaxy contrails observed here ( e.g. , cortese et al 2006 ; owen et al 2006 ; sun et al 2007 ; yoshida et al 2012 ) . in the present case ,
a number of observational details closely resemble hydrodynamic features predicted from simulations of supersonic motions in galaxy clusters ( e.g. , roediger , brggen , & hoeft 2006 ; zavala et al 2012 ) : the main tadpole 1 shows a shape reminiscent of a bow shock , with a projected tangential angle of close to 45 degrees ( indicating , at face value , motion with a moderate mach number @xmath41 1.5 ) , and what looks like a turbulent wake consisting of a vortex street with a longitudinal extent of about 12 kpc and a maximum width of about 5 kpc ( fig .
[ bowshock ] ) . a faint
, linear structure vaguely resembling a `` helix '' extends almost perpendicular to the axis of symmetry ( i.e. , the head - tail direction ) to both sides of the galaxy , with a maximum traceable extent of @xmath42 kpc to the south and somewhat less to the north .
the nature of this feature is not clear
. some of the extended emission seen in fig .
[ juxtaspec ] appears to arise from that region , suggesting that it may be associated with both low and highly ionized gas .
the pattern with the `` helical '' appearance could perhaps arise through hydrodynamic instabilities . a broadening tail of vortices ( e.g. , roediger , br " uggen , & hoeft 2006 , their fig .
4 ) , may look similar , if the tadpole were viewed almost frontally , moving toward the observer , and the `` helical '' structure were on the far side of the tadpole .
similar , curly structures also occur in the wakes of _ interacting _ galaxies undergoing stripping ( e.g. , kapferer et al 2008 , their fig . 9 ) .
alternatively , the apparent spiral pattern may suggest a passing , rotating small galaxy with an asymmetric outflow of gas that traces out a helical pattern .
if the structure were an outflow from the tadpole galaxy 1 , e.g. , the helical jet of an agn , a large outflow velocity faster than the relative velocity between galaxy and ambient gas could explain why the `` helices '' are not being swept back at a sharper angle .
tadpole 2 does not show a discernible bow shock , and has a narrower wake .
the object 4 that we associated above with high equivalent width ly@xmath0 emission exhibits a linear , thin tail ( fig .
[ tads ] , [ tad4 ] ) , not unlike the expected outcome of bondi - hoyle accretion ( sakelliou 2000 ) .
the proximity of the galaxies to each other , the presence of a bow shock in one but not in the others , and the various degrees of turbulence in their tails may imply that the flow of galaxies is encountering inhomogeneous conditions . specifically , the spatial sequence from tadpole 4 to 2 to 1 ( fig . [ tads ] ) could be interpreted as indicating passage through a zone with a significant temperature gradient ( perhaps an accretion shock ) from a colder to a hotter gaseous medium : the tadpole 4 , which is most advanced in the direction of the flow , has a linear , apparently non - turbulent tail , suggesting that it may be experiencing the higher viscosity and thermal pressure of a hotter environment ( e.g. , roediger & brggen 2008 ) .
tadpole 2 is more turbulent , but the tail is still narrow , and it does not have a bow shock either .
tadpole 1 has a bow shock , presumably because it is on the colder side of the interface between hot and cold gas , and its velocity relative to the intergalactic medium exceeds the sound speed for the colder gas
. this condition would easily be satisfied when passing through the general filamentary igm with even moderate velocity , as the typical temperature at @xmath43 is only a few times @xmath44 ( rauch et al 1996 ) , corresponding to sound speeds of a few tens of kms@xmath6 .
the tail of tadpole 1 is flaring up and shows a turbulent vortex pattern , consistent with the lower viscosity expected in a colder medium .
a particular interesting consequence of ram - stripping is the formation of stars in the stripped gas , which has been the subject of recent observational ( e.g. , yoshida et al 2012 ; smith et al 2010 : hester et al 2010 ; fumagalli et al 2011 ; boissier et al 2012 ; owers et al 2012 ) and theoretical study ( e.g. , kapferer et al 2009 , tonnesen & bryan 2012 ; see also the ly@xmath0 emitting filament in paper ii , which may have a similar origin ) .
several strands of evidence suggest that the current situation is indeed an instance of star formation in galactic wakes : the tails of most of the objects have blue colors , indicative of very young stars .
the finding of extended metal line emission far out from the galaxies suggests the presence of a stripped , or re - created , interstellar medium that is being excited by the newly forming stars . the high ly@xmath0 equivalent width suggested by the tail of galaxy 4 may be another sign of hot , young stars , as discussed in paper ii .
the usual condition for ram - pressure stripping to take place is that the ram - pressure on the gas in a galaxy , experienced when passing through the intergalactic medium , needs to exceed the gravitational binding force per surface area , @xmath45 ( e.g. , mccarthy et al 2008 ) . here
@xmath46 is the ambient gas density , @xmath47 the relative velocity of galactic gas and ambient intergalactic medium , and @xmath48 , @xmath49 , @xmath50 and @xmath51 are the gravitational constant , the distance of a given gas volume element from the center of the galaxy , the total gravitating mass internal to that radius , and the galactic gas density , respectively . it has often been assumed that ram pressure stripping is most relevant for low redshift , massive clusters .
however , the hierarchical nature of structure formation , leading to more compact gravitational potential wells , higher gas densities , and higher interaction rates at @xmath43 ( when compared to the local universe ) suggests that one should expect miniature versions of the ram - pressure stripping seen in low redshift clusters to occur among satellites in individual high z galactic halos . with the higher density at high redshift favoring a higher pressure for a given velocity , the in - falling satellites themselves collapse from a denser background as well ,
so for the effect of ram pressure stripping one would have to look to lower mass , dwarf galaxies , perhaps aided by processes that may lower the binding energy of the gas further . the higher merger rate at high redshift may also work to increase the amount of ram - pressure stripped gas ( e.g. , domainko et al 2006 ; kapferer et al 2008 ) , as may stellar or galactic outflows , as long as they can offset a significant part of the gravitational binding energy . in addition , new cosmological hydro - simulation techniques suggest more efficient stripping ( e.g. , hess & springel 2012 ) and the presence of puffed up , high - angular momentum gas ( e.g. , keres et al 2012 ) , `` ready to go '' .
recently it has been argued ( benitez - llambay et al 2012 ) that , in particular , dwarf galaxies do not even require the encounter with fully formed massive halos but can lose gas to ram - pressure stripping in large - scale structure filaments at high redshift when entering terminal nodes like the ( future ) local group pancake . in this case , our spectrograph slit may have intersected a filament of the cosmic web , lit up by the star - formation in the ablated contrails of a swarm of coherently moving galaxies . to attain the high relative velocities required for stripping , the galaxies would have to move highly supersonically with respect to the gas they are plunging into . while we appear to be seeing one galaxy with a bow shock ,
it is not clear if the velocities of the other objects are high enough for this to work .
however , as argued above , the presence of an accretion shock in the terminal node with hot gas on one side may make this scenario consistent with the observations .
a variant of this picture may explain the stripping and the apparent gradient in the properties of the tadpoles as a group of galaxies being `` hosed down '' when obliquely passing an accreting stream of gas .
the existence of such extended structures at high redshift , the relatively large number density of galaxies with tidal or ram - pressure related features ( see also the disturbed halos described in paper i and ii ) , and the presence of multiple sites of star - formation in a common gaseous halo or large scale filament suggest that the stripping of gas from galaxies in interactions could be an important contributor to the metal enrichment of the intergalactic medium , analogous to the lower redshift process leading to the enrichment of the gas in galaxy clusters .
to explain the finding of metal enrichment in the igm at large distances from the nearest bright galaxy , galactic winds from lyman break galaxies have been invoked to drive metal - enriched gas far into intergalactic space ( e.g. , pettini et al 2001 ; steidel et a 2010 ) . among the persistent uncertainties with this scenario is that the actually observed ranges of galactic winds invariably fall short of accounting for the metals seen in qso absorption systems at large distances from such galaxies . however ,
if , as we have argued above , ram pressure stripping of in - falling dwarf galaxies and star formation in the stripped wake operate at high redshift , there may be less need to invoke long - range winds from the central galaxy of a halo . in this alternative picture , the ram - pressure that led to the ablation of gas and subsequent star formation may also act to dispel newly formed metal enriched gas from the tails , aided by stellar winds and supernova explosions that would find it much easier to escape from the weakly bound ( and dark - matter free ) star forming regions of galactic wakes . in any case , differential motion between the lost gas and the parent star forming galaxies will distribute the gas spatially over time , and the assumption that this process mostly occurs in in - falling dwarf satellites implies that the gas is automatically reaching distances from any brightest halo galaxy as large as commonly observed ( in metal absorption lines ; e.g. , chen 2012 ) .
tidal interactions between satellites may lead to a similar result , metals expelled into the gaseous halo of brighter galaxies , that came from the shredded interstellar medium of its satellites or from outflows in tidal dwarfs , and both processes may exist among the extended , asymmetric ly@xmath0 emitters in our study .
there may be differences between the metallicities of the gas ejected from tidal star - forming regions , and the gas lost by star forming regions in galactic contrails .
stars in the former arise from the relatively metal - rich ism of the parent galaxy , as would stars in gas ablated by ram - pressure or viscous stripping .
stars forming in turbulent wakes behind the galaxies may feed on the lower metallicity gas in the halo or intergalactic medium as well , which may contribute to the signatures of hot , young stars described in paper ii .
as argued earlier in paper ii , extragalactic star formation in the wakes of stripped galaxies and in tidal tails may facilitate the production and escape of ionizing photons and may have brought about the reionization of the universe at high redshift . star formation outside of the dense galactic hi cocoons would lead to lines of sight with reduced optical depth for ionizing photons .
the presence of young , massive stars in the wakes , with the weak gravitational binding force enabling easy removal of neutral gas by even moderate amounts of stellar winds or supernova outflows would all tend to enhance the escape of ionizing radiation .
recently , bergvall et al ( 2013 ) , examining selection effects in the search for local galaxies leaking ionizing radiation , have come to similar conclusions as to the likely conditions required , including the importance of stripping .
we have detected a ly@xmath0 emitting halo with several faint filaments stretching over tens of kpc .
the filaments correlate with star - forming regions in the form of mostly blue , faint galaxies , several of which have a distinct tadpole shape and blue , partly turbulent tails , with one object showing what appears to be a bow shock .
the goods - n acs f435w image reveals many such features criss - crossing an area several times bigger than the visible extent of the ly@xmath0 halo .
the emission of the central halo and of the filaments is broadly consistent with being powered by stellar photoionization .
we detect spatially extended emission lines from gas surrounding the main tadpole , including heii 1640 , nv 1240 and probably oiii ] , alii , and feii , suggesting an extended , extragalactic , interstellar medium with current star formation .
the tadpole shapes , partial alignment , and the considerable numbers of unusual broad band objects make it unlikely that the features observed are predominantly tidal in origin ( i.e. , caused by individual two - body encounters ) . instead
, the galaxies may have experienced stripping of gas when moving relative to the intergalactic or intra - halo medium , with stars forming downstream in the galactic contrails .
this process is observationally and theoretically well established in the local universe .
our observations have identified an occurrence of ram - pressure stripping at high redshift , possibly involving dwarf galaxies interacting with the gas in more massive , individual galactic halos .
the filamentary structure trailing behind a galaxy in the z=2.63 halo described in paper ii may be another example of this effect . in the present case ,
the properties of several tadpoles change along their general direction of motion , which may be consistent with these galaxies passing into a hotter gaseous environment , possibly the region behind an accretion shock .
such a stripping scenario may play out on a larger scale when differential motions of galaxies relative to the nodes in the gaseous cosmic web strip galaxies off their gas , as suggested by benitez - llambay et al 2012 . as in the case of local clusters
, the galactic contrails should be able to release metal - enriched gas , perhaps enhanced by local stellar feedback , more easily than normal galaxies . at the very least these objects
should provide a contribution to the intergalactic metal budget of galactic halos .
the loss of enriched gas from galactic contrails may suggest a solution to the long - standing puzzle of how the intergalactic medium at large distances from bright galaxies was polluted with metals .
star formation in galactic contrails would involve young stars , surrounded by lower hi gas columns than stars born in ordinary galaxies , and capable of clearing their environment of dense gas , suggesting a way in which galaxies can ionize the intergalactic medium .
the data were obtained as part of a long term collaboration with the late wal sargent , to whose memory we dedicate this paper .
we acknowledge helpful discussions with guillermo blanc , bob carswell , michele fumagalli and andy mcwilliam .
we thank the staff of the keck observatory for their help with the observations .
mr is grateful to the national science foundation for grant ast-1108815 .
gb has been supported by the kavli foundation , and mgh received support by the european research council under the european union s seventh framework programme ( fp/2007 - 2013 ) / erc grant agreement n. 320596 .
jrg acknowledges a millikan fellowship at caltech .
we acknowledge use of the atomic line list v2.05 , maintained by peter van hoof , and the use of the nist atomic spectra database ( ver . 5.0 ; kramida et al 2012 ) .
this research has further 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 of the vizier catalogue access tool .
arrigoni battaia , f. , gavazzi , g. , fumagalli , m. , boselli , a. , boissier , s. , cortese , l. , heinis , s. , ferrarese , l. , cot , p. , mihos , j. c. , and 8 coauthors , 2012 , a&a , 543 , 112 benitez - llambay , a. , navarro , j. f. , abadi , m. g. , gottlber , s. , yepes , g. , hoffman , y. , steinmetz , m. , 2013 , apj , 763 , 41 bergvall , n. , leitet , e. , zackrisson , e. , marquart , t. , 2013 , arxiv 1303.1576 boissier , s. , boselli , a. , duc , p .- a . , cortese , l. , van driel , w. , heinis , s. , voyer , e. , cucciati , o. , ferrarese , l. , cot , p. , and 3 coauthors , 2012 , a&a , 545 , 142 bowen , i. s. , 1947 , pasp , 59 , 196 cayatte , v. , van gorkom , j. h. , balkowski , c. , kotanyi , c. , 1990 , aj , 100 , 604 chen , h .- w .
, 2012 , mnras , 427 , 1238 chung , a. , van gorkom , j. h. , kenney , j. d. p. , vollmer , b. , 2007 , apj , 659 , 115 cooke , j. , berrier , j. c. , barton , e. j. , bullock , j. s. , wolfe , a. m. , 2010 , mnras , 403 , 1020 cooke , j. , omori , y. , ryan - weber , e. 2013 , arxiv1305.0562 cortese , l. , gavazzi , g. , boselli , a. , franzetti , p. , kennicutt , r. c. , oneil , k. , sakai , s. , 2006 , a&a , 453 , 847 cortese , l. , marcillac , d. , richard , j. , bravo - alfaro , h. , kneib , j .-
p . , rieke , g. , covone , g. , egami , e. , rigby , j. , czoske , o. , davies , j. , 2007 , mnras , 376 , 157 domainko , w. , mair , m. , kapferer , w. , van kampen , e. , kronberger , t. , schindler , s. , kimeswenger , s. , ruffert , m. , mangete , o. e.,2006 , a&a , 452 , 795 elmegreen , bruce g. ; elmegreen , debra m. , 2010 , apj , 722 , 1895 fossati , m. , gavazzi , g. , boselli , a. , fumagalli , m. , 2012 , a&a , 544 , 128 fumagalli , m. , gavazzi , g. , scaramella , r. , franzetti , p. , 2011 ,
a&a , 528 , 46 giavalisco , m. and the goods team , 2004 , apj , 600 , l93 gunn , j. e. , gott iii , j. r. , 1972 , apj , 176 , 1 hamana , t. ouchi , m. , shimasaku , k. , kayo , i. , suto , y. , 2004 , mnras , 347 , 813 hayashino , t. , matsuda , y. , tamura , h. , yamauchi , r. , yamada , t. , ajiki , m. , fujita , s. , murayama , t. , nagao , t. , ohta , k. , okamura , s. , ouchi , m. , shimasaku , k. , shioya , y. , taniguchi , y. , 2004,aj , 128 , 2073 hester , janice a. , seibert , m. , neill , j. d. , wyder , t. k. , gil de paz , a. , madore , b. f. , martin , d. c. , schiminovich , d. , rich , r. m. , 2010 , apj , 716 , 14 kapferer , w. , kronberger , t. , ferrari , c. , riser , t. , schindler , s. , 2008 , mnras , 389 , 1405 kapferer , w. , sluka , c. , schindler , s. , ferrari , c. , ziegler , b. , 2009 , a&a , 499 , 87 keres , d. , vogelsberger , m. sijacki , d. , springel , v. , hernquist , l. , 2012 , mnras , 425 , 2027 kramida , a. , ralchenko , yu . , reader , j. , and the nist asd team ( 2012 ) .
national institute of standards and technology , gaithersburg , md lin , d. n. c. ; faber , s. m. , 1983 , apj , 266 , 21 kovac , k. , somerville , r. s. , rhoads , j. e. , malhotra , s. , wang , j.x . , 2007,apj , 668 , 15 madau , p. , pozzetti , l. , dickinson , m. , 1998 , apj , 498 , 106 marcolini , a. , brighenti , f. , dercole , a. , 2003 , mnras , 345 , 1329 matsuda , y. , yamada , t. , hayashino , t. , yamauchi , r. , nakamura , y. , morimoto , n. , ouchi , m. , ono , y. , umemura , m. , mori , m. , 2012 , mnras , 425 , 878 mccarthy , j. k. , cohen , j. g. , butcher , b. , cromer , j. , croner , e. , douglas , w. r. , goeden , r. m. , grewal , t. , lu , b. , petrie , h. l. , weng , t. , weber , b. , koch , d. g. , rodgers , j. m. , 1998 spie , 3355 , 81 mccarthy , i. g. , frenk , c. s. , font , a. s. , lacey , c. g. , bower , r. g. , mitchell , n. l. , balogh , m. l. , theuns , t. , 2008 , mnras , 383 , 593 mcconnachie , a. w. , venn , k. a. , irwin , m. j. , young , l. m. , geehan , j. j. , 2007 , apj , 671 , 33 meiksin , a. , 2006 , mnras , 365 , 807 nulsen , p. e. j. , 1982 , mnras , 198 , 1007 oke , j. b. , cohen , j. g. , carr , m. , cromer , j. , dingizian , a. , harris , f. h. , 1995 , pasp , 107 , 375o owen , frazer n. , keel , w. c. , wang , q. d. , ledlow , m. j. , morrison , g. e. , 2006 , aj , 131 , 1974 owers , m. s. , couch , w. j. , nulsen , p. e. j. , randall , s. w. , 2012 , apj , 750 , 23o pettini , m. , shapley , a. e. , steidel , c. c. , cuby , j .- g . , dickinson , m. , moorwood , a. f. m. , adelberger , k. l. , giavalisco , m. , 2001 , apj , 554 , 98 prescott , m. k. m. , dey , a. , jannuzi , b. t. , 2009 , apj , 702 , 554 rakic , o. , schaye , j. , steidel , c. c. , rudie , g. c. , 2011 , mnras , 414 , 3265 rasmussen , j. , ponman , t. j. , mulchaey , j. s. , 2006 , mnras , 370 , 453 rauch , m. , sargent , w. l. w. , womble , d. s. , barlow , t. a. , 1996 , apj , 467 , 5 rauch , m. , haehnelt , m. , bunker , a. , becker , g. , marleau , f. , graham , j. , cristiani , s. , jarvis , m. , lacey , c. , morris , s. , and 3 coauthors , 2008 , apj , 681 , 856 rauch , m. becker , g. d. , haehnelt , m. g. , gauthier , j .-
r . , ravindranath , s. , sargent , w. l. w.,2011 , mnras , 418 , 1115 ( paper i ) rauch , m. becker , g. d. , haehnelt , m. g. , gauthier , j .-
r . , sargent , w. l. w.,2012 , mnras,429 , 429 ( paper ii ) rauch , m. becker , g. d. , haehnelt , m. g. , carswell , r.f . ,
gauthier , j .-
, 2013 , mnras , 431 , 68 ( paper iii ) roediger , e. , brggen , m. , hoeft , m. , 2006 , mnras , 371 , 609 roediger , e. , brggen , m. , 2008 , mnras , 388 , 89 scarlata , c. , colbert , j. , teplitz , h. i. , bridge , c. , francis , p. , palunas , p. , siana , b. , williger , g. m. , woodgate , b. , 2009 , apj , 706 , 1241 shapley , a. e. , steidel , c. c. , pettini , m. , adelberger , k. l. , 2003 , apj , 588 , 65 smith , r. j. , lucey , j. r. , hammer , d. , hornschemeier , a. e. , carter , d. , hudson , m. j. , marzke , r. o. , mouhcine , m. , eftekharzadeh , s. , james , p. , and 3 coauthors , 2010 , mnras , 408 , 1417 steidel et al 2004 , apj , 604 , 534 steidel , c. c. , erb , d. k. , shapley , a. e. , pettini , m. , reddy , n. , bogosavljevic , m. , rudie , g. c. , rakic , o. , 2010 , apj , 717 , 289 straughn , a. n. , cohen , s. h. , ryan , r. e. , hathi , n. p. , windhorst , r. a. , jansen , r. a. , 2006 , apj , 639 , 724 sun , m. , donahue , m. , voit , g. m. , 2007 , apj , 671 , 190 tonnesen , s. bryan , g. l. , 2012 , mnras , 422 , 1609 vollmer , b. , huchtmeier , w. , 2003 , a&a , 406 , 427 willis , a. j. , van der hucht , k. a. , conti , p. s. , garmany , d. , 1986 , a&as , 63 , 417 xue , y. q. , luo , b. , brandt , w. n. , bauer , f. e. , lehmer , b. d. , broos , p. s. , schneider , d. p. , alexander , d. m. , brusa , m. , comastri , a. , and 15 coauthors , 2011 , apjs , 195 , 10 yagi , m. , yoshida , m. , komiyama , y. , kashikawa , n. , furusawa , h. , okamura , s. , graham , a. w. , miller , n. a. , carter , d. , mobasher , b. , jogee , s. , 2010 , yang , y. , zabludoff , a. i. dav , r. , eisenstein , d. j. , pinto , p. a. , katz , n. , weinberg , d. h. , barton , e. j.,2006 , apj , 640 , 539 yoshida , m. , yagi , m. , komiyama , y. , furusawa , h. , kashikawa , n. , hattori , t. , okamura , s. , 2012 , apj , 749 , 43 zavala , j. , balogh , m. l. , afshordi , n. , ro , s. , 2012 , mnras , 426 , 3464 zheng , z. , cen , r. , trac , h. , miralda - escud , j. , 2011 , apj , 726 , 38 | a spectroscopically detected lyman @xmath0 emitting halo at redshift 3.216 in the goods - n field is found to reside at the convergence of several ly@xmath0 filaments .
hst images show that some of the filaments are inhabited by galaxies .
several of the galaxies in the field have pronounced head - tail structures , which are partly aligned with each other .
the blue colors of most tails suggest the presence of young stars , with the emission from at least one of the galaxies apparently dominated by high equivalent width ly@xmath0 .
faint , more diffuse , and similarly elongated , apparently stellar features , can be seen over an area with a linear extent of at least 90 kpc .
the region within several arcseconds of the brightest galaxy exhibits spatially extended emission by heii , nv and various lower ionization metal lines .
the gas - dynamical features present are strongly reminiscent of ram - pressure stripped galaxies , including evidence for recent star formation in the stripped contrails .
spatial gradients in the appearance of several galaxies may represent a stream of galaxies passing from a colder to a hotter intergalactic medium .
the stripping of gas from the in - falling galaxies , in conjunction with the occurrence of star formation and stellar feedback in the galactic contrails suggests a mechanism for the metal enrichment of the high redshift intergalactic medium that does not depend on long - range galactic winds , at the same time opening a path for the escape of ionizing radiation from galaxies .
= -0.8 in [ firstpage ] galaxies : halos galaxies : interactions galaxies : evolution galaxies : intergalactic medium (cosmology : ) diffuse radiation |
let @xmath7 be a compact and connected @xmath8-smooth manifold of dimension @xmath9 and let @xmath10 be a @xmath8-smooth riemannian metric on @xmath7 .
let @xmath11 be a real - valued @xmath8-smooth function on @xmath7 , and denote by @xmath12 the laplace - beltrami operator on @xmath7 .
we consider a hyperbolic inverse problem corresponding to the 2nd order elliptic operator @xmath13 in local coordinates @xmath10 is a positive - definite @xmath8-smooth matrix @xmath14 with the inverse @xmath15 and @xmath16 where @xmath17 .
hence our results cover the setting , where @xmath18 is an open domain with smooth boundary and @xmath19 is an elliptic operator of the form ( [ def : laplace_beltrami ] ) .
let @xmath20 be the sobolev space of @xmath21 times weakly differentiable functions on @xmath7 , and let @xmath22 be the @xmath23 closure of @xmath24 , the space of smooth compactly supported functions .
the operator @xmath25 is self - adjoint in @xmath26 , where @xmath27 is the riemannian volume measure . in local
coordinates @xmath28 .
denote by @xmath29 the solution of the initial boundary value problem @xmath30 for @xmath31 , and define the hyperbolic dirichlet - to - neumann operator @xmath32 where @xmath33 is the normal derivative on @xmath34 . in local
coordinates the exterior conormal @xmath35 is the covector @xmath36 with @xmath37 and @xmath38 denote by @xmath39 the restriction of the dirichlet - to - neumann operator @xmath40 where @xmath41 are open .
furthermore , denote @xmath42 .
it is well known , that the map @xmath43 determine the manifold @xmath44 up to an isometry @xcite .
this is also true for the restriction @xmath45 when @xmath46 is nonempty and @xmath47 is sufficiently large @xcite . in many applications observations of physical fields
can not be done on the same locations where the sources of the fields are .
for instance , in imaging in earth sciences , elastic or acoustic fields are often implemented using explosions @xcite .
in such a case observation devices need to be far away from the sources .
similarly , in electromagnetic imaging , it is technically difficult to use electrodes at the same time as sources and for making observations .
these are typical examples of cases where the observation devices and the sources of the fields are supported on disjoint sets . in this paper
we show , that for certain collections of pairs @xmath48 of open and disjoint subsets of @xmath34 , the operators @xmath39 determine the manifold @xmath44 up to an isometry .
[ thm : touching_main ] let @xmath49 be open sets such that @xmath50 and @xmath51
. then @xmath52 , given as a smooth manifold , and the operator @xmath53 determine the manifold @xmath54 up to an isometry .
[ thm : combining_data_infinite_time ] let @xmath0 , @xmath55 be open and nonempty .
then the smooth manifolds @xmath56 , @xmath57 , and the operators @xmath58 determine the manifold @xmath54 up to an isometry . for measurements on a finite time interval ,
we prove a theorem similar to theorem [ thm : combining_data_infinite_time ] under an additional controllability assumption : * for any @xmath59 there is a boundary value @xmath60 satisfying @xmath61 where @xmath62 is the solution of the equation ( [ def : wave_eq ] ) and @xmath63 .
let us comment the controllability assumption ( a ) when @xmath7 is embedded in @xmath64 .
property ( a ) follows from the geometric control condition of bardos , lebeau and rauch @xcite , which yields exact controllability of the wave equation .
property ( a ) follows also from existence of a strictly convex function @xmath65 on @xmath66 with respect to the riemannian metric @xmath10 .
suppose that @xmath67 is strictly convex and that @xmath68 is a lower bound for the hessian of @xmath65 in the riemannian metric @xmath10 , that is @xmath69 by @xcite , ( a ) holds if @xmath70 where @xmath71 and @xmath72 are the gradient and length with respect to the riemannian metric @xmath10 , @xmath73 is the euclidean unit outward normal to @xmath74 and @xmath75 is the euclidean inner product .
we refer to @xcite for examples of riemannian manifolds @xmath44 having a strictly convex function @xmath65 .
[ thm : combining_data_finite_time ] let @xmath76 be open and nonempty .
if the controllability assumption ( a ) holds for @xmath4 and @xmath77 , and @xmath78 then the riemannian manifolds @xmath79 , @xmath80 , and the operators @xmath81 determine the manifold @xmath44 up to an isometry .
the proofs of these theorems consist of showing that the data determine , up to a gauge transformation , the boundary spectral data of the operator @xmath82 on a part of the boundary .
the manifold is then determined up to an isometry , as can be seen using the boundary control method @xcite .
notice also , that the operator @xmath39 determines the operator @xmath83 by a time reversal argument .
are shown as arrows pointing from the support of sources to the support of observations .
using lemma [ lem : symmetrization_of_data ] we can change the direction of any arrow in the picture on left . hence also the measurements shown on right are covered by theorems [ thm : combining_data_infinite_time ] and [ thm : combining_data_finite_time ] .
, title="fig:",scaledwidth=40.0% ] are shown as arrows pointing from the support of sources to the support of observations .
using lemma [ lem : symmetrization_of_data ] we can change the direction of any arrow in the picture on left .
hence also the measurements shown on right are covered by theorems [ thm : combining_data_infinite_time ] and [ thm : combining_data_finite_time ] .
, title="fig:",scaledwidth=40.0% ] [ lem : symmetrization_of_data ]
let @xmath41 be open and nonempty .
let @xmath77 , and define the time reversal operator @xmath84 .
if @xmath85 and @xmath86 , then @xmath87 hence @xmath88 , @xmath89 , given as riemannian manifolds , and the operator @xmath39 determine the operator @xmath90 .
this result is relatively well known , see e.g. @xcite , but for the sake of completeness , we will give a proof in the appendix .
let us review previous results on the topic .
the inverse problem for isotropic wave equation on a compact manifold with measurements on the whole boundary was solved by belishev and kurylev @xcite .
this was based on the boundary control method originally developed in @xcite for wave equation on a bounded domain of @xmath64 .
the inverse problems for more general hyperbolic equations on a compact riemannian manifold with sources and observations on the same open subset @xmath46 of the boundary has been studied by katchalov and kurylev @xcite , see also @xcite .
similar problem has recently been studied for non - compact manifolds in @xcite .
inverse problems for elliptic equations with data on a part of the boundary have been studied intensively as they are the natural generalization of the caldern s inverse problem for the conductivity equation @xcite . when measurements are given on the whole boundary , the inverse problem for schrdinger equation on a bounded domain of @xmath64 , @xmath91 , and hence for isotropic conductivity equation , was solved by sylvester and uhlmann in @xcite . the corresponding two dimensional problem for isotropic conductivity equation
was solved first by nachman in @xcite for @xmath92 conductivities , and for @xmath93 conductivities , for which caldern s inverse problem was originally posed , by astala and pivrinta in @xcite .
recently , also the inverse problem for schrdinger equation on a bounded domain of dimension two with measurements on the whole boundary was solved by bukgheim in @xcite .
the corresponding problem on a compact riemannian surface was later solved in @xcite .
the inverse problem for schrdinger equation on a bounded domain of @xmath64 , @xmath91 , with observations on an open subset @xmath46 of the boundary was solved in @xcite .
the inverse problem for schrdinger equation on a bounded domain of @xmath64 , @xmath94 , with sources and observations on the same open subset @xmath46 of the boundary was solved by imanuvilov , uhlmann and yamamoto in @xcite .
the corresponding problem on a compact riemannian surface was later solved in @xcite . for related results with measurements on a part of the boundary ,
see @xcite . the inverse problem for
the laplace - beltrami operator @xmath12 on a compact riemannian manifold with sources and observations on the same open subset @xmath46 of the boundary has been studied on analytic riemannian manifolds of dimension @xmath95 in @xcite , and on riemannian surfaces in @xcite , see also @xcite .
the inverse problem for the laplace - beltrami operator in dimensions @xmath95 is open in general , even when measurements are given on the whole boundary . for positive results under certain geometrical conditions
see @xcite .
denote by @xmath96 the increasing sequence of distinct eigenvalues of the operator @xmath82 and let @xmath97 be an orthonormal basis of real - valued @xmath8-smooth eigenfunctions .
moreover , let @xmath98 be a partition of @xmath99 such that @xmath100 is a basis for the space of eigenfunctions corresponding @xmath101 .
let @xmath102 , and consider @xmath103 also as a function in @xmath104 by defining @xmath105 for @xmath106 .
there is a constant @xmath107 such that for @xmath108 , the fourier transform @xmath109 is an analytic function of @xmath110 when @xmath111 .
it is known ( see e.g. @xcite ) , that @xmath109 extends to a meromorphic function of @xmath112 , and that it may have poles only at points @xmath113 .
moreover , the residues at these points are @xmath114 where @xmath115 and @xmath116 is the riemannian surface measure .
if @xmath117 , @xmath118 are constants for @xmath119 , and the linear combination @xmath120 vanish on a nonempty open subset of @xmath34 , then @xmath121 for all @xmath119 by unique continuation , see e.g. @xcite . hence for an open nonempty set @xmath122 and @xmath117 , the functions @xmath123 are linearly independent .
let @xmath41 be open and nonempty . by linear independence and smoothness of @xmath124
, there are @xmath125 and @xmath126 such that @xmath127 moreover , for fixed @xmath112 , the map @xmath128 from @xmath129 to @xmath130 is surjective .
hence the operator @xmath53 determines the eigenvalues @xmath101 and the operators @xmath131
in this section we prove theorem [ thm : touching_main ] . [
lem : symmetry_for_derivatives_1d ] suppose that @xmath132 are such that @xmath133 for all @xmath134 .
then @xmath135 for all @xmath136 or @xmath137 for all positive @xmath138 .
assume that the claim is not valid .
then there exist @xmath117 and @xmath139 such that @xmath140 and @xmath141 .
let us next consider the smallest such integers @xmath142 and @xmath143 . by leibniz s formula @xmath144 where
@xmath145 and the terms with indices @xmath146 and @xmath147 have cancelled each other out . as @xmath143 is the smallest positive integer such that @xmath141
, we have @xmath148 in the sum @xmath149 , and so @xmath150 . as @xmath142 is the smallest integer such that @xmath140
, we have @xmath151 in the sum @xmath152 and @xmath153 in the sum @xmath154 , thus @xmath155 .
hence @xmath156 , which is a contradiction with the assumption that @xmath140 and @xmath141 .
this proves the claim . in the proof of the next lemma
we use the equation @xmath157 where @xmath158 , @xmath159 , @xmath160 and @xmath117 , [ lem : symmetry_for_derivatives ] suppose that @xmath161 are such that @xmath162 for all multi - indices @xmath163 .
then @xmath164 for all multi - indices @xmath165 or @xmath166 for all nonzero multi - indices @xmath167 .
let @xmath134 and @xmath160 . by ( [ eqn : directed_derivative ] ) @xmath168
hence @xmath169 for all @xmath134 and @xmath160 .
define the sets @xmath170 the sets @xmath171 and @xmath172 are closed by smoothness of @xmath173 and @xmath65 , respectively .
lemma [ lem : symmetry_for_derivatives_1d ] gives that @xmath174 .
if @xmath175 , then @xmath176 is open , nonempty and contained in @xmath172 . thus @xmath171 or @xmath172 contains an open nonempty subset .
suppose that @xmath177 is open and nonempty .
let @xmath117 , and define the polynomial @xmath178 by ( [ eqn : directed_derivative ] ) , @xmath179 , and @xmath180 vanish in @xmath181 .
using unique continuation for real analytic functions we see that @xmath182 in @xmath64 , and so the coefficients of @xmath180 vanish .
as @xmath142 can be chosen freely , @xmath164 for all multi - indices @xmath183 . similarly , if there exists an open and nonempty @xmath184 , then @xmath185 for all nonzero multi - indices @xmath183 .
[ rem : derivatives_in_local_coordinates ] let @xmath181 be a @xmath8-smooth manifold of dimension @xmath9 , @xmath186 and @xmath187 .
if @xmath164 for all multi - indices @xmath165 in some local coordinates taking @xmath180 to @xmath188 , then @xmath164 for all multi - indices @xmath165 in all local coordinates taking @xmath180 to @xmath188 .
[ lem : unique_continuation ] let @xmath189 be an eigenfunction of the operator a corresponding to an eigenvalue @xmath190 , and let @xmath191 . then in any local coordinates of @xmath34 taking @xmath192 to @xmath188 , there is a multi - index @xmath193 such that @xmath194 .
assume that the claim is not valid .
then @xmath195 for all @xmath193 in some local coordinates of @xmath34 taking @xmath192 to @xmath188 .
consider boundary normal coordinates of @xmath66 taking @xmath192 to @xmath188 .
we may suppose that the coordinates map a small neigborhood @xmath196 of @xmath192 onto @xmath197 , where @xmath198 is a ball of radius @xmath199 centered at the origin .
then these coordinates take a boundary point @xmath200 to a point @xmath201 , where @xmath202 .
the special property of the boundary normal coordinates is , that a point @xmath203 has coordinates @xmath204 , where @xmath205 and @xmath206 are the coordinates of the unique boundary point @xmath207 such that @xmath208 . moreover , in the coordinates @xmath209 the equation @xmath210 has the form @xmath211 for some @xmath212 , @xmath213 , see e.g. @xcite .
let us show , that @xmath214 .
let @xmath215 and @xmath193 . by applying the operator @xmath216 on the both sides of ( [ eq : eigeneq_in_boundary_normal_coordinates ] ) ,
we get @xmath217 the right hand side of this equation is a linear combination of functions @xmath218 at the point @xmath219 . the equations @xmath220 hold for all @xmath193 by the boundary condition @xmath221 .
furthermore , we have by remark [ rem : derivatives_in_local_coordinates ] , that @xmath222 for all @xmath193 . using induction
we see that @xmath223 for all @xmath193 and @xmath215 .
define the odd and even reflection operators @xmath224 where @xmath225 , @xmath226 and @xmath227 . also , define @xmath228 , @xmath229 , @xmath230 and @xmath231 for @xmath232 .
denote @xmath233 .
as @xmath234 , we see that @xmath235 and @xmath236 moreover , @xmath237 is lipschitz continuous in @xmath181 , @xmath238 for @xmath213 , and @xmath239 where the both sides are considered as functions in @xmath240 .
hence for some constant @xmath107 @xmath241 since @xmath242 vanishes up to arbitrary degree in origin , taylor s formula gives for any @xmath243 a constant @xmath244 such that @xmath245 hence for any @xmath243 , there is a constant @xmath246 such that @xmath247 by hrmander s strong unique continuation result @xcite this yields , that @xmath248 in @xmath181 . in particular ,
@xmath214 around some point @xmath249 .
as @xmath7 is connected , unique continuation gives that @xmath214 in @xmath7 .
this is a contradiction with the assumption that @xmath189 is an eigenfunction , and the claim is proved .
[ lem : spans ] let @xmath250 and @xmath251 be hilbert spaces , and let @xmath252 and @xmath253 be linearly independent .
suppose that @xmath254 is a dense subspace , and define @xmath255 then @xmath256 determines the unique bounded extension @xmath257 , and its adjoint @xmath258 .
hence , @xmath256 determines the spaces @xmath259 [ thm : boudary_data_to_spectral_data ] let @xmath49 be open , @xmath50 , and @xmath51 . then the smooth manifold @xmath52 and the collection @xmath260 determine boundary spectral data up to a constant gauge transformation on @xmath1 .
that is , one can find a collection @xmath261 where for an unknown constant @xmath107 not depending on @xmath142 or @xmath143 , @xmath262 is an orthonormal basis of eigenfunctions in @xmath263 corresponding the eigenvalue @xmath101 .
choose a smooth positive measure @xmath264 on @xmath52 .
then there is a positive function @xmath265 such that @xmath266 . as @xmath267 , the functions @xmath268 are linearly independent for all @xmath117
. for all @xmath117 , denote @xmath269 , and define @xmath270 note , that for the smallest eigenvalue @xmath271 , the space of eigenfunctions is one dimensional ( see e.g. ( * ? ? ?
* thm 6.5.2 ) ) , and so @xmath272 consider a positive function @xmath273 and real - valued functions @xmath274
, @xmath138 , such that the following three conditions hold : * if @xmath275 , then in local coordinates of @xmath52 taking @xmath276 to @xmath188 @xmath277 for all multi - indices @xmath278 . * @xmath279 and @xmath280 for all @xmath117 . * @xmath281 for all @xmath282 , where @xmath283 such functions @xmath284 and @xmath285 exist . for example , @xmath286 and @xmath287 satisfy the conditions .
next we show the following two statements .
* we can verify using the data ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) , whether any given functions @xmath273 and @xmath274 , @xmath138 , satisfy the conditions ( a1 ) , ( a2 ) and ( a3 ) .
* there is an orthonormal basis of eigenfunctions @xmath288 of operator @xmath82 and a constant @xmath107 , not depending on @xmath143 , such that @xmath289 if ( i ) holds , then the data ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) determine the nonempty collection @xmath290 and if ( ii ) holds , then any element from this collection determines a collection of type ( [ eq : boundary_spectral_data ] ) .
so the claim of the theorem is proved after proving ( i ) and ( ii ) .
let us show the claim ( i ) .
clearly , the condition ( a3 ) can be verified using the data ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) . as @xmath291 by varying @xmath125 , we see that the map @xmath292 determines the function @xmath293 .
let @xmath275 . given @xmath294 and @xmath284 , it is possible to compute @xmath295 in any local coordinates @xmath52 taking @xmath276 to @xmath188 . by smoothness of @xmath296 , @xmath284 and @xmath297 ,
these functions are known also at @xmath298 .
hence the condition ( a1 ) can be verified using the data ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) .
taking @xmath299 , @xmath300 and @xmath301 in the formulation of remark [ lem : spans ] , we see that the map @xmath292 determines the spaces @xmath302 and @xmath303 .
hence the condition ( a2 ) can be verified using the data ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) , and the claim ( i ) is proved .
let us show the claim ( ii ) .
let @xmath304 .
lemma [ lem : unique_continuation ] gives that , in local coordinates of @xmath34 taking @xmath276 to @xmath188 , there is a multi - index @xmath193 such that @xmath305 .
hence the condition ( a1 ) and lemma [ lem : symmetry_for_derivatives ] imply , that @xmath306 for all nonzero multi - indices @xmath307 .
fix @xmath117 and , to simplify the notation , drop the subindices @xmath142 from now on . by the condition ( a2 )
@xmath308 for some constant matrices @xmath309 and @xmath310 .
fix @xmath311 , let @xmath312 and define the function @xmath313 we have seen that , in local coordinates of @xmath34 taking @xmath276 to @xmath188 , the equation @xmath306 holds for all nonzero multi - indices @xmath307 .
hence for any multi - index @xmath314 @xmath315 by lemma [ lem : unique_continuation ] , the coefficients @xmath316 vanish , and so @xmath317 .
moreover @xmath318 for all @xmath125 and @xmath319 . on the other hand
, the condition ( a3 ) gives @xmath320 for all @xmath125 and @xmath319 .
denote @xmath321 . by density of @xmath322 in @xmath323 , @xmath324 , @xmath325 for all @xmath326 and @xmath327 .
let @xmath328 be biorthogonal with @xmath329 in @xmath330 , and let@xmath331 be biorthogonal with @xmath332 in @xmath333 , that is , @xmath334 then @xmath335 denote @xmath336 .
we have shown , that @xmath337 .
hence the matrix @xmath338 is orthogonal . to conclude
, we observe that @xmath339 where @xmath340 is an orthonormal basis of eigenfunctions corresponding the eigenvalue @xmath101 .
as discussed in the previous section , the operator @xmath53 determines the collection ( [ eqn : data_of_thm_boudary_data_to_spectral_data ] ) .
so by previous theorem , if @xmath51 , then the operator @xmath53 determines the collection ( [ eq : boundary_spectral_data ] ) .
the collection ( [ eq : boundary_spectral_data ] ) determines the manifold up to an isometry by ( * ? ? ?
* chapter 4.4 ) .
this proves theorem [ thm : touching_main ] .
in this section we prove theorems [ thm : combining_data_infinite_time ] and [ thm : combining_data_finite_time ] .
denote @xmath341 .
it is enough to show , that the collection @xmath342 determines a collection of type ( [ eq : boundary_spectral_data ] ) .
denote @xmath343 , and choose a smooth positive measure @xmath264 on @xmath52 .
there is a positive function @xmath265 such that @xmath344 .
define for all @xmath117 @xmath345 choose a positive function @xmath273 and real - valued functions @xmath274
, @xmath138 , such that the following two conditions hold : * for all @xmath117 @xmath346 * @xmath347 for all @xmath117 and @xmath348 , where @xmath349 again , such functions @xmath284 and @xmath285 exist , as @xmath286 and @xmath287 satisfy the conditions ( b1 ) and ( b2 ) .
it is enough to show the following two statements : * we can verify using the data ( [ eqn : data_of_thm_combining_data_infinite_time ] ) , whether any given functions @xmath273 and @xmath274 , @xmath138 , satisfy the conditions ( b1 ) and ( b2 ) .
* there is an orthonormal basis of eigenfunctions @xmath288 of operator @xmath82 and a constant @xmath107 , not depending on @xmath143 , such that @xmath289 analogously with the proof of theorem [ thm : boudary_data_to_spectral_data ] , the maps @xmath350 , @xmath351 and @xmath352 determine the spaces @xmath302 , @xmath303 , @xmath353 and @xmath354 . hence the claim ( i ) is proved .
let us show the claim ( ii ) .
fix @xmath117 and , to simplify the notation , drop the subindices @xmath142 from now on .
the condition ( b1 ) gives , that for all @xmath312 @xmath355 for some constant matrices @xmath356 for the smallest eigenvalue @xmath271 , the space of eigenfunctions is one dimensional , and so @xmath357 for all @xmath358 . by lemma [ lem : unique_continuation ] ,
the set @xmath359 does not contain a nonempty open set of @xmath34 .
hence @xmath360 .
moreover , @xmath361 in @xmath362 and by continuity also in the whole set @xmath1 .
denote by @xmath363 the constant @xmath364 .
as @xmath365 are linearly independent , @xmath366 for all @xmath117 .
moreover , we may use the condition ( b2 ) as we used the corresponding condition in the proof of theorem [ thm : boudary_data_to_spectral_data ] , and get @xmath367 hence @xmath368 and @xmath369 . to conclude , we observe that @xmath339 where @xmath340 is an orthonormal basis of eigenfunctions corresponding the eigenvalue @xmath101 .
we prove theorem [ thm : combining_data_finite_time ] by reduction to theorem [ thm : combining_data_infinite_time ] using a time continuation argument similar to @xcite .
suppose that @xmath41 are open and nonempty .
denote @xmath370 if @xmath371 , then the smooth manifolds @xmath0 , @xmath1 , the operator @xmath39 and the inner products @xmath372 determine the operator @xmath373 for @xmath374 .
[ lem : continuation_of_data ] denote by @xmath375 the time delay operator @xmath376 as the coefficients of the wave equation ( [ def : wave_eq ] ) are time - independent , @xmath377 and @xmath378 .
let @xmath379 and choose @xmath380 and @xmath381 such that @xmath382 .
let @xmath383 .
as @xmath384 , the operator @xmath39 determine the function @xmath385 in @xmath1 .
therefore , it is enough to show that the given data determine also @xmath386 consider a sequence @xmath387 satisfying the following two conditions . * @xmath388 in @xmath23 , * @xmath389 in @xmath263 .
such a sequence exists , since @xmath390 , and thus we see exactly as in ( * ? ? ?
4.28 ) that the set @xmath391 is dense in @xmath392 . let us prove , that ( c1 ) is equivalent with * for all @xmath393 @xmath394 where @xmath395 .
as @xmath396 and @xmath397 , we have that @xmath398 .
hence @xmath399 where @xmath400 is the exterior derivative on @xmath7 .
if ( c1 ) holds , then @xmath401 for large enough @xmath393 there is a constant @xmath402 such that @xmath403 . hence if ( c1 ) holds , then @xmath404 as @xmath405 , and ( c1 ) holds
therefore ( c1 ) and ( c1 ) are equivalent .
next we observe that @xmath406 hence the condition ( c1 ) can be verified for given functions @xmath407 and @xmath65 using the inner products ( [ eqn : continuation_inner_products ] ) .
similarly , @xmath408 and condition ( c2 ) can be verified for given functions @xmath407 and @xmath65 using the inner products ( [ eqn : continuation_inner_products ] ) .
as @xmath409 on @xmath410 , conditions ( c1 ) and ( c2 ) together with the continuous dependency of the solution of the wave equation on the initial data , see e.g. ( * ? ?
* thm . 2.30 ) ,
give @xmath411 we have seen that , the inner products ( [ eqn : continuation_inner_products ] ) determine for any @xmath380 the nonempty set @xmath412 and that any sequence in this set satisfies @xmath413 in @xmath414 .
as @xmath415 , the inner products ( [ eqn : continuation_inner_products ] ) and the operator @xmath39 determine @xmath416 pointwise for @xmath126 and @xmath383 . next we prove the last of the three main theorems formulated in the introduction . by lemma [ lem : symmetrization_of_data ]
the operators @xmath417 determine the operators @xmath418 we use the time delay operator @xmath375 defined in the proof of lemma [ lem : continuation_of_data ] .
define @xmath419 : = \int_0^t \int_{{\partial}m } \left ( { \partial}_\nu v^f \bar{v^h } - v^f \bar{{\partial}_\nu v^h } \right ) ds_g dt , \quad f , h \in c_0^\infty({\partial}m \times ( 0 , \infty)),\ ] ] and let @xmath420 .
we recall blagovestchenskii identity ( * ? ? ?
4.16 ) , originating from @xcite , @xmath421 ds,\ ] ] where @xmath422 . by this identity
, the operators ( [ eqn : ops_of_finite_time_thm ] ) determine the inner products @xmath423 for @xmath424 , @xmath425 .
next we will show that the operators ( [ eqn : ops_of_finite_time_thm ] ) determine the inner products ( [ eqn : inner_products_t0 ] ) also for @xmath426 , @xmath57 .
let @xmath427 and consider a sequence @xmath428 satisfying the following two conditions .
* for all @xmath429 @xmath430 * the sequence @xmath431 is bounded in @xmath432 . by assumption ( a )
, there is @xmath433 such that @xmath434 .
thus , there is a sequence @xmath428 such that @xmath435 by ( * ? ? ?
* lem . 2.42 ) , @xmath436 in @xmath263 as @xmath405 .
hence a sequence satisfying the conditions ( d1 ) and ( d2 ) exists .
let us next show that @xmath437 as @xmath438 for all @xmath439 and @xmath440 , ( * ? ? ?
* thm . 3.10 )
gives that the set @xmath441 is dense in @xmath263 .
let @xmath199 and choose @xmath442 such that @xmath443 by ( * ? ? ?
2.42 ) there is @xmath107 , and by the condition ( d1 ) there is @xmath444 such that for @xmath445 @xmath446 by the condition ( d2 ) @xmath447 hence the equation ( [ eqn : inner_product_is_determined ] ) is valid . by ( * ?
3.10 ) the functions @xmath448 , @xmath427 , are dense in @xmath263 .
hence @xmath449 where @xmath450 , @xmath451 , and the supremum is taken over all @xmath427 such that @xmath452 .
the condition ( d1 ) can be verified for any @xmath173 and @xmath453 using the inner products ( [ eqn : inner_products_t0 ] ) for @xmath454 , @xmath455 .
therefore , these inner products determine for any @xmath427 the nonempty set @xmath456 by equation ( [ eqn : inner_product_is_determined ] ) any sequence in this set together with inner products ( [ eqn : inner_products_t0 ] ) for @xmath457 and @xmath458 determine @xmath459 . as @xmath427
can be chosen freely , the inner products ( [ eqn : inner_products_t0 ] ) for @xmath454 , @xmath455 and for @xmath457 , @xmath458 together with polarization identity determine the inner products ( [ eqn : inner_products_t0 ] ) for @xmath460 . the equation ( [ eqn : inner_product_is_determined2 ] ) , polarization identity and the inner products ( [ eqn : inner_products_t0 ] ) for @xmath461 , @xmath462 determine the inner products ( [ eqn : inner_products_t0 ] ) for @xmath463 .
therefore , the operators ( [ eqn : ops_of_finite_time_thm ] ) determine the inner products @xmath464 choose @xmath465 , where @xmath466 is defined as in lemma [ lem : continuation_of_data ] . by lemma [ lem : continuation_of_data ] the operators ( [ eqn : ops_of_finite_time_thm ] ) and the inner products ( [ eqn : all_inner products ] ) determine the operators @xmath467 repeating this construction
, we see that the operators @xmath468 are determined for all @xmath243 .
the claim follows from theorem [ thm : combining_data_infinite_time ] .
next we prove the lemma [ lem : symmetrization_of_data ] stating that the operator @xmath39 determines the operator @xmath83 .
define @xmath469 , where @xmath470 is the solution of the equation ( [ def : wave_eq ] ) with the boundary data @xmath471 .
then @xmath472 satisfies the equation @xmath473 integration by parts gives @xmath474_{t=0}^{t = t } = 0,\end{aligned}\ ] ] since @xmath475 and @xmath476 .
* acknowledgements . *
the authors were partly supported by finnish centre of excellence in inverse problems research , academy of finland coe 213476 .
lo was partly supported also by finnish graduate school in computational sciences .
99 k. astala , l. pivrinta , caldern s inverse conductivity problem in the plane . ann . of math .
( 2 ) 163 ( 2006 ) , no .
1 , 265299 .
m. belishev , an approach to multidimensional inverse problems for the wave equation .
( russian ) dokl .
nauk sssr 297 ( 1987 ) , no .
3 , 524527 ; translation in soviet math . dokl .
36 ( 1988 ) , no .
3 , 481484 . a. p. caldern , on an inverse boundary value problem .
seminar on numerical analysis and its applications to continuum physics ( rio de janeiro , 1980 ) , pp .
6573 , soc .
brasil . mat .
, rio de janeiro , 1980 .
a. katchalov , y. kurylev , m. lassas , energy measurements and equivalence of boundary data for inverse problems on non - compact manifolds .
i m a volumes in mathematics and applications ( springer verlag ) `` geometric methods in inverse problems and pde control '' ed . c. croke , i. lasiecka , g. uhlmann , m. vogelius , 2004 , pp .
183 - 214 .
kurylev , m. lassas , hyperbolic inverse problem with data on a part of the boundary .
differential equations and mathematical physics ( 1999 ) , 259272 , ams / ip stud . adv .
, 16 , amer .
soc . , 2000 .
i. lasiecka , r. triggiani , p. yao , exact controllability for second - order hyperbolic equations with variable coefficient - principal part and first - order terms .
proceedings of the second world congress of nonlinear analysts , part 1 ( athens , 1996 ) .
nonlinear anal .
30 ( 1997 ) , no .
1 , 111122 . | we consider an inverse problem for a hyperbolic partial differential equation on a compact riemannian manifold . assuming that @xmath0 and @xmath1 are two disjoint open subsets of the boundary of the manifold we define the restricted dirichlet - to - neumann operator @xmath2 .
this operator corresponds the boundary measurements when we have smooth sources supported on @xmath0 and the fields produced by these sources are observed on @xmath1 .
we show that when @xmath0 and @xmath1 are disjoint but their closures intersect at least at one point , then the restricted dirichlet - to - neumann operator @xmath2 determines the riemannian manifold and the metric on it up to an isometry . in the euclidian space , the result yields that an anisotropic wave speed inside a compact body is determined , up to a natural coordinate transformations , by measurements on the boundary of the body even when wave sources are kept away from receivers .
moreover , we show that if we have three arbitrary non - empty open subsets @xmath3 , and @xmath4 of the boundary , then the restricted dirichlet - to - neumann operators @xmath5 for @xmath6 determine the riemannian manifold to an isometry .
similar result is proven also for the finite - time boundary measurements when the hyperbolic equation satisfies an exact controllability condition .
_ keywords _ : inverse problems , wave equation , partial data . |
in recent years , the investigation of the spin structure of the nucleon by using deep inelastic lepton - hadron scattering ( dis ) experiments has been an exciting and controversial field in hadron physics .
the experiments performed by emc , smc at cern @xcite , and e142 , e143 at slac @xcite provide direct information on the matrix elements of spin - dependent operators in the nucleon .
the spin - dependent dis cross section is determined by the antisymmetric part of the hadronic tensor @xmath10g_2(x , q^2)\ } \eqno ( 1)\ ] ] where @xmath11 and @xmath12 are the four vectors of the nucleon and virtual photon momentums . @xmath13 and @xmath14 are two spin - dependent structure functions for the bjorken variable @xmath15 with @xmath16 is the transfered four momentum squared and @xmath17 is the covariant spin vector of the nucleon . according to operator product expansion ( ope ) analysis , to order @xmath18 , the lowest two moments of @xmath13 and @xmath14 can be written ( qcd radiative corrections are not included ) @xcite @xmath19 \eqno(2a)\ ] ] @xmath20 @xmath21 @xmath22 where @xmath23 , @xmath24 and @xmath25 depend on the nucleon forward matrix elements of twist-2 , twist-3 and twist-4 operators respectively .
for example @xmath26 @xmath27 ( @xmath28 ) are axial charges defined by the above axial - vectorial matrix elements .
the singlet axial charge is proportional to the total quark helicity @xmath29 in the nucleon . using the semileptonic weak decay data ,
which are related to nonsinglet axial charges , the 1988 emc result seemed to indicate that @xmath30 is surprisingly small and led to so - called `` spin crisis '' .
since then , an intensive study of @xmath31 has been conducted .
the efforts both from experimental and theoretical works led to a deeper understanding of internal spin structure of the nucleon , although many questions remain .
the most recent reviews on this subject can be found in @xcite . neglecting the terms of order @xmath18 in eqs.(2a - d ) , the longitudinal polarized structure function @xmath13 receives only twist-2 contributions . on the other hand ,
the structure function @xmath14 contains not only twist-2 but also twist-3 contributions corresponding to the matirx element @xmath32 .
the twist-3 contributions coming from spin - dependent quark gluon correlations do not vanish even in the large @xmath33 limit @xcite . to separate the twist-2 and twist-3 contributions of @xmath14 ,
one can write ( wandzura - wilczek@xcite ) @xmath34 where the twist-2 piece ( see , however , discussion in section iii ) is @xmath35 and the twist-3 piece is @xmath36 it is easy to check that , to order o(@xmath37 , we have @xmath38 @xmath39 hence the twist-2 and twist-3 contributions of @xmath14 are separated .
measuring @xmath13 and @xmath14 as functions of @xmath40 and @xmath33 , one can obtain @xmath41 , @xmath42 and their third and even higher moments . up to order @xmath18 , @xmath14 uniquely measures the twist-3 contributions without involving any model - dependent analysis .
hence @xmath43 provides more detailed information about nucleon structure than does @xmath31 .
very recently , a preliminary experimental result of @xmath43 has been published @xcite and more precise data obtained by e143 experiment @xcite at slac should be published soon . since @xmath43 was discussed only briefly in a previous paper @xcite , it is appropriate to give a detailed analysis in the modified center - of - mass ( cm ) bag model , using newly obtained data on @xmath43 . in section
ii , the cm bag model is briefly reviewed and the results for @xmath1 , @xmath2 and @xmath3 , and their moments @xmath23 , @xmath24 are presented . discussion and comparison with most recent data and other model results
are given in section iii .
a brief summary is given in section iv .
as mentioned in a previous paper @xcite , the bag model does not contain gluon fields explicitly , but the boundary of the bag - confined quarks simulates the binding effect coming from quark - gluon and gluon - gluon interactions ( the gluon contribution to the proton spin in the bag model has been discussed by jaffe @xcite recently ) .
hence , the structure function @xmath43 calculated in the bag model does include higher twist effects .
we have reported the cm bag model results @xcite for the unpolarized and polarized structure functions @xmath44 , @xmath45 , and @xmath31 , and briefly for @xmath43 . in this paper
, we will focus our attention on @xmath43 . for reader s convenience , we give a brief review for cm bag model calculation of the structure functions .
one needs to calculate the hadron tensor @xmath46|p , s>\ ] ] and separate the antisymmetric part to obtain @xmath47 which can be expressed in general as @xmath48 in the bjorken limit , where @xmath49 depends on the model of the nucleon and the approximations used in the calculation .
the basic assumptions and approximations of the cm bag model are : ( i ) the nucleon electromagnetic current @xmath50 ( or @xmath51nn vertex , see ( 2.7 ) in @xcite ) , can be approximately expressed by incoherent sum of single quark electromagnetic currents .
it implies that the virtual photon interacts with only one quark at a time and the other two quarks are spectators ; this is an impulse approximation .
the current includes not only the contribution of the struck quark but also those of the spectator quarks . since the current satisfies translational invariance , four momentum is conserved .
( ii ) the nucleon consists of three valence quarks in their @xmath52-wave state ; higher excited states and higher fock states which include gluons and sea quark pairs in addition to three valence quarks are neglected .
( iii ) su(3)@xmath53su(2)@xmath54 wave functions for the proton and neutron are used .
symmetry - breaking effect is described in terms of a parameter @xmath55 , which simulates the smaller spatial size for the scalar @xmath56 quark pair than that for the vector @xmath57 and @xmath58 quark pairs in the nucleon @xcite .
( iv ) the effect of quark confinement due to nonperturbative quark - gluon and gluon - gluon interactions is described in terms of bound - state quark spatial wave functions , for instance the quark bag wave function in the cavity approximation in the mit bag model or gaussian - type quark wave function used in some other models .
all necessary formulae for the cm bag model calculation can be found in @xcite .
a formal and general discussion on the theoretical basis of the cm bag model has been given in @xcite .
we note that in addition to the cm bag model calculation , the transverse spin structure functions @xmath43 has also been computed in the original mit bag model by jaffe and ji @xcite , and other modified versions of the mit bag model by schreiber , signal and thomas ( sst bag model @xcite ) , and by stratmann ( mod model @xcite ) .
experimentally , @xmath13 and @xmath14 are measured by combining two different cases of deep inelastic scattering of polarized leptons on polarized nucleons : ( i ) the beam and target spin orientations are parallel , and ( ii ) the beam and target spin orientations are perpendicular . experimentally , @xmath59 can be calculated from various models of the nucleon . in this case
it is convenient to choose suitable projection operators to extract @xmath31 and @xmath43 from model results of @xmath49 .
one of possible projections is to extract @xmath60 and @xmath61 by choosing the nucleon spin parallel ( ` l ' ) or perpendicular ( ` t ' ) to the virtual photon momentum as we did in @xcite . it should be noted that it is not necessary to choose the same projection as those used in the experimental analysis .
several parameters have been used in the cm bag model calculation .
they are : ( a ) ` bag ' radius r=5 gev@xmath62 , ( b ) su(3 ) symmetry breaking parameter @xmath63 and ( c ) maximum momentum of quarks inside the nucleon @xmath64 gev / c .
@xmath65 and @xmath66 were determined from the fit of the @xmath67 radius of the neutron and proton , and the ratio @xmath68 @xcite .
the model with these two parameters gives a fairly good result for the electromagnetic form factors of the nucleon and the magnetic moments of octet baryons .
in particular , the neutron charge form factor is well reproduced . for the dis
parton distributions , the third parameter @xmath69
gev / c has been introduced .
it constrains the unpolarized valence quark distributions to satisfy the sum rules @xmath70 as mentioned in @xcite , to compare the model results with data , the qcd evolution technique has to be used to evolve the parton distributions from the renormalization scale @xmath71 to higher @xmath33 where the experiments are performed .
we choose @xmath71=0.81 ( gev / c)@xmath72 and qcd scale parameter @xmath73=0.3 ( gev / c ) .
the cm bag model with these parameters gives a good description for both unpolarized and polarized structure functions at @xmath74 , where the valence quark contributions dominate . for small-@xmath40 region , @xmath75 ,
the sea quark contributions are necessary . using some qcd inspired phenomenological sea distributions ( see ( 4.19 ) in @xcite )
, we found that the first moment of @xmath76 is consistent with the experimental value and the first moment of @xmath77 satisfies the bjorken sum rule .
for the higher moments ( @xmath78 ) , the sea contributions coming mainly from the small @xmath40 region are highly suppressed by the factor @xmath79 and are thus less important .
hence we neglect possible sea contributions for the higher moments of @xmath43 , @xmath80 and @xmath81 . for qcd evolution of the structure function @xmath31 and twist-2 piece @xmath80 from @xmath71 to @xmath82 , the ordinary altarelli - parisi equations @xcite can be used . for twist-3 part of @xmath43 ,
however , it has been shown @xcite that due to mixing of twist-3 quark operators and quark - gluon operators with same twist and quantum numbers , the number of independent operators contributing to @xmath81 increases with @xmath83 , where @xmath83 refers to the @xmath83-th moment .
it implies that one can not write down an altarelli - parisi type evolution equation for @xmath43 .
this feature has been confirmed by several later calculations in @xcite and most recently in @xcite .
hence there is no simple evolution equation for @xmath43 in the general case .
one has to look for some approximate solutions .
two approximate evolution approaches under the limits @xmath84 or power @xmath85 were suggested by ali , braun and hiller @xcite .
we use the approach in the large @xmath86 limit rather than the approximation in the large @xmath83 limit , which only provides the asymptotic behavior of @xmath14 in the region @xmath87 .
we also note that as far as the third moments @xmath88 are concerned , the @xmath33 evolution is straightforward , i.e. a single power behavior of @xmath89 ( for instance see @xcite ) . in fig.1@xmath905 , we present the results of @xmath76 , @xmath91 , @xmath92 , @xmath93 ( which is determined by @xmath76 ) , and @xmath94 respectively . for the deuteron target , @xmath95 , @xmath96 and @xmath97
are shown in fig .
all theoretical curves are calculated in the cm bag model at @xmath71=0.81 ( gev / c)@xmath72 and evolved to @xmath33=5.0 ( gev / c)@xmath72 except for fig . 4 and fig . 6 , where @xmath33=4.0 ( gev / c)@xmath72 and @xmath33=3.0 ( gev / c)@xmath72 respectively . comparisons of our results for the third moments of @xmath31 and @xmath43 with recent data and other model predictions are listed in tables i , ii and iii .
the data for @xmath31 are taken from @xcite and those for @xmath43 are taken from @xcite .
* 1*. for the leading term of first moments of @xmath0 , we obtain @xmath98 which can be compared with smc data @xcite at @xmath99=5 ( gev / c)@xmath72 : @xmath100 , @xmath101 , @xmath102 and e143 data @xcite at @xmath99=3 ( gev / c)@xmath72 : @xmath103 , @xmath104 , @xmath105 .
* 2*. for the leading term of third moments of @xmath0 , we get @xmath106 while preliminary data at @xmath99=5 ( gev / c)@xmath72 @xcite show : @xmath107 and @xmath108 .
comparisons with other models are listed in tables i and iii .
we note that no @xmath109 data are available yet .
* 3*. since @xmath76 is always positive in the range @xmath110 , hence @xmath111 must be positive . from eq.(2c ) , the leading term of @xmath111 , i.e. @xmath112 must also be positive as shown in eq.(9 ) .
however , the @xmath113 as function of @xmath40 is mostly negative except for large @xmath40 region , where @xmath113 is positive but very small .
hence its third moment , or @xmath114 , is negative . for the deuteron ,
since the positive contribution from the proton is larger than the negative contribution from the neutron in the @xmath115 , hence our model predicts a positive @xmath116 as shown in eq.(9 ) and table iii . * 4*. as mentioned in section i , in the ope approach including twist-2 and twist-3 operators in the presence of qcd corrections , the burkhardt - cottingham @xcite sum rule @xmath4 is known to hold .
the cm bag model predicts @xmath117 comparing to the numerical values for @xmath118 in eq.(8 ) , they are numerically consistent with zero . hence in our model ,
the burkhardt - cottingham sum rule is satisfied within numerical errors .
most recent data @xcite give @xmath119 which are consistent with zero . it should be noted that in the mit bag model with su(6 ) symmetry , @xmath120 is identically zero by itself . however , the cm bag model with su(6 ) symmetry breaking effects ( @xmath121 ) predicts a nonzero @xmath120 . *
5*. since @xmath1 is not identically zero in the model , the burkhardt - cottingham sum rule implies that @xmath1 must change its sign at some @xmath122 , where @xmath123 , _ i.e. _ @xmath1 must have at least one non - trivial zero .
the question is whether the @xmath40-behavior of @xmath43 satisfies @xmath124 or just opposite @xmath125 for case 1 , one has @xmath126 , while for case 2 , @xmath127 .
the cm bag model predictions for @xmath128 are all negative ( see table i , ii and iii or fig.3 and fig.7 ) .
this implies the @xmath40-behavior of @xmath1 fits case 2 . the preliminary data @xcite seem to favor our predictions ( see , for instance , fig.2 for the proton ) .
* 6*. for both the proton and deuteron , we now have @xmath129 and @xmath130 . in the cm bag model , the moment @xmath7 ( @xmath131/3 $ ] ) is about one order of magnitude smaller than @xmath8 ( @xmath132 ) .
it implies @xmath133 ( recall @xmath134 ) , or @xmath135 .
hence the twist-3 matrix element @xmath136 approximately equals to the twist-2 matrix element @xmath137 and the sign of @xmath136 should also be positive for the proton and deuteron targets .
this agrees with data @xcite but disagree with the negative sign predicted by the qcd sum rules @xcite and quenched lattice qcd@xcite . for the neutron ,
since @xmath138 is negative and much larger than @xmath139 in magnitude , hence @xmath140 is negative .
this negative sign is consistent with the results given by other approaches @xcite . in magnitude
, our result agrees with that given by the quenched lattice qcd , but much less than those given by the qcd sum rules . *
7*. our results show @xmath135 , i.e. the twist-3 contribution is almost the same magnitude as twist-2 contribution . from ( 6a ) and ( 6b )
, it implies that the @xmath141 and @xmath3 have opposite sign and approximately same magnitude .
they almost cancel each other and lead to a very small @xmath1 ( see also fig .
12a , b in @xcite ) . for the same reason the original wandzura - wilczek relation @xmath142 is not a good approximation and the higher twist contributions may not be neglected . as pointed by cortes , pire and ralston @xcite that the original wandzura - wilczek relation was derived by using the dirac equation for free and massless quarks .
including quark mass effect and gluon dependent term , an extended wandzura - wilczek relation was given ( similar formula without quark mass term has been given by jaffe @xcite ) : @xmath143 where the mass dependent term is another twist-2 piece which is related to ` transversity ' @xmath144 .
the last term is the ` true ' twist-3 piece arising from quark - gluon correlation . as emphasized in @xcite that when the quark mass term is included the twist-3 term can not be isolated in a model independent way with a measurement of @xmath31 and @xmath43 .
however , neglecting the strange quark contribution , the quark mass term ( @xmath145 ) should be negligible for up and down quarks .
hence the separation of twist-2 and twist-3 pieces in eq.(4 ) seems to be a reasonable approximation . *
8*. according to the ope analysis , neglecting the @xmath18 corrections , the general formulae for the moments of the structure functions are @xmath146 @xmath147 \qquad ( n=2,4,6 , ... ) \eqno ( 11b)\ ] ] for @xmath148 , they reduced to ( 2c ) and ( 2d ) . from ( 11a ) and ( 11b ) , one obtains @xmath149dx = { 1\over 2}d^{(n)}(q^2)\qquad ( n=2,4,6 , ... ) \eqno ( 12)\ ] ] hence one has for @xmath148 @xmath150dx \eqno ( 13)\ ] ] the cm bag model prediction for the function @xmath151 $ ] for the proton is shown in fig . 9 and that for the deuteron is shown in fig .
one can see that in both cases , the model predictions for @xmath136 are nonzero and positive .
this seems to be consistent with recent data @xcite . *
9*. if one assumes that ( 12 ) holds also for @xmath152 one obtains @xmath153dx={1\over 2}d^{(1)}(q^2 ) \eqno ( 14a)\ ] ] to lowest order in @xmath154 , it was shown @xcite by using the field theoretical parton model that @xmath155 vanishes in the chiral limit , one has the efremov - leader - teryaev ( elt ) sum rule @xmath153dx=0 \eqno ( 14b)\ ] ] however , our results of @xmath31 and @xmath43 do not satisfy this sum rule . to demonstrate this , we plot @xmath156 as functions of @xmath40 in fig.11 and fig . 12 and
compare them with recent proton and deuteron data @xcite respectively .
one can see that our model predictions are consistent with the slac e143 data .
however , the elt sum rule seems not to be supported by the data ( at least for the proton data ) .
the study of transverse spin structure function @xmath14 has both theoretical and experimental interest . in the most naive parton model ,
the quark is asymptotic free and has no transverse momentum , @xmath1 is identically zero .
however , quarks inside the nucleon are not free , the binding effect , which arising from quark - gluon interactions , causes a nonzero transverse momentum for quarks and leads to @xmath157 .
measurements of @xmath43 or @xmath158 allow us to get more information about binding effects , which are mainly formulated as ` higher twist effects ' .
on the other hand , since @xmath33 is large in the deep inelastic scattering , quark binding effects should not be significant and @xmath43 should be small , especially compared to @xmath31 .
this seems to agree with most recent exprimental result @xcite .
our model calculation is consistent with this conclusion .
it should be noted , however , that the theoretical predictions given by different models or approaches seem not to fully agree with each other because of different approximations . in addition , as mentioned in section iii that another twist-2 piece which is related to the quark mass and ` transversity ' @xmath144 has been neglected in our discussion . on the experimental side ,
the errors of data are still quite large .
we hope that several new experiments @xcite to be performed in the next few years will provide more precise data and tell us more about the quark spin ( including longitudinal and transverse ) distributions in the nucleon .
i wish to thank x. ji for useful comments and suggestions , and thank p. k. kabir , j. s. mccarthy and h. j. weber for helpful discussions .
i also thank o. rondon - aramayo for providing the updated e143 data and suggestions .
this work has been supported by the u.s .
department of energy and the institute of nuclear and particle physics , university of virginia , usa . | the spin - dependent structure functions @xmath0 , @xmath1 , @xmath2 and @xmath3 and their moments are studied in the cm bag model .
the results show that ( i ) @xmath4 , i.e. the burkhardt - cottingham sum rule holds , hence @xmath1 must have at least one non - trivial zero besides @xmath5 and @xmath6 .
( ii ) @xmath7 is negative for the proton , neutron and deuteron .
( iii ) @xmath7 is about one order of magnitude smaller than @xmath8 , hence the twist-3 matrix element is approximately equal to the twist-2 matrix element .
the results are compared with most recent data and predictions from the mit bag model , lattice qcd and qcd sum rules .
# 1@xmath9#1 |
during the last decade after discovery of top quark @xcite , top quark has been inclusively studied . by now
, the mass of the top quark has been measured to be [email protected] @xmath2 @xcite which is the most precisely measured quark mass and @xmath3 pair production cross section has been measured as less than 10 % of uncertainties @xcite . however , many of another top quark property have not yet been well explored due to the limited statistics . in the ongoing data taking at fermilab s tevatron proton - antiproton collider with collider detector at fermilab ( cdf ) , an increasing of integrated luminosity can make us to measure the property and also discover the unexpected phenomena from top sector .
we describes a few of the cdf s progress of top quark property measurements in the following .
the predominant production of top quark in the tevatron is the @xmath3 pair production .
the standard model ( sm ) predicts the @xmath3 production processes to be @xmath4 annihilation ( @xmath5 ) and @xmath6 fusion ( @xmath7 ) , occurring at the tevatron with relative fractions of @xmath885% and @xmath815% , respectively @xcite .
a measurement of this fraction tests the sm predictions and our understanding of gluon parton distribution functions in the proton .
we measure this quantity both lepton jets and dilepton final state . in the lepton jet channel
, we have two different measurement based on different discriminant .
one method builds distriminant using number of low - momentum track to take advantage of the higher probability for a gluon than for a quark to radiate a low - momentum gluon @xcite .
the other method builds discriminant with artificial neural net using eight variables those are sensitive to the production mechanism @xcite . both two measurement use @xmath9 of cdf data , and is combined by using the feldman - cousin prescription @xcite .
figure [ ggfrac_fc ] shows feldman - cousin bands for the combination of two analysis with 68% and 95% confidence level ( c.l . ) for the fraction of @xmath6 to produce @xmath3 .
we measure the fraction to be @xmath10 , and we find the 95% c.l .
limit to be @xmath11 @xcite .
this is consistent with sm .
feldman - cousins bands for the combination of the analysis for @xmath6 fraction in the lepton jet channel with statistical and systematic uncertainties for 68% c.l . and 95% c.l
.. ,width=302 ] we use angle between two lepton to build discriminant for extraction of @xmath6 fraction in the dilpeton channel @xcite .
we build feldman - cousin bands and fit data using @xmath12 of @xmath13 collision as you can see in fig .
[ ggfrac_fc_dil ] .
we measure the @xmath6 fraction in the dilepton channel to be @xmath14 which is consistent with sm .
feldman - cousins bands for @xmath6 fraction in the dilepton channel with statistical and systematic uncertainties for 68% c.l . and
95% c.l .. ,width=302 ] since new production mechanism for top quark pairs can make the shape of @xmath3 invariant mass as resonances or general shape distortions , the generic method to search the such contribution is to compare the shape of the observed differential @xmath3 cross section @xmath15 with sm expectation .
the mass of the top - antitop system is reconstructed for each event by combining the four vectors of the four leading jets , lepton , and missing transverse energy .
the unfolding technique implemented to correct the reconstructed distribution as for direct comparison with theoretical differential cross section . in the update with @xmath16 data
, we have in - situ jet energy scale ( jes ) measurement using di - jet mass of @xmath17 boson decay , which have been used in the top quark mass measurement @xcite , that we can significantly reduce the jes systematics . as one can see in fig .
[ dsigma ] , we do not find any significant difference with sm expectation .
we check the consistency using the anderson - darling ( ad ) statistics @xcite .
we calculate a p - value of 0.28 using ad statistics which have a good agreement with the sm @xcite
. unfolded differential cross section of @xmath3 invariant mass using cdf data is compared with sm prediction.,width=302 ] we search for resonant top - antitop pair production and subsequent decay in the all - hardronic channel with @xmath18 of data @xcite .
we use the probability from per - event matrix element calculation as discriminant to reduce and control the large background from qcd - multijet as an input of neural net with other kinematic variables .
we reconstruct invariant mass of @xmath3 with matrix element technique and have consistent result with sm prediction as one can see in fig .
we then set the 95% c.l .
limits on z production as 805 gev in case of a leptophobic topcolor resonance candidate .
invariant mass of @xmath3 from data and sm expectation.,width=302 ] due to the large mass of the top quark , the super - symmetric partner of top quark ( stop ) can be lighter than the top quark even to be a lightest squark . for a light stop and r - parity conservation of super - symmetry particles , stop quark dominantly decay to b - quark and chargino , and chargino decay to w boson and neutralino . while neutralino can not be detected , the decay of pair produced stop quark have same final state with @xmath3 decay .
we search the pair production of stop in the dilepton final state with @xmath16 of data @xcite .
we reconstruct the stop mass in the underconstraint system to extract the stop components in the @xmath3 dilepton decay events . figure [ stop ]
shows our data is compatible with sm prediction without stop , and then we set the 95% c.l .
upper limit of stop in certain condition of susy parameter space @xcite .
reconstructed stop mass comparing data to monte carlo.,width=302 ] one of basic quantities of top quark is the electric charge , which is expected to have a value of 2/3@xmath19 in the sm .
however , one of exotic model have decay of top to a @xmath20 instead of @xmath21 having a charge of 4/3@xmath19 @xcite .
we have a measurement using @xmath22 of data in the lepton jet channel .
the measurement identifies the charge of the two w bosons and two b - quarks in each data event , and then determine which w bosons and b - quarks decayed from the same parent top quark .
the charge of the top is then obtained by multiplying the charge of the w with the charge of the jet associated with a b - quark .
figure [ charge ] shows the measured distribution of pairs of charge product which is compatible with sm like top charge .
we then exclude exotic model - like top at 87% c.l .
@xcite .
product of the w charge and the associated jet charge for data and mc ( sm signal mc distribution).,width=302 ] top quark width have been measured with @xmath23 of data in the lepton jet channel .
main idea is to use top mass reconstruction and templates for different top width and to fit it to data .
we use the top quark mass considered known as @xmath24 and templates are produced for range of different top width .
we extract top width from reconstructed top mass distribution compared to signal with different top width and background using unbinned likelihood fit . we then have measurement consistent with sm as one can see in fig .
[ width ] , so we set a limit on top width using feldman - counsins @xcite prescription to be top width(@xmath25)@xmath2613.1 gev of 95% c.l .
upper limit @xcite .
likelihood fit to the data with 1-tag ( up ) and 2-tag ( down ) categories.,title="fig:",width=226 ] likelihood fit to the data with 1-tag ( up ) and 2-tag ( down ) categories.,title="fig:",width=226 ] the sm predicts that the top quark decays almost entirely to a w - boson and a bottom quark , and that the wtb vertex is a v - a charged weak current interaction .
a consequence of this is that the top quark is expected to decay 70.4% of the time of longitudinal and the rest to left handed polarized w - bosons @xcite .
any new particles involved in the same decay topologies and non - standard coupling could create a different mixture of polarized w - bosons .
therefore , a measurement of this fraction is a test of the v - a nature of the wtb vertex . in the cdf
, there are several measurements using different technique in the lepton jet channel with @xmath27 of data .
one method builds template using @xmath28 @xcite , where the @xmath29 is the angle between lepton and b - quark in the w rest frame , this is sensitive to w helicity .
the other method use matrix element technique @xcite which we calculate a likelihood for each event then product per event likelihood to build total likelihood .
figure [ whel ] shows w - helicity measurements at cdf .
all of measurements are consistent with sm prediction .
the summary of w - helicity measurement at cdf.,width=226 ] several exotic physics models , such as susy and two higgs doublet , predict flavor changing neutral currents ( fcnc ) in top decay . in the standard model ,
this decay mode is highly suppressed .
therefore , any signal from fcnc decay chain indicate an evidence of new physics . a search for fcnc decays has been performed at cdf with @xmath27 .
this analysis utilizes a template fit to a mass @xmath30 variable constructed from kinematic constraints present in fcnc top quark decays .
a simultaneous fit is performed to the data using two signal and one control region as one can see in fig .
the control region constrains uncertainties in the shape and normalization of the templates .
as one can see in this plot , our data is well explained without fcnc components and then , we set 95% c.l . upper limit on the branching fraction of ( @xmath31 ) @xmath26 3.7% @xcite .
mass @xmath30 distribution for signal and control regions.,width=302 ] charged higgs @xmath32 bosons are predicted in supersymmetric and gut extensions of the sm .
if a charged higgs boson in sufficiently light , it can be produced in top quark decays . in the presence of a charged higgs boson ,
the @xmath33 decay would compete with the sm top quark decay , thereby altering the expected number of events in different final states of @xmath3 . in the certain final state , which is low @xmath34 ,
the dominant decay of charged higgs is @xmath35 .
we has searched for the decays in the lepton + jets events with @xmath36 by fully reconstructing @xmath3 decay and exploiting the difference between the dijet mass spectra in @xmath37 and @xmath38 decays @xcite . the invariant dijet mass spectrum in data is shown in fig .
[ chiggs ] ( up ) which no significant deviation from the sm is observed .
therefore we set the limits on branching fraction of @xmath39 as one can see in fig .
[ chiggs ] ( down ) . in this plot ,
we extend our search to generic charged boson search which possibly have smaller mass than w boson .
distribution of di - jet mass in @xmath3 decay ( up ) and the upper limit on the branching fraction of @xmath40 at 95% c.l . as a function of charged higgs mass ( down).,title="fig:",width=226 ] distribution of di - jet mass in @xmath3 decay ( up ) and
the upper limit on the branching fraction of @xmath40 at 95% c.l . as a function
of charged higgs mass ( down).,title="fig:",width=226 ]
number of top quark properties not only standard model top signature but also exotic model signature have been searched and measured .
however many measurements are still limited by the statistical uncertainty .
although we do not find evidence conflicting with sm top quark , we expect to have interesting measurement with more data in near future . | we present the top property measurements in the cdf .
most of measurements utilize close to the integrated luminosity of 3 fb@xmath0 . |
the nervous system is an extremely complex system @xcite comprising nerve cells ( or neurons ) and gial cells . by electrical and chemical synapses of different polarity neurons
form a great variety large - scale networks . therefore
, modeling of brain s key functional properties is associated with study of collective activity of complex neurobiological networks .
dynamical modeling approach @xcite is effective tool for the analysis of this kind of networks .
first of all this approach takes building dynamical models of single neurons . on the one hand
, such models should describe large quantity of various dynamical modes of neural activity ( excitable , oscillatory , spiking , bursting , etc . ) .
this complexity is associated with the large number of voltage - gated ion channels of neurons .
it takes employment of complex nonlinear dynamical systems given by differential equations .
the canonical representative of this type of models is hodgkin - huxley system .
it describes dynamics of the transport through membrane of neuron in detail . on the other hand , to model neural network consisting of the large number of interconnected units it is necessary to create simplified models for single neuron to avoid problems that are induced by high dimension and nonlinearity .
for example , one which is commonly used in simulations is integrate - and - fire model @xcite .
it represents one - dimensional nonlinear equation with some threshold rule .
that is , if the variable of the model crosses a critical value , then it is reset to new value and the neuron is said to have fired . to solve the contradiction between the requirements of complexity and simplicity of neuron models phenomenological models were introduced .
they describe basic properties of neuron dynamics , but these models do not take into account the large number of voltage - gated ion channels of neurons . as a rule they involve generalized variable which mimic the dynamic of some number of ionic currents at the same time .
the examples of this type models are fitzhugh - nagumo @xcite , hindmarsh - rose @xcite , morris - lecar @xcite .
they have the form of differential equations systems .
however , there is another class of phenomenological models of the neural activity .
these are discrete - time models in form of point maps . in the last decade
this kind of neural models has attracted much attention@xcite .
for example using a map - based approach rulkov et .
@xcite have studied dynamics of one- and -two dimensional large - scale cortical networks .
it has been found that such map - based models produce spatiotemoral regimes similar to those exhibited by hodgkin - huxley -like models .
neuron oscillatory activity can take a variety of forms @xcite .
one of the most interesting oscillatory regimes is spiking - bursting oscillations regime , which is commonly observed in a wide variety of neurons such as hippocampal pyramidal neurons , thalamic neurons , pyloric dilator neurons etc .
a burst is a series of three or more action potential usually riding on a depolarizing wave .
it is believed that the bursting oscillations play crucial role in informational transmitting and processing in neurons , facilitate secretion of hormones and drive a muscle contraction .
this oscillation can be regular or chaotic depending on the concentration of neuromodulators , currents and other control parameters .
another interesting oscillatory regime is an oscillation of membrane potential below the excitation threshold , so - called subthreshold oscillation .
for example , these oscillations with close to 10 hz frequency are observed in olivo - cerebellar system providing highly coordinated signals concerned with the temporal organization of movement execution @xcite ( see more discussion in the conclusion ) .
the best known spiking - bursting activity model is the hindmarsh - rose system @xcite .
it is three - dimensional ode - based system involving two nonlinear functions .
spiking - bursting dynamics of map - based models has recently been investigated by cazelles et.al @xcite , rulkov @xcite , shilnikov and rulkov @xcite , tanaka @xcite .
a piecewise linear two - dimensional map with a fast - slow dynamics was introduced in @xcite .
it was shown that depending on the connection ( diffusively or reciprocally synoptically ) , the model demonstrates several modes of cooperative dynamics , among them phase synchronization .
two dimensional map is used for modeling of spiking - bursting neural behavior of neuron @xcite .
this map contains one fast and one slow variable .
the map is piecewise nonlinear and has two lines of discontinuity on the phase plane .
modification of this model is presented in @xcite .
the further advancement of rulkov model is presented in @xcite .
a quadratic function has been introduced in the model .
using these modifications authors obtained the dynamical regimes of subthreshold oscillation , corresponding to the periodical oscillation of neuron s transmembrane potential below the excitation threshold .
in @xcite the dynamics of two coupled piece - wise linear one - dimensional monostable maps is investigated .
the single map is associated with poincar section of the fitzhugh - nagumo neuron model .
it is found that a diffusive coupling leads to the appearance of chaotic attractor .
the attractor exists in an invariant region of phase space bounded by the manifolds of the saddle fixed point and the saddle periodic point .
the oscillations from the chaotic attractor have a spike - burst shape with anti - phase synchronized spiking .
a map - based neuron model involving quasi - periodic oscillation for generating the bursting activity has been suggested in @xcite .
izhikevich and hoppenstead have classified @xcite map - based one- and two - dimensional models of bursting activity using bifurcation theory .
our goal here is to introduce a new map - based model for replication of many basic modes of neuron activity
. the greater part of our paper deals with regimes that mimic chaotic spiking - bursting activity of one real biological neuron .
we construct a discontinuous two - dimensional map based on well - known one - dimensional lorenz - type map @xcite and a discrete version of the fitzhugh - nagumo model@xcite .
this is the system of two ode : @xmath0 where @xmath1 is the membrane potential of the neuron and @xmath2 is the recovery variable describing ionic currents , @xmath3 is a cubic function of @xmath1 and @xmath4 is constant stimulus .
this model takes into account the excitability and regular oscillations of neuron , but not spiking - bursting behavior .
we shall introduce a discontinuity in the discrete version for this purpose .
we find conditions under which this two - dimensional map has an invariant region on the phase plane , containing chaotic attractor .
in addition we show that depending on the values of the parameters , our model can produce phasic spiking and subthreshold oscillations also .
the paper is organized as follows . in sec .
[ sec : model ] we describe the map - based model . then in sec .
[ sec : oneddyn ] we study one - dimensional dynamics in the case when the recovery variable @xmath2 is fixed . in sec .
[ sec : twoddyn ] we analyze the relaxation two - dimensional dynamics of the model . then in sec .
[ sec : invr ] we find an invariant region bounding the chaotic attractor in the phase plane of the model . in sec .
[ sec : othermodes ] we observe other modes of neural activity which could be simulated by using this model .
let @xmath5 be a map @xmath6 of the form @xmath7 where the @xmath1-variable describes the evolution of the membrane potential of the neuron , the @xmath2 - variable describes the dynamics of the outward ionic currents ( the so - called recovery variable ) .
the functions @xmath8 and @xmath9 are of the form @xmath10 where @xmath11 the parameter @xmath12 @xmath13 defines the time scale of recovery variable , the parameter @xmath4 is a constant external stimulus , the parameters @xmath14 @xmath15 and @xmath16 @xmath17 ) control the threshold property of the bursting oscillations . here
we have chosen this linear piece - wise approximation of @xmath8 in order to obtain a simple hyperbolic map for chaotic spiking - bursting activity .
however , any cubic function can be also used .
the map @xmath18 is discontinuous map and @xmath19 is the discontinuity line of @xmath18 .
we consider only those trajectories ( orbits ) which do not fall within a discontinuity set @xmath20 , where @xmath21 is the union of points of discontinuity of @xmath18 and its derivative @xmath22 . besides
, we assume that @xmath23 , then @xmath24 for any @xmath25 and the map @xmath18 is one to one . we restrict consideration of the dynamics of the map @xmath18 to the following parameter region @xmath26 note that under such conditions we have @xmath27 .
this condition is very important for forming chaotic behavior of the map @xmath18 as we shall see bellow . for convenience ,
we rewrite the map @xmath18 in the following form @xmath28 where @xmath29 , @xmath30 are the maps @xmath31
let us start with the dynamics of the map @xmath18 when the parameter @xmath32 . in this case
the map @xmath18 is reduced to a one dimensional map : @xmath33 where @xmath34 is a constant and it plays the role of a new parameter .
the map ( [ eq : onedmap ] ) can be rewritten as @xmath35 where @xmath36 .
let us fix the parameters @xmath37 , @xmath16 , @xmath38 , @xmath39 and consider the dynamics of the map ( [ eq : onedmapfl ] ) in the parameter plane @xmath40 .
we restrict our study of the map @xmath18 to the following parameter region @xmath41 @xmath42 where @xmath43 .
these conditions allow to obtain interesting properties of the map ( [ eq : stpmp ] ) .
let us find the conditions on the parameter values for which the map @xmath18 acts like a lorenz - type map @xcite .
for that we require that ( see fig . [
pic : kl ] ) @xmath44 it follows from ( [ eq : cndlrnc0 ] ) the following condition on the parameter @xmath14 : @xmath45 where @xmath46 the inequalities ( [ eq : cnd1dbeta ] ) and ( [ eq : cndlrnc1 ] ) define on the @xmath47 plane the region @xmath48 ( see fig.[pic : bd_dbeta ] ) .
let us take the parameters @xmath16 and @xmath14 inside the @xmath48 region , and let us consider the @xmath49 plane . in this plane the inequalities ( [ eq : cnd1dbeta ] ) ,
( [ eq : cnd1dy0 ] ) and ( [ eq : cndlrnc1 ] ) are satisfied simultaneously in region @xmath50 . in this plane the boundary of @xmath50
consists of the three lines ( fig .
[ pic : bd_betay0 ] ) @xmath51 consider the dynamics of the map @xmath52 for @xmath53 .
this region is separated on four subregions by the bifurcation lines @xmath54 corresponding to different dynamics of the map @xmath52 . the line @xmath55 coincides with appearance of an unstable fixed point @xmath56 through crossing of the discontinuity point @xmath19 .
line @xmath57 corresponds to the fold ( tangent ) bifurcation of the fixed point @xmath58 ( see fig . [
pic : kl](a , d ) ) .
line @xmath59 corresponds to the condition @xmath60 note that for @xmath61 there exists a bifurcation corresponding to appearance of homoclinic orbit @xcite to the unstable fixed point .
the dynamics of the map @xmath52 corresponding to subregions @xmath62 is shown in fig.[pic : kl ] . if @xmath63 the trajectories of the map @xmath52 tend to stable fixed point @xmath64 for any initial conditions different from an unstable fixed point ( fig.[pic : kl ] ( a ) , ( b ) ) .
if @xmath65 the map @xmath18 has invariant interval @xmath66 , where @xmath67 for parameters @xmath68 the map @xmath52 exhibits bistable property , that is there exists two attractors , one is a stable fixed point and the second is an invariant set of the interval @xmath69 whose basins of attraction are separated by an unstable fixed point ( fig.[pic : kl](c ) ) . for @xmath70 there
exists the interval @xmath69 ( fig.[pic : kl ] ( d ) ) which attract all trajectories of the map @xmath52 . check that the map @xmath52 on the @xmath69 acts like a lorenz - type map .
the map @xmath52 will be a lorenz - type if @xcite 1 .
[ lst : cndlrnc1 ] the derivative @xmath71 for any @xmath72 ; 2 .
[ lst : cndlrnc2 ] the set of preimages of the point of discontinuity , @xmath73 is dense in @xmath69 ; 3 .
[ lst : cndlrnc3 ] @xmath74 .
one can see that ( [ lst : cndlrnc1 ] ) and ( [ lst : cndlrnc3 ] ) are satisfied . according to @xcite
the condition ( [ lst : cndlrnc2 ] ) is satisfied if @xmath75 for the map @xmath52 on the interval @xmath69 we have @xmath36 and inequality ( [ eq : cndlrnciii ] ) is obviously satisfied .
therefore the map @xmath52 on the interval @xmath69 acts like a lorenz - type map .
the possible structure of the invariant set of interval @xmath69 is controlled by value @xmath76 .
let us find conditions under which the map @xmath52 is strongly transitive .
recall @xcite that a lorenz - type map @xmath52 is strongly transitive if for any subinterval @xmath77 there is @xmath78 such that @xmath79 under the condition ( [ eq : cndlrnciii ] ) the sufficient condition for strong transitivity on the interval @xmath69 are ( @xcite ) @xmath80 where @xmath81 are such that they satisfy the following conditions @xmath82\ ] ] @xmath83 , g_3^{n_2 + 1}(c ) \in [ b , d).\ ] ] now let us find condition for the parameter values of the map @xmath52 under which @xmath84 .
consider the condition ( [ eq : cndn1 ] ) .
let us take @xmath85 , where @xmath86 .
it is clear that ( [ eq : cndn1 ] ) holds if the parameter @xmath34 satisfies the following conditions @xmath87 let us require that ( [ eq : cnd : strtransn1rngy0 ] ) for @xmath88 is satisfied for @xmath89 from inequalities ( [ eq : cnd1dbeta ] ) , ( [ eq : cndlrnc1 ] ) and the definitions of the boundaries @xmath90 and @xmath59 , it follows that this requirement holds if @xmath91 similarly , for @xmath92 we get @xmath93 by the same argument as indicated above we obtain that for @xmath94 inequalities ( [ eq : cnd : strtransn2rngy0 ] ) hold if the conditions ( [ eq : cndn2 ] ) are satisfied .
for example , let us fix @xmath95 , that is @xmath96 . in this case
the map @xmath97 is strongly transitive and therefore it follows from the theorem 3.1.1 . of @xcite
that the periodic points are dense in @xmath69 .
we note that all of these periodic points are unstable @xmath98 and @xmath69 is a chaotic attractor .
fig.[pic : kl ] ( c),(d ) illustrates the dynamics of the map @xmath52 on the interval @xmath69 for regions @xmath99 and @xmath100 respectively .
in this section consider the case @xmath101 and @xmath102 . this case corresponds to instability of the unique fixed point @xmath103 . since parameter @xmath12 is sufficiently small ,
the dynamic of the map @xmath18 is a relaxation @xcite similarly by to the case of ode ( [ eq : modelfhn ] ) .
the distinctive characteristic of these systems is two time and velocity scales , so - called `` fast '' and `` slow '' motions .
basically fast motions are provided by `` frozen '' system in which slow variables are regarded as a parameters , and it is assumed that small parameter of the system equals to zero .
slow motions with size of order of the small parameter are given by evolution of `` frozen '' variable . in case of the map @xmath18
, @xmath1 is the fast variable and @xmath2 is the slow one .
let us study the fast and slow motions in our system .
the fast motions of the model ( [ eq : stpmp ] ) is approximately described by the map ( [ eq : onedmap ] ) . as indicated above , the dynamics of the map ( [ eq : onedmap ] )
can be both , regular and chaotic according to the parameter value ( fig.[pic : kl ] ) .
consider now under conditions ( [ eq : cnd1dbeta ] ) , ( [ eq : cndlrnc1 ] ) slow motions of the map @xmath18 on the phase plane @xmath104 in the region separated by the following inequalities @xmath105 in the case @xmath101 the motions of the map @xmath18 have slow features within thin layer @xmath106 ( thickness is of the order @xmath107 , @xmath108 ) @xcite near invariant line @xmath109 where @xmath110 directly from the map @xmath18 it can be obtained that @xmath111 is invariant line not only for @xmath112 but for @xmath113 also .
one can see that the dynamics on the line @xmath111 is defined by one - dimensional linear map @xmath114 it is clear that the map ( [ eq : smonedmap ] ) has stable fixed point @xmath115 .
therefore for @xmath116 the trajectories on @xmath111 with initial conditions @xmath117 moves to the line @xmath58 .
all trajectories from layer @xmath106 behave in the same way .
let us now consider the stability of the slow motions from @xmath106 relatively to the fast ones . since in the case
@xmath118 each point of the @xmath119 is stable fixed point of the fast map ( [ eq : onedmap ] ) then invariant curve @xmath111 is stable with respect to the fast motions .
it is follows from the previous description that when @xmath12 is small enough , the structure of the partition of the phase @xmath104-plane into trajectories does nt significantly change with respect to case of equations ( [ eq : onedmap ] ) , ( [ eq : smonedmap ] ) .
the trajectories of the map @xmath18 are close to the trajectories of ( [ eq : smonedmap ] ) within the layer of the slow motions near @xmath111 and to the trajectories of ( [ eq : onedmap ] ) outside these layer .
therefore , the motions of the map @xmath18 are also formed by the slow - fast trajectories .
let the initial conditions of @xmath18 belong to neighborhood @xmath120 of the invariant curve @xmath111 .
any of these trajectories moves within the layer of the slow motions down to the neighborhood of the critical point @xmath121 : @xmath122 , and continue their motions according to the fast motions ( see fig .
[ pic : kl](a ) ) , along @xmath123 . since @xmath124 the trajectory of the map @xmath52 with initial condition @xmath125 tends to invariant interval @xmath69 ( see fig . [
pic : kl ] ( d ) ) .
therefore the fast motions of the map @xmath18 with initial conditions @xmath121 falls into some region @xmath126 ( see fig [ pic : chatt ] ) , @xmath127 if @xmath128 , where @xmath129 is the parallelogram : @xmath130 in other words the region @xmath131 is one parametrical family @xmath34 - indexed of invariant intervals . as @xmath69 is attractor , then @xmath126 is also two - dimensional attracting region .
since the map @xmath52 has interval @xmath69 for @xmath132 , then a trajectory involving the map @xmath18 belongs to the region @xmath126 as long as its variable @xmath2 do not culminate approximately to the value corresponding to the line @xmath133 ( fig .
[ pic : bd_betay0 ] ) . at the same time
variable @xmath2 is slowly increasing for @xmath134 as @xmath135 .
thus , within the region @xmath126 the variable @xmath2 continues to increase and variable @xmath1 evolution is close to chaotic trajectory of the map @xmath52 . over line @xmath136
the map @xmath52 has stable fixed point which attracts all trajectories ( see fig.[pic : kl ] ( b ) ) . hence if the magnitude of the variable @xmath2 becomes about @xmath59 then trajectory of the map @xmath18 returns into neighborhood of @xmath106
. then the process is repeated . as a result of these slow - fast motions
the attractor @xmath137 of the system @xmath18 @xmath104 phase plane appears as in ( fig.[pic : chatt](a ) ) . to characterize the complexity of the attractor a we calculated numerically its fractal dimension @xmath138 . at
appears that @xmath138 takes non - integer values between 1.35 and 1.9 ( fig.[pic : chatt](b ) ) .
therefore @xmath137 is chaotic attractor .
for the parameter values from fig .
[ pic : chatt ] ( b ) maximum of the fractal dimension @xmath139 is accomplished then @xmath140 .
let us prove that the system ( [ eq : stpmp ] ) has an attractor @xmath137 for different values @xmath12 and let us find conditions under which the map @xmath18 has an invariant region . to do that , we construct some ring - like region @xmath141 .
denote by @xmath142 the outer boundary and by @xmath143 the inner boundary of the @xmath141 .
the @xmath141 is an invariant region if from the conditions @xmath144 and @xmath145 follows that @xmath146 .
its should be fulfilled if 1 .
[ lst : cndinv1 ] the vector field of the map @xmath18 at the boundary @xmath142 and @xmath143 is oriented inwards to @xmath141 ; 2 .
[ lst : cndinv2 ] the images @xmath147 of the boundaries @xmath142 , @xmath143 and the discontinuity line @xmath16 belong to @xmath141 .
we construct boundaries @xmath142 and @xmath143 in the form of some polygons .
taking into account the condition ( [ lst : cndinv1 ] ) and analyzing the vector field of the map @xmath18 at the lines with uncertain slope we have found the shape of @xmath142 and @xmath143 ( see fig .
[ pic : pm_invrg ] ( a ) ) .
the equations of the boundaries of @xmath142 and @xmath143 are presented in the appendix .
analysis of the position of the images @xmath148 on the phase plane @xmath104 show that the condition ( [ lst : cndinv2 ] ) holds if @xmath149 ( see section [ sec : oneddyn ] ) and inequalities @xmath150 @xmath151 } { j - j_{min } } \right\}\ ] ] @xmath152 @xmath153 @xmath154 @xmath155 @xmath156 are satisfied ( the parameters @xmath157 , @xmath158 and @xmath159 have been introduced in appendix ) .
fig.[pic : pm_invrg](b),(c ) illustrates the transformation of @xmath141 by the action of the map @xmath18 under conditions ( [ eq : cndinvepsj ] ) .
the inequality ( [ eq : cndinvepsj ] ) determine the parameter region @xmath160 in the parameter plane @xmath161 ( fig .
[ pic : bdinvrg ] ( a ) ) . since @xmath162
then the trajectories with initial conditions @xmath144 execute rotation motion around the fixed point @xmath163 forming some attractors @xmath137 .
we calculated numerically fractal dimensional @xmath138 ( fig.[pic : bdinvrg](b ) ) in terms of @xmath12 .
its shows that @xmath137 is chaotic attractor .
the possible structure of the attractor @xmath137 in the phase plane is shown in fig.[pic : pm_bursts](a ) .
[ pic : pm_bursts](b ) illustrates time evolution of the variable @xmath1 corresponding to chaotic attractor @xmath137 .
it shows chaotic spiking - bursting neural activity .
[ pic : bdinvrg ] ( b ) shows that fractal dimension @xmath138 , on average , tends to decrease with increasing @xmath12 .
there is a critical value , @xmath164 , for which fractal dimension has a minimal value @xmath165 .
the mechanism of this decreasing can be accounted for by the different types of the dynamics of the variable @xmath2 for different @xmath12 .
as the parameter @xmath12 increases , the velocity of the variable @xmath2 is expected to climb .
therefore `` life time '' of the trajectories in the strip corresponding to lorenz - map dynamics is reduced . as a result ,
the chaotical motions are reduced .
at previous sections it was shown that system ( [ eq : stpmp ] ) allows to simulate spiking - bursting behavior of the neuron . here
we show that other regimes of the neural activity ( phasic spiking and burstings threshold excitation , subthreshold oscillation , tonic spiking and chaotic spike generation ) can be obtained by using the map @xmath18 also . to do that we neglect the first inequality in ( [ eq : condd ] ) , inequality ( [ eq : cnd1dbeta ] ) and condition @xmath166 .
studying response of the neurons to the influence of external stimulus is one of important task of neuroscience @xcite associated with the problem of information transmission in neural system .
usually external stimulus is represented as the injection of electrical current into the neuron .
let us suppose that the neuron is not excited initially , that is , it is in steady state ( rest ) . in the model ( [ eq : stpmp ] ) such state of neuron corresponds to stable fixed point @xmath163 .
consider the response of the system ( [ eq : stpmp ] ) to pulse type stimulus .
we assume that the duration of each pulse is small enough ( see fig . [
pic : pm_ex2t](a ) ) and its action is equal to the instantaneous changing of the variable @xmath1 on the pulse amplitude .
besides , we suppose here that @xmath101 and therefore the dynamics of the system ( [ eq : stpmp ] ) is a relaxation
. for this parameter region the system ( [ eq : stpmp ] ) has two thresholds .
the first threshold is determined ( see fig . [
pic : pm_ex2t](c ) ) by the thin layer of the slow motions near the following invariant line @xmath167 where @xmath168 analogously , the second threshold is defined ( see fig . [
pic : pm_ex2t](c ) ) as the thin layer of the slow motions near invariant line @xmath169 where @xmath170 denote by the @xmath171 and @xmath172 ( @xmath173 ) the values of the variables @xmath1 and @xmath2 after stimulation respectively .
let @xmath174 be the trajectory of the system ( [ eq : stpmp ] ) with this initial conditions .
in other words @xmath174 is response of the system ( [ eq : stpmp ] ) to pulse input .
\(i ) if the amplitude of stimulus is not enough ( fig .
[ pic : pm_ex2t](a),(i ) ) for overcoming the first threshold , then the maximum of the response will be about amplitude of the stimulus .
therefore , in this case the generation of the actions potential does not take place .
\(ii ) let us increase the amplitude of the stimulus as it breaks the first threshold but at the same time it is not enough for overcoming of the second threshold ( fig .
[ pic : pm_ex2t](c),(ii ) ) . in this case
the fast motions of the map @xmath18 will be close to the fast motions of the nap @xmath52 on interval @xmath69 for @xmath132 . and so , the trajectory of the map @xmath18 perform some number of irregular oscillations around discontinuity line @xmath19 ( fig .
[ pic : pm_ex2t](c)(ii ) ) .
after that , the trajectory @xmath174 within layer near @xmath175 tends to fixed point @xmath163 ( fig .
[ pic : pm_ex2t](c)(ii ) ) .
such trajectory @xmath174 forms the region of phasic bursting activity @xcite with irregular number of spikes ( fig .
[ pic : pm_ex2t](b),(ii ) ) .
\(iii ) if the amplitude of the stimulus ( fig .
[ pic : pm_ex2t](a),(iii ) ) is enough for overcoming the second threshold , then the point @xmath176 belongs to the region of attractor of the invariant line @xmath177 , where @xmath178 with @xmath179 therefore trajectory @xmath174 tends to thin layer of slow motions near invariant line @xmath177 .
it moves within thin layer to the neighborhood of the point @xmath180 ( fig .
[ pic : pm_ex2t](c),(iii ) ) and its motions continue along fast motions .
these motions are close to the trajectories of the map @xmath52 for @xmath181 ( see fig . [
pic : kl](a ) )
. therefore the trajectory @xmath174 tends to the layer near stable invariant line @xmath175 .
after that the trajectory @xmath174 moves within thin layer near @xmath182 and it tends to the fixed point @xmath163 ( fig . [ pic : pm_ex2t](c),(iii ) ) . in this case trajectory
@xmath174 corresponds to phasic spike @xcite ( fig .
[ pic : pm_ex2t](b),(iii ) ) .
let us consider dynamics of the map @xmath18 under following conditions on the parameters .
@xmath183 one can see from the jacobian matrix that in this case the fixed point @xmath163 has a complex - conjugate multipliers .
this point is stable for @xmath184 and unstable for @xmath102 .
therefore the piece - wise map @xmath18 produces neimark - sacker like bifurcation ( in classical case of neimark - sacker bifurcation the map is smooth ) .
the fixed point @xmath163 is surrounded for @xmath116 by an isolated stable attracting close curve @xmath185 ( fig .
[ pic : pm_so](a ) ) .
the oscillations corresponding to the @xmath185 occur under the threshold of excitability of the neuron and therefore it is called in neuroscience @xcite , subthreshold oscillations ( fig .
[ pic : pm_so](b ) ) .
let us consider again the relaxation ( @xmath101 ) dynamics of the map @xmath18 in the case @xmath102 , that is when fixed point @xmath163 is unstable .
additionally we assume that the parameters of the map @xmath18 has satisfied the following conditions @xmath186 @xmath187 in this case the invariant line @xmath188 separate the fast motions on two flows ( fig .
[ pic : pm_cts](a ) ) in the neighborhood of the discontinuity line @xmath19 .
the first flow forms the trajectory performing the chaotic oscillations near the line @xmath19 ( fig . [
pic : pm_cts](a ) ) .
their dynamics are close to the dynamics of the map @xmath52 on interval @xmath69 for @xmath189 .
the second flow consists of the trajectories overcoming the second threshold ( fig .
[ pic : pm_cts](a ) ) .
it moves to neighborhood of the stable invariant line @xmath177 .
after that these trajectory tends to the stable invariant line @xmath175 and the described process is repeated .
these trajectories form chaotically switching flow from one to other . as a result
is the appearance of a chaotic attractor @xmath190 on phase plane ( fig .
[ pic : pm_cts](a ) ) .
the fractal dimension @xmath191 .
the attractor @xmath190 determines chaotic regime of spiking activity over the background of the subthreshold oscillations ( fig .
[ pic : pm_cts](b ) ) .
let the parameters of the map @xmath18 satisfy the same conditions as in the case of previous subsection [ sec : closeinvcurve ] with exception of inequality ( [ eq : cndforkmn ] ) . in this case
the parameter @xmath14 is small enough .
therefore the trajectories with initial conditions from neighborhood of @xmath192 do not change direction of motion when they intersect the discontinuity line @xmath19 .
this leads to the appearance on the phase plane of the motions between layer near @xmath175 and @xmath177 .
such dynamics leads to forming close invariant curve @xmath193 ( fig .
[ pic : pm_ts](a ) ) .
so there exists only one attractor on the phase plane formed by this invariant closed curve .
this determines tonic spiking regimes of neural activity ( fig .
[ pic : pm_ts](b ) ) .
a new phenomenological model of neural activity is proposed .
the model can reproduce basic activity modes such as spiking , chaotic spiking - bursting , subthreshold oscillations etc . of real biological neurons .
the model is a discontinous two - dimensional map based on the discrete version of the fitzhugh - nagumo system and dynamical properties the lorenz - like map .
we have shown that the dynamics of our model display both regular and chaotic behavior . we have studied the underlying mechanism of the generation of chaotic spiking - bursting oscillations .
sufficient condition for existence chaotic attractors in the phase plane are obtained . in spite of idealization ,
the dynamical modes which are demonstrated in our model are in agreement with the neural activity regimes experimentally found in real biological systems .
for example , subthreshold oscillations ( see fig . [
pic : pm_so](b ) ) is a basic regime of inferior olive ( i.o . )
neurons @xcite .
inferior olive neurons belong to the olivo - cerebellar network which plays a key role @xcite in organization of vertebrate motor control .
it is also typical for i.o .
neuron @xcite a spiking regime over the chaotic subthreshold oscillations ( see fig . [
pic : pm_cts ] ) .
the spiking - bursting activity is significant for many types of neurons , in particular in hippocampal pyramidal cell @xcite and thalamic cells @xcite .
the table summarizes results on gallery of behavior of neural activity showed by our model .
we hope that our model will be useful to understand the mechanism of neural pattern formation in large networks . & * the regimes of neuronal activity * + & + @xmath184 , @xmath101 @xmath194 , @xmath195 ( spike ) @xmath196 ( bursts ) & phasic spikes and bursts + & + @xmath102 , @xmath197 @xmath198 @xmath199 & subthreshold oscillations + & + inequalities ( [ eq : cndinvepsj ] ) . & chaotic bursting oscillations + & + @xmath102 , @xmath101 @xmath200 @xmath201 & chaotic spiking + & + @xmath102 , @xmath101 @xmath202 .
& tonic spiking +
this work was supported partly by university paris 7-denis diderot and in part by the russian foundation for basic research ( grant 06 - 02 - 16137 ) and leading scientific schools of the russian federation ( scientific school 0 7309.2006.2 ) .
the boundary @xmath142 is given by @xmath203 with @xmath204 @xmath205 @xmath206 @xmath207 + f(d ) } { m_0},\ ] ] @xmath208 the boundary @xmath143 is given by @xmath209 with @xmath210 _ e.r .
kandel , j.h .
schwartz , t.m jessell , _ principles of neural science , prentice - hall int .
inc . , 1991 .
rabinovich , p. varona , a.i .
selverston , h.d.i .
abarbanel , _ dynamical principles in neuroscience , reviews of modern physics , 78(4 ) ( 2006 ) , 1213 . _ h.r .
wilson , j.d.cowan_ excitatory and inhibitory interaction in localized population of model neurons . biophys .
j. , 12 ( 1972 ) , pp 1 - 24 .
_ r. fitzhugh _ mathematical models of the threshold phenomena in the nerve membrane . bull math .
v.17 , pp .
257 - 287 ( 1955 ) _ j.l .
hindmarsh , r.m . rose _
, a model of neuronal bursting using three coupled first order differential equations : philos .
trans r. soc .
london , ser .
b221 , 87 - 102 ( 1984 ) . _ c. morris , h. lecar _ , voltage oscillations in the barnacle giant muscle fiber .
j. v25 , 1981 , p. 87
chialvo , _ generic excitable dynamics on the two - dimensional map , chaos solitons fract . 5 ( 1995 ) 461 - 480 .
_ o. kinouchi , m. tragtenberg , _ modeling neurons by simple maps , int
. j. bifurcaton chaos 6 ,
n 12 a ( 1996 ) , 2343 - 2360 . _ s.kuva , g.lima , o.kinouchi , m.tragtenberg , a. roque , _ a mimimal model for excitable and bursting element .
neurocomputing 38 - 40 , pp .
255 - 261 ( 2001 ) .
_ g. de vries , _ bursting as an emergent phenomenon in coupled chaotic maps , phys .
e 64 ( 2001 ) 051914 .
rulkov , i.timofeev , m. bazhenov , _ oscillations in large - scale cortical networks : map - based model .
j. of computational neuroscince 17 ( 2004 ) , 203 - 223 .
traub , j.g.r .
jefferys , m.a .
whittington , _ fast oscillations in cortical circuits . the mit press , massachusettes , 1999 .
_ llinas r. and yarom , y. , _ oscillatory properties of guinea - pig inferioir olivary neurines and their pharmacological modulation : an in vitro study . j. physol .
, lond . , 315 , 569 - 84 , 1986 . _ llinas r. , _ i of vortex
. from neurones to self.the mit press , massachusettes , 2002 . _
b. cazelles , m. courbage , m. rabinovich , _ anti - phase regularization of coupled chaotic maps modelling bursting neurons .
europhysics letters , 56 ( 4 ) , pp .
504 - 509 ( 2001 ) .
rulkov , _ regularization of synchronized chaotic bursts , phys .
86 , 183 - 186 ( 2001 )
_ n.f . rulkov .
_ modeling of spiking - bursting neural behavior using two - dimensional map .
rev.e , v.65 , p. 0.41922
shilnikov , n.f .
_ origin of chaos in a two dimensional map modeling spiking - bursting neural activity .
j. bifurc .
chaos , v.13 , n11 , pp .
3325 - 3340 ( 2003 ) .
shilnikov , n.f .
rulkov . _ subthreshold oscillations in a map - based neron model .
physics letters a 328 , pp .
177 - 184 ( 2004 ) .
_ h. tanaka _ design of bursting in two - dimensional discrete - time neron model.physics letters a 350 , pp .
228 - 231 ( 2006 ) .
_ m.courbage , v.b .
kazantsev , v.i .
nekorkin , v. senneret .
_ emergence of chaotic attractor and anti - synchronization for two coupled monostable neurons .
chaos 12 , pp .
1148 - 1156 ( 2004 ) .
izhikevich , f. hoppensteadt , _ classification of bursting mappings .
j. bifurcation and chaos , v.14 , n11 , pp 3847 - 3854 , ( 2004 ) .
afraimovich , sze - bi hsu .
_ lectures on chaotic dynamical systems , american mathematical society .
press , 354 p. ( 2003 ) _
afraimovich , l.p .
_ strange attractors and quasiattractors . in book `` nonlinear dynamics and turbulence '' ( eds .
barenblatt , g. iooss , d.d .
joseph , pitam , boston , 1983 , pp .
1 - 34 _ v.i .
arnold , v.s .
afraimovich , yu . s. ilyashenko , l.p .
_ bifurcation theory , dyn .
sys . v. encyclopaedia mathematics sciences , springer , berlin , 1994 _ bernardo l.s . , foster r.p . _ oscillatory behaviour in the inferior olive neurons : mechanism , modulation , cell agregates .
brain res . bull .
1986 , v.17 , p.773 r.s.k .
wang and d.a .
prince afterpotential generation in hippocampal pyramidal cells .
j. neurophysiol .
v45 , 1981 , p.86 m. deschenes , j.p .
roy and m. steriade thalamic bursting mechanism : an inward slow current revealed by membrane hyperpolarization .
brain res . 239 , 1982 , p.289 . with @xmath213 for regions : ( a ) - @xmath214 , @xmath215 , ( b ) - @xmath216 , @xmath217 , ( c ) - @xmath100 , @xmath218 , ( d ) - @xmath99 , @xmath219.,title="fig:",scaledwidth=75.0% ] with @xmath213 for regions : ( a ) - @xmath214 , @xmath215 , ( b ) - @xmath216 , @xmath217 , ( c ) - @xmath100 , @xmath218 , ( d ) - @xmath99 , @xmath219.,title="fig:",scaledwidth=75.0% ] + on the phase plane @xmath104 ; ( b ) waveform of relaxation spike - bursting oscillations generated by the map @xmath18
parameter values : @xmath220 . ,
title="fig:",scaledwidth=75.0% ] + on the phase plane @xmath104 ; ( b ) waveform of relaxation spike - bursting oscillations generated by the map @xmath18 .
parameter values : @xmath220 . , title="fig:",scaledwidth=75.0% ]
+ on the parameter plane @xmath161 , parameter values : @xmath224 .
( b ) fractal dimension @xmath138 of the attractor @xmath137 versus parameter @xmath12 , @xmath225 , title="fig:",scaledwidth=75.0% ] + on the parameter plane @xmath161 , parameter values : @xmath224 .
( b ) fractal dimension @xmath138 of the attractor @xmath137 versus parameter @xmath12 , @xmath225 , title="fig:",scaledwidth=75.0% ] + on the phase plane @xmath104 .
( b ) spike - bursting oscillations generated by the map @xmath18 . parameters
value @xmath226 , title="fig:",scaledwidth=70.0% ] + on the phase plane @xmath104 .
( b ) spike - bursting oscillations generated by the map @xmath18 .
parameters value @xmath226 , title="fig:",scaledwidth=70.0% ] + ) to positive pulse @xmath227 : ( a ) - three different amplitude of the stimulus ; ( b ) - the behavior of variable @xmath1 ( membrane potential of neuron ) ; ( c ) - the phase plane .
parameter values : @xmath228 , title="fig:",scaledwidth=70.0% ] + ) to positive pulse @xmath227 : ( a ) - three different amplitude of the stimulus ; ( b ) - the behavior of variable @xmath1 ( membrane potential of neuron ) ; ( c ) - the phase plane .
parameter values : @xmath228 , title="fig:",scaledwidth=70.0% ] + .
( b ) subthreshold oscillations generated by map @xmath18 .
parameter values : @xmath229 , title="fig:",scaledwidth=70.0% ] + .
( b ) subthreshold oscillations generated by map @xmath18 .
parameter values : @xmath229 , title="fig:",scaledwidth=70.0% ] + .
( b ) chaotic spiking against the background subthreshold oscillations .
parameter values : @xmath230,title="fig:",scaledwidth=70.0% ] + .
( b ) chaotic spiking against the background subthreshold oscillations .
parameter values : @xmath230,title="fig:",scaledwidth=70.0% ] + | we propose a discrete time dynamical system ( a map ) as phenomenological model of excitable and spiking - bursting neurons .
the model is a discontinuous two - dimensional map .
we find condition under which this map has an invariant region on the phase plane , containing chaotic attractor .
this attractor creates chaotic spiking - bursting oscillations of the model .
we also show various regimes of other neural activities ( subthreshold oscillations , phasic spiking etc . )
derived from the proposed model .
* the observed types of neural activity are extremely various .
a single neuron may display different regimes of activity under different neuromodulatory conditions .
a neuron is said to produce excitable mode if a `` superthreshold '' synaptic input evokes a post - synaptic potential in form of single spikes , which is an order of magnitude larger than the input amplitude . while a `` subthreshold '' synaptic input evokes post - synaptic potentials of the same order . under some conditions
a single spike can be generated with arbitrary low frequency , depending on the strength of the applied current .
it is called spiking regime .
an important regime of neural activity is bursting oscillations where clusters of spikes occur periodically or chaotically , separated by phases of quiescence .
other important observed regimes are phasic spikes and bursts , subtreshold oscillations and tonic spiking . understanding dynamical mechanisms of such activity in biological neurons has stimulated the development of models on several levels of complexity .
to explain biophysical membrane processes in a single cell , it is generally used ionic channel - based models @xcite .
the prototype of those models is the hodgkin - huxley system which was originally introduced in the description of the membrane potential dynamics in the giant squid axon .
this is a high dimensional system of nonlinear partial differential equations .
another class of neuron models are the phenomenological models which mimic qualitatively some distinctive features of neural activity with few differential equations .
for example , the leaky integrate - and - fire model , hindmarsh - rose and fitzhugh - nagumo model etc .
a new important subclass of phenomenological models is the map - based systems .
basically such models are designed with the aim of simulating collective dynamics of large neuronal networks .
the map - based models possess at least the same features of ordinary differential equations ( ode ) models , and have more simple intrinsic structure offering an advantage in describing more complex dynamics . in order to model basic regimes of neural activity we design new family of maps that are two - dimensional and based on discrete fitzhugh - nagumo system
in which we introduce heaviside step function .
the discontinuity line determines the excitation threshold of chaotic spiking - bursting oscillations . for some domain of the parameters
, we found on phase plane an invariant bounded region containing chaotic attractor with spiking - bursting activity .
the interesting fact is that the dynamical mechanism , leading to chaotic behavior of our two - dimensional map is induced by one - dimensional lorenz - like map .
we demonstrate also that our model can display rich gallery of regimes of neural activity such as chaotic spiking , subthreshold oscillations , tonic spiking etc.all these modes play important role in the information processing in neural systems . * |
graphene , a novel two - dimensional form of carbon,@xcite displays exciting new physics , distinct from that of the two - dimensional electron gas ( 2deg ) .
a striking example in the presence of a magnetic field is the anomalous integer quantum hall effect , which has been observed at room temperature.@xcite the treatment of electronic interactions is also challenging in graphene , due to the structure and symmetry of the dispersion relations at the two inequivalent dirac points .
when scattering between these points is negligible , the chirality of the electrons results in a suppression of backscattering.@xcite this together with graphene s reduced level of defects and impurities , makes it highly efficient at charge transport and a promising candidate for use in nanotechnology.@xcite however , in order for this potential to be realized , we need to better understand the precise nature of defects in graphene .
optical measurements are a particularly useful tool for probing this.@xcite in this paper , we study the _ magneto - optical _ response of graphene in a strong perpendicular magnetic field in the presence of a low density of charged impurities . infrared studies of landau level ( ll ) transitions have reported significant departures from the bare ( non - interacting ) cyclotron resonance.@xcite whether this can be attributed solely to interaction effects and a possible role of disorder remains unclear . here
we develop a general formalism for studying localized collective modes of magnetoplasmon and spin - wave types and determine their optical signatures .
we treat the electron - electron ( @xmath3-@xmath3 ) interactions beyond the conventional mean - field / rpa level and the electron - impurity interaction exactly in the high magnetic field regime .
our results are robust for a range of impurity screening strengths .
we establish the existence of an exact symmetry for collective excitations , which should be observable by magneto - optical spectroscopy.@xcite this symmetry , briefly speaking , connects the magneto - optical _
electron_-like excitations of , e.g. a _ positively _ charged donor @xmath0 , with _
excitations for a _ negatively _ charged acceptor @xmath1 .
this duality is a consequence of the electron - hole symmetry between single - particle states in the lower and upper cones of graphene.@xcite furthermore , we establish exact optical selection rules , which demonstrate that the `` dual '' collective excitations with total angular momenta @xmath4 are active in two different circular polarizations @xmath5 and , besides having the same energies , exhibit the same oscillator strengths .
therefore , a qualitative distinction of graphene from the conventional 2deg,@xcite is that there are strong dipole - allowed transitions in _ both _ circular polarizations sensitive to the charge of impurity .
we show an example of this symmetry for the lowest ll @xmath6 in fig .
[ fig - emz ] .
each ll in graphene consists of four sublevels , due to spin and valley ( pseudospin ) splitting .
we denote by @xmath7 a many - electron ground state corresponding to the sublevel filling factor @xmath8 of a particular ll .
for sublevels @xmath9 of ll with number @xmath10 , the eigenstates and eigenenergies of excitations localized on the @xmath0 with @xmath11 , coincide precisely with those with @xmath12 formed at filling factor @xmath13 of the ll with number @xmath14 , localized on the @xmath1 .
this effect represents what remains of the electron - hole symmetry after it has been broken by a charged impurity .
we consider collective excitations in a system of electrons interacting via a screened coulomb potential @xmath15 [ ] in the presence of an additional external field @xmath16 .
all the results presented here are for a coulomb impurity , @xmath17,@xcite but the method is applicable for any non - singular axially symmetric potential @xmath18 . the parameters @xmath19 and @xmath20 are the effective dielectric constants for @xmath3-@xmath3 interaction and electron - impurity screening,@xcite respectively . a composite index @xmath21 is used to designate the ll number @xmath10 , the spin @xmath22 , and pseudospin @xmath23 projections .
low - energy collective excitations correspond to the promotion of one electron from one of the uppermost filled levels @xmath24 to a higher lying empty level @xmath25 ( see fig .
[ fig - emz ] ) .
electrons in graphene follow a linear dispersion relation close to the zeroes of energy ( dirac points ) , which occur at two inequivalent points in the brillouin zone , the @xmath26 and @xmath27 valleys.@xcite we describe a perpendicular magnetic field @xmath28 by the symmetric gauge @xmath29 , consistent with the axial symmetry of @xmath18 . a single electron wavefunction in , e.g. the @xmath30 valley ( pseudospin @xmath31 ) , is a four - component spinor @xmath32 here
, @xmath10 is an integer ll number , @xmath33 is a 2deg wavefunction with oscillator quantum number @xmath34 , @xmath35 , @xmath36 ( with @xmath37 ) and @xmath38 is the spin part corresponding to two spin projections @xmath39.@xcite the wavefunction in the @xmath40 valley ( pseudospin @xmath41 ) is obtained by changing the order of the spinor components .
the spinors are the eigenstates of the generalized [ orbital plus isospin ( sublattice ) ] angular momentum projection @xmath42 [ ] with half - integer eigenvalues @xmath43 .
the single - electron energies are given by @xmath44 , where @xmath45 is the cyclotron energy in graphene , @xmath46 is the zeeman splitting and @xmath47 is a possible valley splitting.@xcite using the hole representation for all filled levels , @xmath48 and @xmath49 for @xmath50 , we introduce operators of collective excitations as @xmath51 with expansion coefficients satisfying the condition @xmath52 . an exact quantum number @xmath11 is an eigenvalue of the total @xmath53 ; for neutral collective excitations @xmath11 is _ integer _ and of purely _
orbital _ nature .
it has a direct geometrical meaning@xcite determining the average positions of the electron and the hole relative to the impurity , i.e. @xmath54 + \delta_{n_2,0 } - \delta_{n_1,0 } \right ) \ell_b^2\ ; , \end{aligned}\ ] ] where the states are defined as @xmath55 . considering the total hamiltonian @xmath56 matrix elements @xmath57 , we find that the effective hamiltonian is given by @xmath58 here @xmath59 denotes the single - particle ll energy renormalized by @xmath3-@xmath3 exchange self - energy corrections @xmath60 [ ] . since kohn s theorem is not applicable in graphene , these corrections lead to the renormalization of the bare cyclotron energy , @xmath61 [ ] . for the @xmath62 transition ( denoted hereafter by @xmath63 ) , @xmath64 is due only to exchange interactions with the lower cone and @xmath65 . here
@xmath66 is the characteristic energy of coulomb interactions in strong @xmath28 , @xmath67 . due to the spinor form of the single - particle wavefunctions ,
the impurity matrix elements in graphene are connected with those in the conventional 2deg,@xcite @xmath68 , according to @xmath69 the two - body interaction in consists of the direct electron - hole ( @xmath3-@xmath70 ) attraction and exchange @xmath3-@xmath70 repulsion , i.e.@xmath71 in electron representation , @xmath72note that we neglect the intervalley scattering in graphene by long - range ( coulomb ) potentials .
therefore the two - particle graphene matrix elements are given by @xmath73 \
, , \end{split}\ ] ] where @xmath74 are those used in the conventional 2deg .
we compute the matrix elements for lowest ll s analytically@xcite and those for arbitrary ll s numerically using eq .
( [ u_2d ] ) . in general , an infinite number of excitations ( [ q2 ] ) with the same @xmath11 are mixed by the @xmath3-@xmath3 interactions .
however , those with different single - particle cyclotron energies are only weakly ( @xmath75 ) mixed in strong magnetic fields in graphene and can be neglected.@xcite let us suppose all ll s with @xmath76 are completely filled , all ll s with @xmath77 are empty , and the four ll s with @xmath6 become successively completely filled .
we designate the corresponding filling factors as @xmath2 . for each @xmath8
, there are sixteen possible inter - ll excitations involving the @xmath6 ll as an initial or final state and which have single particle energies of magnitude @xmath78 . here
we concentrate on the four excitations for which no spin or pseudospin - flip occurs ( see the insets in fig .
[ fig - emz ] ) , as these are the only excitations which are optically dipole active .
these have the same single particle energy @xmath79 and , therefore , are strongly ( resonantly ) mixed .
let us discuss some general features of our approach . in the absence of an external potential
all magnetoplasmon states are extended and the corresponding hamiltonian matrix is infinite , see eq.([hamd ] ) .
the magnetoplasmons can be labeled by a continuous quasimomentum @xmath80 and their eigenenergies fill a band of width @xmath81.@xcite in the presence of an impurity , however , some states become localized .
importantly , the basis states are localized two - particle orbitals whose distances from the impurity increase @xmath82.@xcite hence , for localized excitations the scheme is convergent so that the basis can be truncated .
we include the first @xmath83 basis states for each excitation @xmath84 with the total matrix size being @xmath85 for four strongly mixed excitations .
the achieved accuracy in eigenergies of bound states is better than @xmath86 .
figure [ fig - emz](a ) shows for @xmath87 four low - energy branches of magnetoplasmons bound on the @xmath0 for @xmath88 ; two of these branches are degenerate . for large positive @xmath11 ,
the hole is on average much farther from the impurity than the electron , see eq . .
therefore , the @xmath89-@xmath0 attraction dominates over the @xmath90-@xmath0 and @xmath3-@xmath70 interactions . generally ,
for an excitation with the electron in the @xmath91 ll , we find branches with asymptotic @xmath92 energies equal to @xmath93 , when counted from the corresponding renormalized cyclotron energy . as an example , notice the three branches approaching energy @xmath94 and the single branch approaching zero energy in fig .
[ fig - emz](a ) .
these originate , respectively , from the three @xmath95 transitions ( denoted hereafter as @xmath96 ) and from the single @xmath63 transition for @xmath97 .
similar asymptotic behavior can be seen for other filling factors .
the high - energy ( i.e. , above the band ) magnetoplasmons develop for @xmath98 , when the hole is on average closer to the @xmath0 than the electron .
such unusual excited states are bound in 2d because of the confining effect of @xmath28.@xcite fig .
[ fig - emz](b ) shows the spectra for @xmath99 for the @xmath1 which is a `` mirror reflection '' of fig .
[ fig - emz](a ) because of the aforementioned symmetry .
generally , due to the symmetry , results for the @xmath1 at @xmath100 can be obtained from those for the @xmath0 by changing @xmath101 and @xmath102 . for all filling factors ,
the spectra of bound states are qualitatively similar to those shown in fig .
[ fig - emz ] . for @xmath103
, the states can be classified according to the total spin and pseudospin , so the states are either spin and pseudospin singlets or triplets.@xcite only spin and pseudospin singlets are optically active .
let us consider the magneto - optical response in graphene.@xcite in the electric dipole approximation , the interaction of electrons with light of frequency @xmath104 and left ( @xmath105 ) and right ( @xmath106 ) circular polarizations is described by the hamiltonian @xmath107 where @xmath108 is the electric field amplitude and @xmath109 are the pauli matrices acting in the space of the graphene crystal sublattices .
the exact optical selection rules for the collective excitations that follow are : only those with no spin or pseudospin flips and with @xmath4 and @xmath110 are optically active in the two circular polarizations @xmath111 .
we quantify the rate of microwave absorption in the @xmath111 polarization by calculating the dipole transition matrix elements @xmath112 to final states of magnetoplasmons obtained by numerical diagonalization .
figure [ fig - enu ] shows optical properties of states bound on @xmath0 for the @xmath113 and @xmath114 polarizations .
the results for the @xmath1 at @xmath100 can be obtained from those reported here with the change @xmath115 and @xmath116 .
two types of localized states can be optically observed : ( i ) truly bound states , which are split off the continuum and have normalizable wavefunctions , ( ii ) quasibound states within the continuum , which have high probability amplitudes on the impurity and long - range oscillating tails .
the latter may exhibit _ asymmetric _ fano - type optical signatures.@xcite for both polarizations , the upper branch originates mostly from the @xmath63 transitions with some small ( zero at @xmath103 ) admixture of the @xmath96 . with increasing @xmath8 , the number of @xmath63 transitions increases , which leads to the enhanced contribution of the repulsive @xmath3-@xmath70 exchange interactions
this explains the _ blue shift _ of the upper branch to higher energies with increasing @xmath8 . also , its optical strength @xmath117 increases ( fig .
[ fig - enu]a ) while @xmath118 decreases ( fig . [ fig - enu]b ) .
this is explained by the fact that in the @xmath113 ( @xmath114 ) polarization only the electron - like @xmath63 ( hole - like @xmath96 ) transitions are optically active . for the @xmath0 in @xmath113 polarization for @xmath99 .
symbols are as in fig .
[ fig - enu ] . ]
screening was shown to be relevant in graphene,@xcite although the situation is not yet fully understood , particularly for a strong magnetic field .
coulomb impurities with charge @xmath119 belong to the subcritical regime and hence screening effects due to the substrate and the electron system in graphene can be modelled via an effective charge @xmath120,@xcite or by an effective dielectric constant @xmath20 .
however the value relative to the @xmath3-@xmath3 screening constant , @xmath121 is unknown for graphene.@xcite we use @xmath122 in figs .
[ fig - emz ] and [ fig - enu ] and show in fig .
[ fig - emz - don ] how the energies of the excitations for , e.g. the @xmath99 case of fig .
[ fig - enu](a ) , are modified by @xmath123 . as this
ratio increases , the branches of optically active bound states are pushed away from the band of extended magnetoplasmons .
in conclusion , we established the spectra and the symmetries of collective excitations bound on charged impurities in graphene in magnetic fields . our single - impurity theory is applicable to samples with finite impurity density @xmath124 , i.e. when the mean separation between impurities exceeds the size of bound magnetoplasmons .
the intensity of impurity peaks will then be @xmath125 .
recent progress in fabrication of large graphene films@xcite with sizes exceeding the wavelength @xmath126 ( @xmath127@xmath128 ) opens the way for detailed studies of the effects predicted here .
polarization resolved magneto - optical spectroscopy and cyclotron resonance detection using the photoconductive response may be very effective experimental probes .
our results demonstrate the breaking of particle - hole symmetry in a sample with predominantly positive or negative impurities , which may partly explain the observed asymmetry in ll transitions.@xcite the developed method can be extended for defects with short - range potentials , results will be published elsewhere.@xcite we thank d.n .
basov , v.i .
falko , and m.m .
fogler for useful discussions .
we acknowledge funding by epsrc and the warwick north american travel fund ( amf ) .
amf is grateful for hospitality at csu bakersfield .
abd acknowledges cottrell research corporation and the scholarship of kitp , uc santa barbara .
the characteristic size of the localized excitations is of the order of the magnetic length , @xmath129 . since @xmath130 ( @xmath131 , the lattice spacing ) , the exact position of the coulomb impurity in a primitive cell is irrelevant . | we consider collective excitations in graphene with filled landau levels ( ll s ) in the presence of an external potential due to a single charged donor @xmath0 or acceptor @xmath1 impurity . we show that localized collective modes split off the magnetoplasmon continuum and , in addition
, quasibound states are formed within the continuum .
a study of the evolution of the strengths and energies of magneto - optical transitions is performed for integer filling factors @xmath2 of the lowest ll .
we predict impurity absorption peaks above as well as below the cyclotron resonance .
we find that the single particle electron - hole symmetry of graphene leads to a duality between the spectra of collective modes for the @xmath0 and @xmath1 .
the duality shows up as a set of the @xmath0 and @xmath1 magneto - absorption peaks having same energies , but active in different circular polarizations . |
although the x - ray emission from seyfert galaxies is still not fully understood , there are indications that the mystery is starting to unfold .
it is now believed that the emission mechanism is most likely thermal ( see , e.g. , svensson 1996 ; zdziarski et al . 1996 ) .
a flattening of the spectrum above about 10 kev ( e.g. , nandra & pounds 1994 ) strongly suggests the presence of a component due to the reflection of the intrinsic x - ray spectrum by nearby cold matter , which means that the x - ray emitting region is probably located above the accretion disk .
in addition , the fact that the uv and x - ray luminosities from these sources are comparable and that their time variations are tightly correlated suggests that the uv and x - ray emitting regions are located very close to each other .
there are also indications that the broad and redshifted fe k@xmath5 line originates within the cold accretion disk which extends inwards to the last stable orbit ( e.g. , iwasawa et al .
1996 ) .
many of these observational characteristics fit within the currently popular two - phase patchy accretion disk - corona model ( haardt & maraschi 1991 ; haardt & maraschi 1993 ; haardt et al .
1994 ) , in which the x - ray emitting region is located within ` active regions ' ( ar ) above the disk , where the energy dissipation rate greatly exceeds the local gravitational dissipation inside the disk at least for the lifetime of the region .
recent calculations by zdziarski et al .
( 1996 ) show that the thomson optical depth @xmath1 in seyfert galaxies must be close to unity . from a theoretical perspective , this is highly significant given that the sample of sources covers a broad range of black hole masses , accretion rates , and other system parameters .
thus , it is not easy to understand why @xmath2 , especially since the energy transport mechanism has not yet been identified .
haardt et al .
( 1994 ) suggested that the ars may be magnetic flares occurring above the accretion disk s atmosphere and showed that their compactness @xmath6 may be quite high ( @xmath7 ) , so that pairs can be created .
the local compactness parameter @xmath6 is defined as @xmath8 , where @xmath9 is the source luminosity , @xmath10 is the source size and @xmath11 is the thomson cross section . in principle , it is possible to obtain still larger values for the compactness parameter ( @xmath12 few hundred ; see nayakshin & melia 1997a ) , thus creating enough pairs to account for the observed @xmath13 ( zdziarski et al .
this explanation for the observed value of @xmath2 based on the pair equilibrium condition , however , relies on the assumption that the particles are confined to a rigid box , so that no pressure constraints need to be imposed .
this is unphysical for a magnetic flare where the particles are free to move along the magnetic field lines .
therefore , as far as the two - phase model _ without _ a proper pressure equilibrium condition is concerned , the thomson optical depth is a parameter , rather than a calculable quantity . in this _ letter _ , we attempt to address this question of pressure equilibrium during an energetic magnetic flare above the surface of a cold accretion disk .
we show that under the assumptions ( i ) that the energy is supplied to the x - ray emitting region by magnetohydrodynamic waves with a group velocity close to the speed of light and ( ii ) that the emitting region has a very large compactness parameter @xmath14 , the pressure equilibrium between the outwardly directed x - ray flux and the inwardly directed mhd waves leads to the requirement that @xmath15-@xmath16 , in accordance with the apparently `` universal '' value of @xmath1 seen in seyfert galaxies
. we will also show that both assumptions are likely to be valid during a physically plausible magnetic flare above the surface of a cold accretion disk .
the ` universal ' x - ray spectral index of seyfert galaxies ( e.g. , nandra & pounds 1994 ) suggests that the emission mechanism is thermal comptonization with a @xmath17-parameter close to one ( haardt & maraschi 1991 ; haardt & maraschi 1993 ; fabian 1994 ) . for the emission to be dominated by comptonization
, the compactness parameter needs to be large ( @xmath14 ; e.g. , fabian 1994 ) .
observations of seyfert galaxies point to a global compactness parameter @xmath18 ( svensson 1996 and references cited therein ) . for electrons and protons at a single temperature @xmath19 and an electron number density @xmath20 with the assumption of neutrality ,
the gas pressure is @xmath21 .
the radiation energy density may be recast in terms of the luminosity @xmath9 of the source , and therefore its compactness @xmath6 .
the total pressure is @xmath22{\rm , } \ ] ] where @xmath23 is the dimensionless electron temperature .
the compton @xmath17-parameter for thermal electrons is @xmath24 ( see rybicki & lightman 1979 ) and is of order 1 .
thus , since the dimensionless gas pressure in equation ( 1 ) is always smaller than @xmath17 , the radiation pressure dominates over the gas pressure in a one - temperature plasma when @xmath14 .
one consequence of this is that the amount of energy escaping from the source even during one light crossing time is larger than the total particle thermal energy .
thus , there must be an agent that energizes the particles to enable them to radiate at this high rate , and the presence of this agent must be dynamically consistent with the state of the system .
we foresee two possibilities for the nature of this ` agent ' : ( i ) the gravitational field , and ( ii ) an external flow of energy into the system .
these two cases are quite distinct physically . insofar as the first possibility is concerned , the gravitational potential energy of the plasma ( primarily that of the protons ) is dissipated as the gas sinks deeper into the well of the black hole .
the gravitational field does not provide a pressure , but it does compress the gas . however , this leads to an internal ( radiation plus gas ) pressure that varies from source to source as the physical conditions change .
there does not appear to be a scale that sets @xmath1 to have a value of @xmath25 .
for example , in standard accretion disk theory , the inner radiation pressure - dominated regime has an optical depth that depends on several parameters , such as the accretion rate and the @xmath5-parameter ( shakura & sunyaev 1973 ) .
the @xmath5-parameter reflects the rate at which the protons ` use up ' their potential energy , and so a change in this rate leads to a change in the equilibrium optical depth .
it is even less obvious why @xmath1 should be @xmath26 in the gas pressure dominated regimes since there the pressure has no reference to @xmath1 at all .
it seems that when the pressure equilibrium is dictated by the gravitational field ( e.g. , due to a compression of the x - ray emitting region ) , the thomson optical depth should span a range of values depending on the source geometry , the specific parameter values and the particle interactions assumed to operate in the source .
this is not so when the energy is supplied to the x - ray emitting region by an inflow of energy , e.g. , via a magnetic field .
the principal difference between the two cases is that the dynamic portion of the magnetic field supplies a `` ram '' pressure that is related in a known way to its energy density .
if the magnetic energy flux into the x - ray emitting region is known , this also constrains the inwardly directed momentum flux ( the compressional force ) into the system .
thus , the compressional force exerted on the active region by the magnetic field is expected to correlate with the source luminosity .
what makes this useful in terms of setting the optical depth of the system is that a similar correlation exists between the luminosity and the outwardly directed radiation pressure in the emitting region .
but in this case , the pressure also depends on @xmath1 . assuming a spherical geometry for simplicity , the radiation pressure is @xmath27 , where @xmath28 in terms of the source radius @xmath10 and radiation flux @xmath29 .
thus , since all the balance equations are to first order linear in @xmath29 , it is anticipated that the pressure and energy equilibria of the system point to a unique value of @xmath1 .
we explore this possibility in the next section .
let us first suppose that the x - ray source is a sphere with thomson optical depth @xmath1 , and that the energy is supplied radially by magnetohydrodynamic waves .
the waves carry an energy density @xmath30 and propagate with velocity @xmath31 . for definitiveness , we assume that these are alfvn waves , in which case the momentum flux that enters the x - ray source is @xmath32 .
the magnetic energy of the alfvn waves is in equipartition with the oscillating part of the particle energy density , and so we can estimate the gas pressure as being of the same order as the ram pressure of the oscillating part of the magnetic field , i.e. @xmath32 . finally , we assume that all of the wave energy and momentum are absorbed by the source
. the energy equilibrium for the ar is then given by @xmath33 whereas in pressure equilibrium @xmath34 dividing the latter equation by the former , one obtains for the equilibrium thomson optical depth : @xmath35 this value does not depend on luminosity , but it does of course depend on the geometry and @xmath31 .
nayakshin & melia ( 1997a ) show that the alfvn velocity is close to @xmath0 for typical conditions in an accretion disk if its luminosity is a sizable fraction of the eddington luminosity and the magnetic field strength in the ar is comparable to the disk thermal energy density .
suppose now that the geometry is not perfectly spherically symmetric , and that instead the alfvn waves can enter the x - ray source through an area @xmath36 , but the radiation leaves through an area @xmath37 , which is plausibly just the total area of the ar .
this situation may occur if part of the x - ray source is confined by other than the alfvn wave ram pressure , e.g. , by the underlying ( non - dynamic ) large - scale magnetic field ( see below ) . in this case , since the energy balance is now @xmath38 , the equilibrium @xmath1 is changed to @xmath39 to understand the scale represented by the bracketed quantities in this equation , let us consider the physical conditions that are likely to be attained during a short - lived and very energetic magnetic flare above the standard @xmath5-disk .
the magnetic field energy density is a fraction of the underlying disk energy density and the typical size @xmath40 of the flare is expected to be of the order of the disk scale height ( galeev et al .
1979 ; haardt et al .
1994 ; nayakshin & melia 1997a ) .
now , the confinement of the plasma inside the flare , and the observed condition @xmath14 , require that @xmath41 .
since the magnetic stress is much larger than any other stress , the magnetic flux tube adjusts to be in a stress - free vacuum configuration .
the tube is thick ( meaning that its cross sectional radius is of the order of its length ) , since the pressure in the disk s atmosphere is insufficient to balance the tube s magnetic field pressure .
this is due to the fact that the magnetic field is presumably anchored in the disk s midplane , where the pressure is much greater than the atmospheric pressure .
the magnetic waves propagate upwards along the magnetic flux tube , while radiation pressure from the ar is pushing the gas along the magnetic lines , i.e. , downwards to the disk .
this downward direction of the radiation pressure arises naturally in a two - phase model ( unlike the situation within the accretion disk ) since here most of the energy is released above the disk s atmosphere ( see also nayakshin & melia 1997b ) with this in mind , we may now describe heuristically how the magnetic flare develops and how pressure equilibrium is established . as is well known ( parker 1979 ; galeev et al .
1979 ) , magnetic flux tubes are buoyant in a stratified atmosphere , and so they rise to the surface of the accretion disk .
as the tube is rising , the particles slide along the magnetic field lines downward to the disk in response to gravity .
the magnetic flux tube becomes more and more particle - free , @xmath31 is increasing , and so the conditions become more and more favorable for the dissipation of magnetic field energy . if magnetohydrodynamic waves are generated and propagate up to the top of the flux tube , and if the conditions there for effective reconnection are satisfied , the different ( incoherent ) waves reconnect and thus produce highly energetic particles .
the particles in turn produce x - radiation by up - scattering the uv radiation from the disk .
since the radiation pressure @xmath42 is very much smaller than the stress in the underlying magnetic flux tube , we may neglect the sideways expansion of the flux tube .
we need to consider the pressure equilibrium along the magnetic field lines , however , since the plasma can in principle move freely in that direction .
the balance of radiation pressure with the magnetic ram pressure then sustains the ar optical depth as discussed above .
since the flux tube is geometrically thick , the corresponding ratio @xmath43 is probably of order @xmath44 one to a few , and with @xmath45 , we therefore expect @xmath46 the lowest values of the equilibrium @xmath1 can be reached due to the fact that @xmath47 in this equation is not necessarily the total area of the source , because some of the x - ray flux can be reflected by the underlying disk and re - enter the ar
. some of this re - entering flux can be parallel to the incoming alfvn energy flux , and thus the effective @xmath47 is smaller than the full geometrical area of the source .
furthermore , we have assumed a one - temperature gas , and have neglected the gas pressure in our calculation .
it is possible that the protons are much hotter than the electrons , and that they account for a sizable fraction of the total pressure in the ar , which then leads to a reduction in the value of the equilibrium @xmath1 as compared to equation ( 5 ) . in the analysis developed here ,
there is nothing specific to alfvn waves .
we could have instead invoked the influx of energy by other waves or even energetic particles accelerated by a magnetic reconnection process .
all that matters is that they have a well - defined relationship between their momentum and energy densities , and that their propagation speed is close to @xmath0 .
therefore , there may occur a different physical setting where the mechanism for the pressure equilibrium described here can be at work .
it seems unavoidable that magnetic fields are involved , and that they therefore provide some of the confinement , e.g. , in a direction perpendicular to the magnetic field lines .
we encourage future work on the range of possible geometries for the ars .
zdziarski et al . ( 1996 ) produced a fit of the average _
ginga_/osse spectrum of seyfert 1 galaxies assuming that the active regions form hemispheres above the disk .
they found that the radial optical depth of the hemispheres is @xmath2 . here
, we will examine whether the seyfert spectrum can be due to a combination of spectral components from flares with different @xmath1 , but the same @xmath17-parameter ( set arbitrarily at @xmath48 ) .
the latter assumption is introduced to ensure that the x - ray spectral index does not vary considerably from flare to flare .
a constant @xmath17-parameter is a natural consequence of the fixed geometry of the flare , in the sense that the cooling of the plasma is fixed by how much of the x - ray flux re - enters the emitting region after it is reflected from the disk ( see haardt & maraschi 1991 ) . as an illustration of the method ,
we first compute the spectrum from flares with a range of thomson optical depths assuming that they all have the same luminosity .
we then convolve these spectra with a gaussian probability distribution that a flare occurs with @xmath1 .
the composite spectrum @xmath49 ( in energy / sec / kev ) is @xmath50 f(e , \tau_t)\,{\rm , } \ ] ] where @xmath51 is the spectrum from a single flare with @xmath1 .
we take @xmath52 and adopt several values of @xmath53 to represent the possible spread in @xmath1 between different flares .
the individual spectra are computed assuming a slab geometry using an eddington frequency - dependent approximation for the radiative transfer , using both the isotropic and first moments of the exact klein - nishina cross section ( nagirner & poutanen 1994 ) .
although this geometry is clearly different from that of a realistic flare , our point here is to test the possibility of co - adding spectra with different @xmath1 , in order to see what range in @xmath1 may be permitted by current observations .
we expect that a more accurate calculation with the correct geometry will yield qualitatively similar limits on @xmath1 , though the exact values should be inferred using a @xmath54 fit to the data .
figure 1 shows the results of our calculation for @xmath52 and two values of @xmath55 : @xmath56 and @xmath57 . the spectrum from a single flare ( solid curve ) with @xmath58 is also shown for comparison .
it can be seen that the plot for @xmath59 is hardly distinguishable from that for @xmath58 .
moreover , these curves differ the most above @xmath60 kev , where the osse data typically have error bars larger than this difference ( see , e.g. , fig . 1 in zdziarski et al .
on the basis of this simple test , we would expect that seyfert spectra may be comprised of contributions from many ars encompassing a range ( @xmath61 ) of @xmath1 .
we note , however , that a broader range in @xmath1 is unlikely because of the considerable flattening to the spectrum for @xmath62 .
the conclusion that @xmath1 is allowed to vary within the range of @xmath61 is very important for the magnetic flare model , since it is otherwise difficult to see how different flares could produce exactly the same @xmath1 . it may also happen that a flare evolves through many phases and that its thomson optical depth therefore varies with time .
however , these calculations demonstrate that as long as that variation is restricted to the range @xmath63 , the resulting spectrum is consistent with the observations .
this is not to say that flares with @xmath1 greatly in excess of @xmath25 are not permitted . according to equation ( 5 )
, a large optical depth probably corresponds to a flare with @xmath64 .
the magnetic field is therefore probably smaller and the energy is transported with a smaller @xmath31 .
that is , magnetic flares with large @xmath1 may be significantly dimmer than the more energetic ones with @xmath2 , and so they do not contribute noticeably to the observed seyfert spectrum .
we have considered the consequences of imposing a pressure equilibrium on the active regions of seyfert galaxies , in addition to the more often studied energy equilibrium , under the assumption that the emission arises within energetic magnetic flares above the surface of a cold disk .
we showed that if the energy is supplied to the x - radiating plasma by the influx of some energy source with a group velocity @xmath65 , then @xmath1 probably falls within the range @xmath66 .
the current x - ray/@xmath3-ray observations are consistent with this range of thomson optical depths .
we conclude that magnetic flares on the surface of the cold disk remain a viable explanation for the spectra observed in seyfert galaxies .
alternative explanations , based on a gravitational confinement of the ars , appear to be incapable of accounting for the observed ` universality ' in the value of @xmath1 .
this work was partially supported by nasa grant nag 5 - 3075 .
we are very grateful to the anonymous referee for pointing out a serious error in our original manuscript .
nayakshin & melia 1997a , submitted to apj nayakshin & melia 1997b , apjl , in press ( also astro - ph/9705010 ) parker , e.n .
1979 , cosmical magnetic fields , clarendon press , oxford .
nagirner , d. j. , & poutanen , j , 1994 , astrophys .
space phys . ,
* 9 * , 1 rybicki , g. b. , & lightman , a.p . , 1979 , radiative processes in astrophysics , john wiley and sons : new york .
shakura & sunyaev 1973 , a&a , 24 , 337 * figure 1 . * the spectrum resulting from co - adding the components due to different magnetic flares with a gaussian distribution in the thomson optical depth ( see text ) , centered on @xmath52 with a width @xmath67 ( long - dashed curve ) and @xmath57 ( short - dashed curve ) . the spectrum from a single flare with optical depth @xmath68
is shown as a solid curve .
the curves are normalized to the same integrated flux .
these spectra demonstrate that current observations may not be able to differentiate between a single - flare spectrum and one comprised of many different flares if their optical depth is in the range @xmath69 . | we here consider the pressure equilibrium during an intense magnetic flare above the surface of a cold accretion disk . under the assumption that the heating source for the plasma trapped within the flaring region is an influx of energy transported inwards with a group velocity close to @xmath0 , e.g. , by magnetohydrodynamic waves
, this pressure equilibrium can constrain the thomson optical depth @xmath1 to be of order unity .
we suggest that this may be the reason why @xmath2 in seyfert galaxies .
we also consider whether current data can distinguish between the spectrum produced by a single x - ray emitting region with @xmath2 and that formed by many different flares spanning a range of @xmath1 .
we find that the current observations do not yet have the required energy resolution to permit such a differentiation .
thus , it is possible that the entire x - ray/@xmath3-ray spectrum of seyfert galaxies is produced by many independent magnetic flares with an optical depth @xmath4 .
#1#2#3#1 , a&a , # 2 , # 3 # 1#2#3#1 , a&as , # 2 , # 3 # 1#2#3#1 , aj , # 2 , # 3 # 1#2#3#1 , _ ap .
j. _ , * # 2 * , # 3 .
# 1#2#3#1 , _ ap .
j. ( letters ) _ , * # 2 * , # 3 .
# 1#2#3#1 , apjs , # 2 , # 3 # 1#2#3#1 , ara&a , # 2 , # 3 # 1#2#3#1 , baas , # 2 , # 3 # 1#2#3#1 , icarus , # 2 , # 3 # 1#2#3#1 ,
_ m.n.r.a.s .
_ , * # 2 * , # 3 .
# 1#2#3#1 , _ nature _ , * # 2 * , # 3 .
# 1#2#3#1 , pasj , # 2 , # 3 # 1#2#3#1 , pasp , # 2 , # 3 # 1#2#3#1 , qjras , # 2 , # 3 # 1#2#3#1 , science , # 2 , # 3 # 1#2#3#1 , soviet astr . ,
# 2 , # 3 # 1#2#3#1 , soviet astr .
lett .
, # 2 , # 3 # 1#2#3#4#1 , # 2 , # 3 , # 4 # 1#1 # 1#2 submitted to the editor of the astrophysical journal letters revised may 28 , 1997 0.3 in |
the abc effect , a mesonic structure of isospin @xmath14 , was first observed by abashian , booth and crow in the reaction @xmath15he@xmath16 @xcite .
later on , it was also shown to appear in several other hadronic reactions , like e.g. in @xmath0h@xmath17 @xcite and in @xmath18h@xmath19h@xmath20he@xmath17 @xcite . in all these cases ,
the abc structure shows up as a cross section peak at missing masses of about 300 350 mev , and with a position and width that varies quite rapidly according to the kinematic conditions .
the only possible explanation for the effect seems to be an enhancement in the double pion production which is caused by a strongly energy dependent production amplitude .
since the abc effect is present in reactions of different type , it seems to be quite natural to assume that there exists a common underlying mechanism for the @xmath5 production which should be able to explain all the different experiments . from a theoretical point of view
, @xmath0h@xmath21 has to be the reaction to analyze first since it involves only two - nucleon states beside the pions and can thus be treated in the most accurate and direct way . for this reaction
, the abc enhancement corresponds to a beam momentum of approximately 2 gev / c in the laboratory system and to a center of mass energy of @xmath22 , i.e. twice the mass of the @xmath23 resonance .
this fact strongly suggests that the two pions are dominantly produced via a double @xmath24 excitation .
indeed , as it has already been shown by bar nir et al .
@xcite , the main features of the experimental spectrum can be understood by assuming the @xmath6 mechanism to be the most efficient one for @xmath25 production in the energy region under consideration . in the present work ,
we will present a new calculation of @xmath6 and @xmath7 excitations in the @xmath0h@xmath21 reaction which explicitly includes the effects of residual interactions in the intermediate resonance states @xcite .
our aim is to show that the @xmath5 production may serve as an interesting tool for further examination of direct @xmath6 and @xmath7 potentials which are not very well known . for this purpose
, we will also examine how the spin structure of the @xmath6 excitation influences the angular distribution of the two pions .
our model is based on a coupled channel approach in a non relativistic framework .
the potentials of the residual interactions are adopted from a meson exchange model @xcite . in this work ,
contributions from @xmath10 , @xmath11 , @xmath12 , and @xmath13 exchange are included .
the @xmath24 resonance is treated thereby as a quasi particle with a given mass and an energy
dependent , intrinsic width .
experimental results for the @xmath0h@xmath1 cross section in the abc energy region have been obtained by plouin et al .
@xcite at @xmath2 gev . in order to allow for a direct comparison with our theoretical results ,
the contribution from one pion production has been subtracted from the spectra .
it has been stressed that there might be some additional background from @xmath26 production @xcite due to momentum spread of the beam , but since there is no microscopic model for this contribution available , we decided not to perform any further corrections .
one should keep in mind , however , that not all of the uncertainties have been included in the experimental error bars . in section [ sec:1 ]
, we will present the theoretical framework of our model .
the results of our calculation are presented and discussed in section [ sec:2 ] .
the paper concludes with a summary in section [ sec:3 ] .
we are interested here in the calculation of the deuteron spectrum for the @xmath0h@xmath1 reaction , i.e. in the double differential cross section @xmath27 . using relativistic kinematics ,
the cross section is given as @xmath28 where @xmath29 denotes the relative momentum of the two pions and @xmath30 is the sum of their energies . the indices @xmath31 and @xmath32 refer to the neutron projectile ( proton target ) and the deuteron ejectile , respectively .
the center of mass momentum of the @xmath5 system @xmath33 is fixed by momentum conservation and equals @xmath34 . as usual , the function @xmath35 is defined to be @xmath36 \left [ s-(m_p - m_n)^2 \right]$ ] .
finally , @xmath37 denotes the transition amplitude for the reaction ; in the case of the unpolarized cross section , an average over the initial and a sum over the final spin orientations has to be performed . from eq
( [ eq:1 ] ) one easily obtains the deuteron spectrum @xmath38 since we are going to calculate the transititon matrixelement in a non relativistic approach , the above expression has to be evaluated in the deuteron rest frame ( drf ) where the deuteron is well described by the usual non relativistic wave function @xcite .
afterwards , the transformation of the spectrum into the laboratory frame ( lab ) is performed by using the simple relation @xmath39 in our microscopic model for the @xmath0h@xmath1 reaction , we take into account two different reaction mechanisms corresponding to the feynman diagrams presented in fig . [
fig : diagrams ] .
the first mechanism ( a ) involves the excitation of a double delta ( @xmath6 ) intermediate state where both @xmath24 resonances subsequently decay into a pion and a nucleon . in the case of the second mechanism ( b ) , each of the two pions couples to the same @xmath24 , hence only one resonance (
i.e. a @xmath7 system ) is excited .
for the explicit calculation of the two matrix elements , wa take advantage of the principle of detailed balance .
this allows us to assume that the deuteron is in the initial state and absorbes the two pions , which finally will lead to a @xmath40 state .
the absorption of only one pion with momentum @xmath41 is described by the operator @xmath42 the exponential term represents the plane wave of the pion field .
the structure of the @xmath43 coupling follows from the usual interaction lagrangian @xcite in the non relativistic reduction .
@xmath44 ( @xmath45 ) is the spin ( isospin ) transition matrix @xcite for the @xmath24 excitation , and @xmath46 is the relative coordinate of the two nucleons in the deuteron .
the symbol @xmath47 in eq .
( [ eq:4 ] ) shall indicate that the operator has to be properly symmetrized with respect to the nucleonic coordinates , i.e. each of the two nucleons in the deuteron may be excited to a @xmath24 resonance . following refs .
@xcite , the quantum mechanical state @xmath48 is called the uncorrelated source function for the @xmath7 intermediate system , with @xmath49 being the deuteron wave function .
after propagation of the @xmath7 system and absorption of a second pion , one obtains the source function for the @xmath6 state , @xmath50 in our model , the full propagator @xmath51 is given by @xmath52 it contains the excitation energy @xmath53 , the energy
dependent width @xmath54 of the resonance @xcite , the operator of the kinetic energy @xmath55 and the interaction potential @xmath9 of the intermediate @xmath7 system . @xmath53 and
@xmath56 are fixed by imposing energy conservation at the @xmath57 vertex .
the potential @xmath9 is constructed in a meson exchange model with @xmath10 , @xmath11 , @xmath12 and @xmath13 exchange taken into account . a diagrammatic representation of the residual @xmath7 interaction is given in fig . [ fig : potential ] ( a , b ) . for more details , we refer the reader to refs .
@xcite where also explicit expressions for @xmath58 can be found . in eq .
( [ eq:6 ] ) we have not yet considered that the @xmath6 source function has to be symmetric under exchange of the pions @xmath59 and @xmath60 . according to the value @xmath61 of total isospin in the @xmath62 system , we define the symmetrized operator @xmath63 the complete matrix element for the @xmath6 mechanism can now be written as @xmath64 here , @xmath65 denote the distorted waves of the projectile neutron ( target proton ) which are calculated in eikonal approximation . in analogy to eq .
( [ eq:7 ] ) , the full propagator for the @xmath6 system is given by @xmath66 - t_\dd - v_\dd } \ : .\ ] ] especially , this propagator includes the residual interaction @xmath8 in the excited @xmath6 system , see also fig .
[ fig : potential](c ) .
@xmath8 , as well as the transition potential @xmath67 between the intermediate @xmath6 system and the @xmath68 state , are both constructed within the same meson exchange model as @xmath9 . to the @xmath69 transition potential , of course , only the @xmath10 and @xmath11 meson exchanges contribute .
the @xmath6 excitation can thus be separated into a spin longitudinal part ( with @xmath10like coupling to @xmath68 ) and a spin transverse part ( with @xmath11like coupling to @xmath68 ) . as we will demonstrate later , this spin structure of the @xmath6 excitation is important for the angular distribution of the pions . the matrix element corresponding to the second mechanism where only one @xmath24 is excited , fig .
[ fig : diagrams](b ) , is found to be @xmath70 here , the operator @xmath71 is obtained from @xmath72 by simply replacing for the second pion @xmath44 and @xmath45 with their hermitian adjungates @xmath73 and @xmath74 , respectively . in order to obtain the @xmath5 production amplitude , the coherent sum of the two contributions of eqs .
( [ eq:9 ] , [ eq:10 ] ) has to be taken . due to the fact that the residual interactions @xmath9 and @xmath8 couple states of different quantum numbers , the concrete evaluation of the matrix elements leads to a system of coupled integro
differential equations .
it turns out that they can be solved in a very effective way with the so called lanczos method @xcite .
this procedure has already been described in ref .
@xcite and can easily be applied to the present calculation .
input parameters of the model are the meson and baryon masses , coupling constants , and formfactor cutoffs at each vertex .
the values we used in our calculations are given in table [ tab : parameters ] .
most of these quantities , of course , are physical observables that are already determined from other experiments , e.g. from @xmath68 scattering @xcite .
only the sigma meson and cutoff parameters remain to be fixed .
they have been adjusted to fit the experimental data for the one pion production @xmath75h@xmath76 .
we made sure that not only the total cross section but also the angular distributions and analyzing powers are correctly reproduced in the @xmath24 resonance energy region @xcite .
after this consistency check , we may now continue and discuss our results for the @xmath5 production @xmath0h@xmath1 . in fig .
[ fig : expspect ] , we present the experimental deuteron spectrum @xmath27 as it is measured in the reaction @xmath77 at a beam momentum @xmath2 gev and in forward direction @xmath3 .
the production of both neutral ( @xmath78 ) and charged ( @xmath79 ) pion pairs is possible .
the available phase space for the @xmath5 production is indicated by the dashed line .
obviously , the cross section does not follow the phase space but shows a characteristic structure which three pronounced peaks .
the two outer maxima correspond to a kinematical situation where the invariant mass @xmath80 of the @xmath5 system is minimal , i.e. where @xmath81 . the broad central maximum , on the other hand , is located at the position where @xmath82 has its maximal value ( 511 mev in the present case ) . in the cms , this situation will be realized if all the kinetic energy is taken by the two pions and the deuteron remains at rest . as can be seen from fig .
[ fig : theospect ] , our model is able to explain the characteristic structure of the experimental spectrum .
the solid line represents the full result with all intermediate interactions included . the @xmath6 mechanism , for which the feynman diagram was presented in fig . [ fig : diagrams](a ) , is in fact the only important contribution to @xmath5 production in the energy range under consideration .
the cross section contribution due to the @xmath7 mechanism of fig .
[ fig : diagrams](b ) is almost negligible .
since the optimal energy for this second mechanism would be @xmath83 , its suppression in the abc energy region is not surprising .
with focus on the @xmath6 excitation and following the arguments of ref .
@xcite , the origin of the different maxima can be understood quite easily .
let us define @xmath84 where @xmath85 denote the four momenta of the two pions .
the excitation of a @xmath6 system is most effective if the invariant masses of the two @xmath24 s are approximately equal ( and hence both close to the resonant mass ) . in this case
we have @xmath86 and since @xmath87 this is equivalent to @xmath88 this condition can be fullfilled in two ways : 1 .
@xmath89 and therefore @xmath90 , or 2 .
@xmath91 in the cms of the two pions ( which means that the deuteron restframe and the @xmath5 restframe are identical ) . in the latter case ,
the two pions carry the whole kinetic energy , and @xmath92 max .
the first situation leads to the outer maxima and corresponds to the parallel decay of the two @xmath24 s because of @xmath93 , i.e. the relative momentum is very small .
the second situation explains the central maximum and corresponds to the antiparallel decay of the @xmath6 excitation for which the relative momentum is maximal . in the deuteron restframe
, the different kinematical configurations for the three maxima can be visualized as depicted in fig .
[ fig : kinematics ] .
the kinematical situations leading to the three maxima of the deuteron spectrum are also clearly visible in the angular distribution of the pions . in fig .
[ fig : angdist1 ] , we show the triple differential cross section @xmath94 for different deuteron laboratory momenta .
it is plotted as a function of @xmath95 which is the angle between the relative momentum @xmath96 of the two pions and the beam axis , as given in the deuteron rest frame ( see also fig . [
fig : kinematics ] ) . for the two outer maxima ( @xmath97 gev and @xmath98 gev ) ,
the pion momenta are nearly parallel and hence the relative momentum is dominantly perpendicular to the beam axis .
consequently , the differential cross section reaches its largest value at @xmath99 . for the central maximum ( @xmath100 gev ) , however , the pion momenta are antiparallel , and all angles @xmath95 are kinematically possible .
if there was no spin dependence of the excitation and the residual interaction , an isotropic distribution would be the result .
the observed angular variation of the cross section is thus reflecting the spin structure of the process which mainly follows from the @xmath10 and @xmath11 exchange contributions to the @xmath101 transition potential . in order to examine this spin structure in more detail
, we will neglect for the moment the residual interaction and the possibility of spin flips .
then , the spin longitudinal @xmath10 exchange leads to an operator proportional to @xmath102 \ ; \left [ ( \vec s_2 \cdot \vec q \ , ) \ , ( \vec s_2^\dagger \cdot \vec k_b ) \right ] \ , \sim \ , ( \vec q \cdot \vec k_a \ , ) \ , ( \vec q \cdot \vec k_b \ , ) \ , ,
] ] where @xmath103 are the pion momenta and @xmath104 is the momentum transfer from the neutron projectile to the proton target . if both the pion momenta are parallel ( or antiparallel ) to @xmath104 , the corresponding cross section will be largest .
this results in a maximum at @xmath105 in the spin longitudinal channel . in complete analogy , the spin
transverse @xmath11 exchange has the operator structure @xmath106 \ ; \left [ ( \vec s_2 \times \vec q \ , ) \ ,
( \vec s_2^\dagger \cdot \vec k_b ) \right ] \ , \sim \ , ( \vec q \times \vec k_a ) \ , ( \vec q \times \vec k_b ) \ , .\ ] ] therefore , the maximum in the spin
transverse channel is reached if both the pion momenta are perpendicular to @xmath104 , i.e. if @xmath99 . indeed
, the theoretical curves in fig . [ fig : angdist2 ] exactly reflect the expected shapes .
we conclude that a measurement of the pion angular distribution would be very helpful in order to reveal the spin structure of the interaction .
effects of residual interactions can be examined in both the @xmath6 and the @xmath7 system since these are the intermediate configurations in our model .
the influence of the corresponding interaction potentials on the theoretical spectrum is shown in fig .
[ fig : vinfl ] for the case of forward scattering ( @xmath3 ) .
the dash dotted line was calculated without any residual interaction , i.e. with @xmath107 for the intermediate states .
the typical abc structure with its three peaks as discussed in section [ sec : spectrum ] is clearly present .
our result comes quite close to the fully relativistic calculation of bar nir et al .
@xcite which also does not include the residual interactions . by choosing @xmath108 but still neglecting the direct @xmath6 interaction , we obtain the dashed line . here , the maxima are even more pronounced .
the enlargement of the peak cross sections is due to the reduction of the excitation energy for the @xmath7 system which is caused by the attractive @xmath7 potential . for @xmath109 as in the present case ,
the whole system is thus closer to the resonance energy .
the relative position of the two invariant @xmath24 masses remains however unchanged since in our model they are fixed independently of @xmath58 .
on the other hand , the @xmath6 potential can directly influence the relative wave function of the @xmath6 system via a redistribution of momentum within the two particle system .
the solid line in fig .
[ fig : vinfl ] has been calculated with the full residual interaction @xmath110 .
the inclusion of @xmath8 results in a transfer of strength from the maxima to the regions between which are kinematically less favoured .
this transfer takes place because the optimal configuration of equal @xmath24 masses can now be reached even if the initial energy distribution of the two pions was asymmetric . as a consequence , the cross section is enhanced between the peaks and simultanously the magnitude of the peaks is reduced .
this reduction is more prominent for the outer maxima since they are related to a small relative momentum of the @xmath6 system . for the central maximum ,
the relative momentum is large and therefore only a slight influence of the @xmath6 potential is observed . obviously , the residual interactions play quite an important role in the @xmath5 production and significantly influence the deuteron spectrum .
the inclusion of these effects in the calculation clearly improves the agreement between theory and experiment .
studying the dependence of the deuteron spectrum on the scattering angle @xmath111 , we find that the theoretical cross section for @xmath6 excitation is decreasing too fast .
exemplatory we present in fig .
[ fig : anguldep ] our calculations for @xmath112 and @xmath113 .
the experimental data @xcite are underestimated by a factor @xmath114 .
other microscopic calculations , e.g. assuming a two nucleon exchange as the dominant reaction mechanism , yield a comparable angular dependence and are also not able to describe the data @xcite .
the reason for this behaviour is not understood .
maybe more than only one production mechanism has to be taken into account in order to solve this problem .
[ fig : energdep ] is demonstrating the energy dependence of the deuteron spectrum in our model .
one recognizes that the central maximum is getting more pronounced if the total energy increases .
we remind the reader that this peak corresponds to the @xmath6 excitation with maximal relative momentum , which already gives the natural explanation of the observed effect . after integration over deuteron momentum and scattering angle ,
the total cross section for the @xmath115 reaction is obtained . in fig .
[ fig : totcs ] , our theoretical result is compared to the experimental data @xcite . in the @xmath6 resonance energy region ,
i.e. for a neutron momentum of @xmath116 gev , the agreement is quite good .
this confirms our assumption that the @xmath5 production in the regime of the abc effect is dominated by the @xmath6 excitation mechanism . on the other hand
, one recognizes that the experimental cross section close to the @xmath5 production threshold ( @xmath117 gev ) is understimated by nearly two orders of magnitude . in this case , the total energy of @xmath118 is simply not sufficient for the excitation of a @xmath6 system , and other production mechanisms will be more important . to mention just two possibilities ,
the pions could be produced by s
wave rescattering or via excitation of the @xmath119 resonance , as recently discussed in @xcite .
since there is also experimental evidence @xcite that the abc structure completely disappears when approaching the @xmath25 production threshold , further theoretical investigations of the low energy region would be of high interest .
to summarize we have shown that the @xmath4 excitation is the dominant reaction mechanism in the @xmath0h@xmath1 two pion production at @xmath2 gev / c .
it is able to explain the abc effect as observed in the experimental deuteron spectrum .
hereby , the residual interactions in the intermediate @xmath6 and @xmath7 states play an important role .
we obtain good agreement with experimental data at forward scattering but too fast a decrease of the cross section for higher scattering angles of the deuteron .
the total cross section is also underestimated close to the @xmath25 production threshold where other reaction mechanisms get more important .
furthermore , we found that the spin structure of the @xmath6 excitation directly influences the angular distribution of the two pions .
we conclude that the @xmath25 production in the abc energy region may well serve as a tool for closer examination of direct @xmath6 interaction and transition potentials .
this paper is dedicated to the memory of prof .
dr . franz osterfeld who deceased in 1997 .
we would like to thank j. speth , ch .
hanhart , o. krehl and c. wilkin for many helpful discussions .
the work was supported in part by the studienstiftung des deutschen volkes . | the deuteron spectrum in @xmath0h@xmath1 at @xmath2 gev / c and @xmath3 is explained by considering a @xmath4 excitation as the dominant reaction mechanism for the @xmath5 production .
we present a new theoretical approach based on a coupled channel formalism which allows to include the residual interaction within the intermediate @xmath6 and @xmath7 systems .
the corresponding interaction potentials @xmath8 and @xmath9 are adopted from a meson exchange model with @xmath10 , @xmath11 , @xmath12 , and @xmath13 exchange taken into account .
the influence of the residual interaction on the deuteron spectrum is studied .
we also predict the angular distribution of the two pions .
it is shown that this distribution is closely connected to the spin structure of the @xmath6 excitation . |
suppose that @xmath19 is an @xmath1 matrix with entries in @xmath20 , each of whose rows contains at least one @xmath21 , for which we seek a left null vector over gf@xmath22 $ ] , i.e. a row vector @xmath23 such that @xmath24 ( mod 2 ) , where here and elsewhere @xmath25 is the all-@xmath26 vector .
more generally , elements of @xmath0 might belong to the finite field gf[@xmath27 of order @xmath28 .
we are interested in the case where @xmath0 is sparse and random , as specified below .
let @xmath29 denote the vectors constituting the rows of @xmath0 , and let @xmath30 denote the co - rank over gf@xmath22 $ ] , namely @xmath31 where here and subsequently ` @xmath32 ' indicates the linear span over gf[2 ] .
then the number of null vectors of @xmath0 , including the zero vector , is @xmath33 which counts the number of distinct solutions in @xmath34 , including the zero solution , to @xmath35 note that for a fixed @xmath6 and a given realization of the sequence of rows @xmath36 , the numbers @xmath2 are nondecreasing as @xmath37 increases .
suppose that @xmath4 , with @xmath5 .
our goal is to examine the limiting behaviour of the expected number @xmath38 $ ] of left null vectors , and the limiting probability @xmath39 $ ] of a mod-@xmath40 linear dependency of the rows of @xmath41 , as a function of the parameter @xmath11 , and especially to derive computable thresholds at which phase transitions occur .
we also study the rate of exponential decay of the probability that @xmath42 is a null vector .
the probabilistic setting that we consider has the rows @xmath29 being independent and identically distributed ( i.i.d . ) with the law of a random vector @xmath43 .
the problem has different flavours depending on the underlying law of @xmath44 , and several regimes have received considerable attention in the literature , including : * the _ dense _ regime in which @xmath44 has order @xmath6 non - zero components ; the standard model studied in this regime has @xmath44 distributed uniformly over @xmath45 . * the classical _ sparse _ regime in which @xmath44 has order @xmath46 non - zero components . * the uniformly ( very ) sparse regime in which @xmath44 has @xmath47 non - zero components .
the main focus of the present paper is regime ( c ) ( albeit our ` @xmath47 ' may be random for each row , and might not even have a mean ) ; in section [ sec : other ] below we briefly discuss other models that have been studied . in the simplest case , @xmath44 contains a fixed number @xmath48 of non - zero components ( the cases @xmath49 and @xmath50 have distinct behaviour from the case @xmath51 ) ; in more generality the number of non - zero components is randomly distributed according to a given _
weight distribution_. before formally describing our model in detail and presenting our main results ( in section [ sec : results ] ) , we make some remarks on motivation , and on the literature .
note that @xmath52 if and only if @xmath0 has row rank @xmath37 , which occurs if and only if @xmath0 has column rank @xmath37 .
thus the absence of non - trivial left null vectors is equivalent to all column vectors in @xmath3 being expressible as a linear combination of the columns of @xmath0 ( with addition modulo 2 ) , or in other words , to there being a solution @xmath53 to @xmath54 for all column vectors @xmath55 . in the special case of @xmath50 , motivation for considering this question
is discussed at the start of ( * ? ? ?
* chapter 3 ) .
the following interpretations help to motivate the general case .
+ _ a scheduling problem .
_ suppose that a tennis club is organizing its annual schedule .
there are @xmath6 playing days , and @xmath37 potential players .
each player wants to play on a given subset of the days ; if there is not a match available on every one of these days , they refuse to pay the annual membership .
each day , in order for nobody to be left out , an even number of players is required .
each possible schedule satisfying these requirements is a left null vector mod @xmath40 ; the one with the most units achieves the maximal income for the tennis club . + _ randomized lights out .
_ this is a variant of the game ` lights out ' @xcite .
each of @xmath37 lamps can be either on or off , and there are @xmath6 switches , each of which is incident to a specified subset of the lamps , as given by the random matrix @xmath0 ; lamp @xmath56 and switch @xmath57 are mutually incident if and only if the entry at @xmath58 of @xmath0 is 1 .
if a switch is toggled , all of the lamps incident to it have their status changed from on to off or off to on .
one may ask whether all states ( i.e. configurations of on and off lamps ) are accessible from the ` all off ' state by using some sequence of switches ( or equivalently , if the ` all off ' state is accessible from all possible starting states ) , and this is equivalent to the question of whether the column rank of @xmath0 is @xmath37 .
+ _ the xorsat problem .
_ this is a variant of the random satisfiability problem @xcite , where there are @xmath6 boolean variables which may be deemed true or false .
each row of @xmath0 represents a clause built as the logical xor ( exclusive or ) involving those boolean variables corresponding to columns incident to this row , so the clause is true if an odd number of the variables incident to the row are deemed true . given a vector @xmath55 , finding a solution @xmath59 to @xmath54 corresponds to finding a truth - assignment for the boolean variables so that each clause @xmath56 is true if @xmath60 and false if @xmath61 .
thus the column rank is @xmath37 if and only if the problem is satisfiable for all possible choices of @xmath62 .
+ _ a spin - glass model . _
the relationship between satisfiability problems and spin glasses has already been noted in @xcite . in the present instance , consider the following variant of the well - known sherington
kirkpatrick mean - field spin - glass model ( see e.g. @xcite ) .
there is a random collection of hyperedges on @xmath6 vertices , represented by the @xmath37 rows of @xmath0 .
each hyperedge @xmath56 has a sign @xmath63 , taking value @xmath64 .
each vertex @xmath57 is assigned a spin @xmath65 taking values in @xmath66 , and ( at zero temperature ) the probability measure on the state - space is concentrated on states of minimal energy , i.e. with maximal value of @xmath67 , where here @xmath68 denotes the product of spins at vertices in hyperedge @xmath56 .
the existence of a configuration with all terms in the sum equal to @xmath69 is equivalent to the existence of a solution to @xmath54 .
+ _ the ehrenfest urn model and the random walk on the hypercube . _ in the ehrenfest model of heat exchange , a box contains @xmath6 particles , some of which are red and the rest blue . at each step
, a particle is sampled uniformly from the box and changes its colour . for a sample of the large literature , see e.g. @xcite , @xcite , @xcite , or @xcite . in the case where @xmath44 has a single unit entry , we may view each row of @xmath0 as selecting which particle is to be changed at that step .
then @xmath70 is a null vector for @xmath0 if and only if the model returns to the initial state after @xmath37 steps .
this may also be interpreted as a random walk on the graph whose vertices are @xmath45 and edges are present between those vertices that differ in a single component ; the event that @xmath70 is null corresponds to the walker being back in his starting state after @xmath37 steps .
the general case , allowing other weight distributions , corresponds to a generalization of the ehrenfest model whereby multiple ` diffusions ' are allowed , i.e. at each step several particles may change colour at once ; cf ( * ? ? ?
* chapter 10 ) . this can be similarly interpreted in terms of a walk on a version of the hypercube with additional edges .
+ there is a large body of work on the properties of random matrices over finite fields and the closely related subject of random linear equations over finite fields .
surveys are provided by the book ( * ? ? ?
* chapter 3 ) as well as the articles @xcite .
the problems may also be formulated in terms of _ random hypergraphs _ : each row represents a hyperedge , and each column represents a vertex ; for details see section [ sec : cores ] below .
they are also related to the _ xorsat _ problem in boolean algebra , as mentioned above ( see also section [ between ] below ) .
generally , such models can be described in the framework of _ random allocation _ or _
occupancy _ problems : see the books @xcite .
the null - vector problem in the fixed row - weight case has received several treatments in the literature , and it is not easy to reconcile all of the existing results , due to differences in presentation and also differences in the underlying probabilistic models . one contribution of the present paper is to clarify some of these issues , including giving a rigorous justification that the results are unchanged under small perturbations of the underlying model .
our main contribution , however , is to treat the case of genuinely _ random _ row weights , which has not previously been studied .
we mention that there has recently been renewed interest in this area in several scientific communities : for example , alamino and saad @xcite give a statistical physics approach to the null - vector problem ; ibrahimi _ et al . _
@xcite treat the related problem of random xorsat ; costello and vu @xcite study the rank of random symmetric ( so in particular , square ) matrices . throughout the paper ,
we extend the function @xmath71 , @xmath72 , continuously to @xmath73 , so that @xmath74 .
we define the _ weight _ of a vector @xmath75 to be @xmath76 , i.e. , the number of unit entries . for @xmath77
we write @xmath78:= \{1,2,\ldots , n\}$ ] .
we write @xmath79 for convergence in distribution .
given @xmath80 , suppose that @xmath81 is a random row vector , selected according to some probability law on @xmath45 .
consider a sequence of i.i.d .
random vectors @xmath36 with the same law as @xmath44 .
let @xmath82 be the @xmath1 matrix whose rows are @xmath29 .
we will consider @xmath44 with law of the following form .
let @xmath8 be an @xmath83-valued random variable ( so @xmath84 = 1 $ ] ) whose law will be the ( limiting ) _ weight distribution _ of our random vector @xmath44 .
let @xmath85 be a sequence of random variables with @xmath86 $ ] such that @xmath87 as @xmath88 .
let @xmath89 have the distribution of @xmath90 , and for each @xmath91 $ ] let the conditional distribution of @xmath44 , given @xmath92 , be uniform over @xmath93 .
let @xmath94 $ ] and @xmath95 $ ] denote the probability generating functions of @xmath8 and @xmath90 , respectively .
we use @xmath96 and @xmath97 for the probability and expectation for the random matrix model with @xmath6 columns and row weight distribution given by @xmath98 .
we shall say @xmath90 are _ uniformly bounded _ if there is a finite constant @xmath99 such that @xmath100 = 1 $ ] for all @xmath6 ( and hence @xmath101 = 1 $ ] as well ) .
we state some fundamental properties of @xmath105 and @xmath106 in the next result , which we prove in section [ secprelim ] .
[ alphaprop ] we have @xmath107 for @xmath108 but @xmath109 for @xmath110 , and @xmath111 is continuous and nondecreasing as a function of @xmath11 .
moreover : * @xmath112 $ ] ; and @xmath113 if @xmath114<\infty$ ] . *
@xmath115 if and only if @xmath116 > 0 $ ] . * if @xmath117 = 1 $ ] , then @xmath118 . *
suppose that @xmath119 is another @xmath83-valued random variable , with @xmath120 $ ] , such that @xmath121 for all @xmath122 $ ] ( which is the case , for example , if @xmath119 stochastically dominates @xmath8 )
. then @xmath123 . in particular , if @xmath124 = 1 $ ] and @xmath114<\infty$ ] , then @xmath125 .
here is our first main result , describing the threshold behaviour of the expected number of null vectors .
we shall prove this in section [ mainproofs ] .
[ thm1 ] suppose that @xmath126 as @xmath88 . then @xmath127 = f_\rho ( \alpha ) ; \ ] ] in particular , the limit in ( [ thm1eq ] ) is strictly positive for @xmath128 and zero for @xmath129 .
moreover , if in addition there exist @xmath130 and @xmath131 such that @xmath132 = 1 $ ] for all @xmath6 , and @xmath133 , then as @xmath88 , @xmath134 = 1 + o ( n^{2-r_0 } ) .\ ] ] for @xmath135 , the expectation in ( [ polybound ] ) is dominated by low - weight null vectors ( short hypercycles ) , in particular by null vectors with only 2 non - zeros .
it may be possible , by extending the argument of lemma [ lem3 ] below , to expand this expectation as a power series in @xmath136 , but we do not pursue this here . for @xmath110 the exponential rate in ( [ thm1eq ] )
is dominated by null vectors using a specific positive proportion @xmath137 of the ( roughly @xmath138 ) available rows ( and also possibly those using proportion @xmath139 of the rows , due to parity effects ) ; in fact , @xmath140 , where @xmath141 is the value of @xmath142 for which the supremum in ( [ fdef2 ] ) is attained .
see section [ mainproofs ] for details .
theorem [ thm1 ] deals with the _ expected _ number of null vectors . also of interest
is , for a fixed @xmath6 , the ( random ) number of rows @xmath37 at which the first non - zero null vector appears .
define @xmath143 the first @xmath37 for which @xmath144 .
standard linear algebra implies that @xmath145 .
we define another threshold , @xmath146 , through an analytic description that needs more notation ( we shall give a probabilistic interpretation later on ) . for @xmath147 set @xmath148 provided @xmath149=1 $ ] we can and do extend @xmath150 continuously to @xmath151 , since @xmath152 as @xmath153 .
note that @xmath154 as @xmath155 provided @xmath156=1 $ ] , and that if @xmath114 < \infty$ ] then as @xmath157 we have @xmath154 and @xmath158 .
set @xmath159 and note that @xmath160 for all @xmath147 , so integrating from @xmath26 to @xmath21 we get @xmath161 , provided @xmath149 = 1 $ ] . for @xmath162 , define @xmath163 with the convention @xmath164 in operation in ( [ gs0 ] ) .
observe that if @xmath165 has unbounded range ( e.g. if @xmath166 = 1 $ ] ) then @xmath167 is the identity map on @xmath168 .
see figure [ fig1 ] for an example .
define @xmath169 in ( [ ubardef ] ) , the set defining @xmath146 is non - empty provided @xmath170=1 $ ] and @xmath114 < \infty$ ] , since as @xmath171 we have @xmath172 and @xmath173 . the relevance of @xmath146 for the null vector problem is shown by the next result . [ thm3 ] suppose @xmath90 are uniformly bounded and @xmath174 = 1 $ ] for all @xmath6 .
then @xmath175 , and for any @xmath176 , @xmath177 = 1.\ ] ] theorem [ thm3 ] is proved in section [ secpoisson ] .
we exclude the case where @xmath178 > 0 $ ] from the statement of theorem [ thm3 ] ; different phenomena occur in that case ( see proposition [ prop3 ] below ) .
this case is also discussed in @xcite , where the functions @xmath150 and @xmath165 also play a role .
typically , @xmath146 defined by ( [ ubardef ] ) will satisfy @xmath179 . in many cases
, there is a single solution in @xmath180 , denoted @xmath181 , to @xmath182 , and @xmath183 .
however , the situation is complicated by the fact that @xmath184 typically has at least one discontinuity .
we defer a more detailed discussion of the properties of the functions @xmath150 , @xmath165 , and @xmath185 , and the corresponding thresholds , to section [ secpoisson ] below .
figure [ fig1 ] provides an example . .
the left plot shows parts of the curves @xmath186 ( all the line ) and @xmath187 ( solid line ) .
the right plot shows parts of the curves @xmath188 ( all the line ) and the locus of @xmath189 ( solid line ) .
the left plot shows that @xmath190 has two discontinuities , one at @xmath191 and one at @xmath192 , with the first corresponding to a jump from @xmath193 to @xmath194 and the second to a jump from @xmath195 to @xmath196 .
the right plot shows the single positive solution of @xmath182 at @xmath197 , so @xmath198 .
it is not a coincidence that the curves @xmath165 and @xmath150 seem to mirror each other : see lemma [ lem : hpsi ] below.,title="fig:",width=264 ] .
the left plot shows parts of the curves @xmath186 ( all the line ) and @xmath187 ( solid line ) .
the right plot shows parts of the curves @xmath188 ( all the line ) and the locus of @xmath189 ( solid line ) .
the left plot shows that @xmath190 has two discontinuities , one at @xmath191 and one at @xmath192 , with the first corresponding to a jump from @xmath193 to @xmath194 and the second to a jump from @xmath195 to @xmath196 .
the right plot shows the single positive solution of @xmath182 at @xmath197 , so @xmath198 .
it is not a coincidence that the curves @xmath165 and @xmath150 seem to mirror each other : see lemma [ lem : hpsi ] below.,title="fig:",width=264 ] theorem [ thm3 ] leaves open the sharp asymptotics of @xmath199 : we believe that the upper bound in theorem [ thm3 ] ( i.e. , @xmath146 ) is sharp : [ conj1 ] if @xmath90 are uniformly bounded and @xmath200 = 1 $ ] for all @xmath6 , then @xmath199 converges in probability to @xmath146 as @xmath88 .
an equivalent statement to the fixed - weight case @xmath201 of this conjecture seems to have been established recently in the random - xorsat literature : see the comments in section [ between ] .
the probabilistic interpretation of the thresholds @xmath202 and @xmath146 is in terms of the _
2-core _ of @xmath41 ; this is the terminal state of an iterative algorithm that deletes every row incident to a column of degree 1 ( see section [ sec : cores ] below for details ) .
let @xmath203 denote the event that @xmath41 possesses a @xmath40-core ( i ) whose number of rows is bounded below by @xmath204 , and ( ii ) which contains more rows than columns of non - zero weight ( all of which have weight 2 or more ) .
in particular , for @xmath176 , @xmath205 implies that @xmath41 has a non - empty @xmath40-core .
if @xmath90 are uniformly bounded , and @xmath206 , then in theorem [ c : alphabar ] below we will show that , under certain additional conditions on @xmath102 , there exists @xmath176 ( allowed to depend on @xmath11 ) such that @xmath207 = 1 $ ] for @xmath11 in some interval of the form @xmath208 with @xmath209 . a more delicate description of the behaviour of the @xmath40-core in terms of the function @xmath150 defined at ( [ psidef ] )
will be given in theorem [ c : alphabar ] below : the limiting aspect ratio of the @xmath40-core being less than or greater than @xmath21 depends on the sign of @xmath210 .
in the example shown in figure [ fig1 ] , and also in the fixed weight setting , @xmath210 changes sign only once , but in the the general random weight setting it may change sign _
multiple _ times ; see the example in figure [ fig2 ] below .
thus the random weight setting gives rise to subtle new phenomena not present in the fixed weight case that has been the focus of previous work .
our techniques may be used to obtain information about the _ weight profile _ of the left null space .
bayes theorem applied to equation ( [ byweight ] ) below shows that the weight of a randomly chosen left null vector has distribution given by @xmath211 = \frac { \binom{m}{k } { { { \mathbb{p}}_{\rho_n } } } [ a ( n , k)]}{{{{\mathbb{e}}_{\rho_n } } } [ { { \mathcal{n}}}(n , m ) ] } ; \ ] ] the quantities on the right - hand side here are studied in detail in section [ sec : occupancy ] .
let @xmath212 denote the event that the row vector @xmath213 is null for @xmath0 , i.e. , @xmath214 one interpretation is in terms of the _ random allocation model_. suppose we have @xmath6 urns , and for each row of @xmath0 we allocate a collection of balls to a chosen set of urns ( determined by the unit entries of that row of @xmath0 ) .
event @xmath215 is the event that all the urns end up with an even number of balls .
random allocations have been extensively studied ; see e.g.@xcite , and the monographs @xcite .
the following theorem , which we prove in section [ sec : occupancy ] , describes the exponential rate of decay for @xmath216 $ ] where @xmath217 has a finite positive limit .
the theorem excludes the case in which _ both _ @xmath8 and @xmath218 only take odd values ; if @xmath37 is odd and @xmath90 is odd a.s .
, then @xmath219 = 0 $ ] since the total number of units in the matrix is odd .
[ allothm ] suppose that @xmath220 as @xmath88 , and that either ( i ) @xmath221 for all @xmath6 ; or ( ii ) @xmath222 > 0 $ ] .
then @xmath223 = - r_\rho ( \alpha ) , \end{aligned}\ ] ] where @xmath224 is continuous and nondecreasing in @xmath225 and is defined by @xmath226 } \left ( \frac { \left ( \rho ( 1 - 2\gamma ) \right)^\alpha } { 2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right).\end{aligned}\ ] ] the relevance of theorem [ allothm ] to theorem [ thm1 ] is clear ; see the formula ( [ eq1 ] ) below .
we present an interesting special case of theorem [ allothm ] that can be understood in isolation .
let @xmath227 denote the probability that all the components @xmath228 of a multinomial @xmath229 random vector @xmath230 are even . here
@xmath228 can be interpreted as the occupancy of urn @xmath57 after @xmath37 balls are independently and uniformly distributed into @xmath6 distinct urns : see e.g. @xcite , @xcite , or @xcite .
then @xmath231 this formula is known in the ehrenfest urn literature ( see @xcite ) and can also be obtained from ( [ pmulti ] ) below .
if @xmath37 is odd , @xmath232 must be zero .
[ multi ] let @xmath233 denote the probability that all the @xmath6 components of a multinomial @xmath234 random vector are even .
suppose that @xmath218 is even for each @xmath6 and @xmath235 as @xmath88 .
then @xmath236 this result follows from theorem [ allothm ] , as we shall show in section [ sec : alloproof ] .
proposition [ multi ] can also be derived from a result of kolchin ( * ? ? ?
* theorem 2 , p. 141 ) .
we now describe the special case of the results in section [ sec : general ] when @xmath237=1 $ ] , for some fixed @xmath238 .
the existing literature is largely concerned with this case ( see the discussion in section [ sec : related ] below ) .
let @xmath239 .
define the thresholds @xmath240 , @xmath241 , and @xmath242 to be the values of @xmath106 , @xmath202 , and @xmath146 , respectively , in the case where @xmath237=1 $ ] ( i.e. where @xmath243 . by ( [ alphastar2 ] ) we have @xmath244 where @xmath245 } \left ( \frac{(1 + ( 1 - 2\gamma)^r ) ^\alpha } { 2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right ) .\ ] ] by proposition [ alphaprop ] , @xmath246 for @xmath247 but @xmath248 for @xmath249 .
if @xmath51 , @xmath250 has a single solution in @xmath180 ( see proposition [ psiprop ] below ) denoted @xmath251 and satisfying @xmath252 and @xmath253 for example , @xmath254 , @xmath255 , and @xmath256 so @xmath257 ( see also figure [ fig8 ] below ) .
the next result is a specialization of theorems [ thm1 ] and [ thm3 ] together with theorem [ c : alphabar ] and proposition [ lessthan1 ] .
[ thm0 ] let @xmath258 .
suppose that @xmath259 in probability .
suppose that @xmath260 as @xmath88 .
then @xmath261 = f_r ( \alpha);\ ] ] in particular , the limit in ( [ meandecay ] ) is strictly positive if and only if @xmath249 . if also @xmath262=1 $ ] ( so @xmath51 ) and @xmath90 are uniformly bounded , then @xmath263 and for any @xmath176 , @xmath264 = 1,\ ] ] and moreover , there exists @xmath265 such that for all @xmath266 , @xmath267 = \begin{cases } 0 & ~\textrm{if}~ \alpha < { \underline{\alpha\mkern-4mu}\mkern4mu } _
r \\ 1 & ~\textrm{if}~ \alpha > { \underline{\alpha\mkern-4mu}\mkern4mu } _ r .\end{cases}\ ] ] the sharp monotone transition displayed by ( [ zeroone ] ) in the fixed weight case was indicated by cooper @xcite . in the general ,
random weight setting , the picture can be more complicated , and there exist examples where the transition is _ non - monotone _ : see the example in figure [ fig2 ] below . in the case
@xmath50 , compare theorem [ thm0 ] to proposition [ prop3](ii ) below : the first cycle in the random graph appears at @xmath268 , where @xmath269 has an asymptotic distribution on @xmath270 .
it is a classical result that for @xmath271 , if @xmath272 , the number of cycles @xmath273 in an erds rnyi graph has a poisson limit with finite expectation , but the limiting expectation is infinite for @xmath274 ( see e.g. @xcite ) ; we could not find in the literature an explicit reference to the fact that the expectation blows up _ exponentially _ with @xmath6 for @xmath275 , at the rate given by theorem [ thm0 ] ( a classical result of erds and rnyi states that at @xmath276 , the expected number of cycles grows as @xmath277 : see theorem 5a of @xcite ) . in this section
we discuss numerical and asymptotic evaluation of the thresholds in our results for the fixed row - weight case described in section [ sec : fixed ] .
table [ tab1 ] shows values of @xmath241 , @xmath240 , and @xmath278 , for @xmath279 .
previous computations of these thresholds are reviewed in section [ sec : related ] .
as suggested by the numerical results , it can be shown that , for @xmath238 large enough , @xmath280 ; this is a consequence of the following result .
[ threshasym ] as @xmath281 , @xmath282 the asymptotic result for @xmath240 in ( [ starsim ] ) is due to calkin @xcite ; we prove the other two parts in section [ secpoisson ] below .
.threshold parameters for @xmath238-uniform random hypergraphs .
note that @xmath283 is not defined when @xmath284 or @xmath40 . [ cols="^,^,^,^,^,^,^,^,^",options="header " , ] one can obtain arbitrarily sharp upper and lower bounds for the solution @xmath285 of @xmath182 in the case @xmath286 , @xmath51 , as follows . in this case , by ( [ mp3 ] ) we have that @xmath287 , where we set @xmath288 for @xmath289 , @xmath290 is strictly increasing for @xmath291 $ ] .
thus if @xmath292 , it follows that @xmath293 . also , @xmath294 , @xmath295 , and @xmath296 , so @xmath297 for @xmath298 but @xmath299 for @xmath300 $ ] . hence starting with @xmath301 ( an inequality proved in proposition [ psiprop ] below ) , we can iterate to obtain an increasing sequence of lower bounds @xmath302 for @xmath251 .
conversely , starting instead with @xmath303 and iterating @xmath304 gives a decreasing sequence of upper bounds @xmath305 for @xmath251 .
for example , after one step we get @xmath306 proceeding up to @xmath307 for the upper bound and @xmath308 for the lower bound is sufficient to obtain the @xmath281 asymptotic expression @xmath309 which will be the main ingredient in the proof of the @xmath283 result in ( [ starsim ] ) .
in fact , this iterative procedure converges , so @xmath310 and @xmath311 . to prove convergence it is sufficient to show that @xmath312 at @xmath313 .
a calculation shows that @xmath314 evaluated at @xmath313 comes to @xmath315 , so for the required inequality it suffices to show that @xmath316 the coefficient of @xmath317 in the power series expansion of the left - hand side of the last inequality is @xmath318 , and for the right - hand side it is @xmath319 , and both series are convergent on the given interval .
an induction shows that @xmath320 for all integers @xmath321 .
so term - by - term comparison of the two power series gives the inequality . to end this section we discuss the numerical evaluations of @xmath240 in table [ tab1 ] . in this discussion
we use some claimed properties of the functions involved that we do not verify rigorously , since here we are only concerned with numerical estimation .
let @xmath322 so that @xmath323 } f_{r,\alpha } ( \gamma)$ ] .
differentiating , we obtain @xmath324 thus @xmath142 is a stationary value for @xmath325 if @xmath142 solves @xmath326 numerical curve sketching shows that ( [ gamma0 ] ) generically has at most 2 solutions in @xmath270 ; of such solutions , the smallest will be the local maximum , since @xmath327 as @xmath328 , by ( [ fdiff ] ) .
if ( [ gamma0 ] ) has no solutions in @xmath270 , then the supremum in ( [ fdef ] ) is either @xmath329 or @xmath330 . thus setting @xmath331 if ( [ gamma0 ] ) has no solutions in @xmath270 and @xmath332 to be the smallest positive solution to ( [ gamma0 ] )
otherwise , we have that @xmath333 whenever @xmath334 . for @xmath334 , @xmath335 if and only if @xmath336
moreover , proposition [ alphaprop ] shows that @xmath337 , so that for @xmath338 such that @xmath339 , @xmath340 .
thus to find @xmath341 , we solve for @xmath342 $ ] the equation @xmath343 where @xmath344 numerical curve plotting shows that @xmath345 is decreasing on @xmath346 $ ] , so ( [ numerical ] ) can be solved using efficient numerical methods ; let @xmath347 denote the solution to ( [ numerical ] ) .
then we compute @xmath240 via @xmath348 . in the simplest case , @xmath349 a.s .
, for a fixed @xmath239 ; then @xmath350 a.s.this fixed row weight ` hypergeometric ' model is studied by cooper @xcite .
a variation is the model in which @xmath238 units are assigned to the row _ independently _ and uniformly at random , with multiplicities reduced mod @xmath40 .
the latter ` binomial ' model corresponds to @xmath351 distributed as the number of odd components in a multinomial @xmath352 random vector ; then @xmath353 ( see lemma [ wnconv ] below ) .
the @xmath51 binomial model is studied by kolchin @xcite .
note that in this model rows of all zeroes may appear , in which case they are ignored ( in other words , empty hyperedges are discounted ) : this is a small effect since @xmath354 = o ( n^{-r/2 } ) $ ] , so a vanishing proportion of rows needs to be discarded .
phase transitions in the null vector problem for random matrices over finite fields with fixed row weight @xmath51 have been studied since the early 1990s . in the case of the binomial model , the threshold @xmath240 , @xmath51 , for @xmath355 $ ] , @xmath272 , was described by balakin _
_ @xcite and kolchin @xcite ( having been announced in @xcite ) ; in these results @xmath240 is characterized by the fact that the expected number of non - trivial null vectors tends to @xmath26 ( @xmath356 ) when @xmath357 ( @xmath358 ) , but the proofs show that the growth is in fact exponential for @xmath358 .
calkin @xcite and cooper @xcite also study @xmath240 , @xmath51 , and in particular calkin @xcite studies @xmath240 as @xmath281 ; both @xcite and @xcite work in the case @xmath359 .
note that cooper s @xcite expression of the matrix problem is transposed compared to ours .
the first part of our theorem [ thm0 ] represents a slight generalization of the results just mentioned above because it allows for any class of sequences @xmath90 provided @xmath259 in probability .
the case of finite fields of order @xmath360 has also been studied : see for instance @xcite . in these previous investigations , the analytic description of the threshold @xmath240 varies .
calkin @xcite gives the same description of @xmath240 as our equation ( [ fdef ] ) .
systems of nonlinear equations for computing @xmath240 have been proposed in @xcite ; these descriptions can be shown to be consistent with ours .
specifically , with @xmath361 as given at ( [ frdef ] ) , one may characterize @xmath341 by the two equations @xmath362 and @xmath363 . on the substitution @xmath364 ,
the first of these equations becomes , after some calculations along the lines of those in the proof of proposition [ multi ] below , @xmath365 the second equation , involving the vanishing of the derivative given at ( [ fdiff ] ) gives , after the same substitution for @xmath366 , @xmath367 the system of equations ( [ old1 ] ) and ( [ old2 ] ) is the same as that given by cooper @xcite , and , after some manipulation , is seen to coincide also with that given by balakin _
_ @xcite and kolchin @xcite . despite this agreement
, there are some small discrepancies in the numerical evaluations for @xmath341 in @xcite , which can presumably be put down to numerical inaccuracies . a similar tabulation to our tabulation of @xmath283 is given by cooper @xcite , who also gives an equivalent analytic description of @xmath283 to our ( [ mp3 ] ) ; see also dietzfelbinger _
@xcite which we discuss further in the next subsection .
we note also that @xmath241 has received considerable attention in its own right : see e.g. @xcite for its role in random xorsat .
the following problem arises in the xorsat literature .
let @xmath239 with @xmath51 .
let @xmath0 be our @xmath1 matrix , with @xmath368 , and suppose @xmath369 a.s . for all @xmath370 .
let @xmath371 denote the number of column vectors @xmath53 such that @xmath372 , where @xmath373 is chosen uniformly at random ( independent of @xmath0 ) .
thus @xmath371 is a random variable .
dubois and mandler @xcite show for @xmath374 , and dietzfelbinger _
@xcite extend to general @xmath239 with @xmath51 ( also providing a more detailed proof ) the following result ( see ( * ? ? ?
* theorem 3.1 ) and ( * ? ? ? * theorem 1 ) , and also @xcite ) : there is a constant @xmath375 such that provided @xmath376 , @xmath377 \to 1 $ ] as @xmath88 .
the proof of this in @xcite is based on a second moment calculation .
the analytical definition of @xmath378 in @xcite is not obviously the same as our definition of @xmath283 , but the definition in terms of cores ( see ( * ? ? ?
* proposition 3 and equation ( 4 ) ) ) seems to match our definition of @xmath283 , and the numerical values in @xcite are consistent with our @xmath283 . if we accept that @xmath379 , this result implies that if @xmath380 there is , for @xmath6 large enough , no non - zero left null vector for @xmath0 , as follows .
suppose that a non - zero @xmath62 satisfies @xmath381 . then @xmath382 implies @xmath383 .
so @xmath384 \geq { \mathbb{p } } [ n > 0 ] \to 1 $ ] , which contradicts the easy observation that @xmath384 = 1/2 $ ] for non - zero @xmath62 .
we may then deduce that in the case with @xmath385 , our theorem [ thm3 ] may be strengthened to @xmath386 in probability .
this implies that for @xmath11 in the interval @xmath387 , a form of substantialism occurs ; existence of any left null vector is unlikely , but if there is one , there are lots of them .
the classical cases of the constant weight model @xmath385 in which @xmath388 exhibit different behaviour from the case @xmath51 . recall the definition of @xmath389 from ( [ tdef ] ) .
[ prop3 ] ( i ) for @xmath49 , for any @xmath390 , @xmath391 = \exp \ { - z^2 / 2 \}$ ] .
\(ii ) for @xmath50 , for any @xmath392 , @xmath393 = ( 1-z)^{1/2 } \exp \left\ { \frac{z}{2 } + \frac{z^2}{4 } \right\}.\ ] ] in the case @xmath49 , @xmath394 correspond to @xmath37 repetitions of the experiment of placing a ball uniformly at random in one of @xmath6 urns .
then @xmath389 is the first trial at which a ball is placed in an urn which is already occupied , and its law is given by solution to the birthday problem @xcite : @xmath395 = \sum _ { j=1}^{m-1 } \log \left(1 - \frac{j}{n}\right ) = -\frac{m ( m-1)}{2 n } + o ( m^3/n^2 ) , \ ] ] provided @xmath396 .
take @xmath397 to obtain the result . in the case
@xmath50 , @xmath389 is the same as the number of edges which have to be added in the erds
rnyi random graph process in order for the first cycle to appear .
the formula ( [ rgform ] ) has been derived by janson ( * ? ? ?
* theorem 8.1 ) . for the sake of comparison with sparse matrix phenomena ,
the case where @xmath44 is selected uniformly at random from @xmath398 is worthy of mention .
in fact we give a result for the more general finite field gf@xmath399 $ ] for arbitrary prime power @xmath28 .
let @xmath400 .
suppose ( for this section only ) that @xmath36 are i.i.d .
random vectors that are uniformly distributed over @xmath401 , and let @xmath402 be the @xmath1 matrix over gf@xmath399 $ ] with rows @xmath403 .
we write @xmath404 for probability associated with this model .
define @xmath389 by ( [ tdef ] ) ( with addition mod @xmath28 ) .
the following elementary result proves that @xmath405 is likely to be a small integer , uniformly in @xmath6 .
write @xmath406 .
[ prop1 ] suppose that @xmath36 are independent and uniformly distributed on @xmath401 . as @xmath88
, @xmath407 converges to a distribution on @xmath408 : for any @xmath409 ,
@xmath410 \to \prod_{j = r}^\infty ( 1 - q^{-j } ) = : p_r \in [ 0,1 ] , \ ] ] where @xmath411 and @xmath412 as @xmath281 .
moreover , the following lower bounds apply : for @xmath80 and @xmath413 , @xmath414 \geq \begin{cases } \exp \ { - q^{1-r } \ } & \textrm{if } ~ q \geq 3 \\ \exp \ { - \frac{4}{3 } 2^{1-r } \ } & \textrm{if } ~ q = 2 \end{cases } .\ ] ] limit results of the form of ( [ tlim ] ) are classical , and apparently date back at least to an 1895 paper of landsberg ( see @xcite ) . the @xmath415 case of ( [ tlim ] ) corresponds to the @xmath416 , @xmath417 case of ( * ? ? ?
* theorem 3.2.1 , p. 126 ) , but with a slightly different probabilistic model : there the @xmath418 are uniform on the whole of @xmath419 , not just @xmath420 ; comparing the results shows that this difference in the probability measures used is negligible in the limit .
let @xmath421 denote the event that @xmath422 is linearly independent .
if @xmath421 occurs , then the span of @xmath423 is a subspace with @xmath424 non - zero elements . since @xmath425 is ( statistically ) independent of @xmath423 and uniform on @xmath426 , which has @xmath427 non - zero elements , @xmath428 = \frac{q^n - q^k}{q^n-1 } , \ ] ] so that , since @xmath429 = 1 $ ] , for @xmath430 , @xmath431 = { \mathbb{p } } _ * [ a_k ] = \prod _ { j=1}^{k-1 } { \mathbb{p } } _ * [ a_{j+1 } \mid a_j ] = \prod _ { j=1}^{k-1 } \frac{q^n - q^j}{q^n - 1 } = ( 1 - q^{-n } ) ^{1-k } \prod_{j=1}^{k-1 } ( 1 - q^{j - n } ) ; \nonumber\ ] ] with the usual convention that an empty product is @xmath21 .
taking @xmath432 we obtain @xmath433 = ( 1 - q^{-n } )
^{r - n } \prod_{j = r}^{n-1 } ( 1 - q^{-j } ) \to \prod_{j = r}^\infty ( 1 - q^{-j } ) , \ ] ] on letting @xmath88 , establishing ( [ tlim ] ) with @xmath434 as stated there .
the asymptotic form given for @xmath434 follows from writing @xmath435 and applying taylor s theorem .
moreover , it follows from the fact that @xmath436 provided @xmath48 that the convergence in ( [ pak ] ) is in fact monotone for @xmath370 , i.e. , @xmath437 \geq \sum _ { j = r}^{n-1 } \log \left ( 1 - q^{-j}\right ) \geq \sum_{j = r}^\infty \log \left ( 1 - q^{-j } \right ) .\ ] ] with @xmath438 , we have @xmath439 and @xmath440 for @xmath441 , so @xmath442 for all @xmath59 with @xmath441 .
thus for @xmath443 and @xmath444 , @xmath445 \geq - \sum_{j = r}^\infty q^{-j } - \sum_{j = r}^\infty q^{-2j } = - \frac{q^{1-r}}{q-1 } \left ( 1 + \frac { q^{1-r}}{q + 1 } \right ) .\ ] ] the stated lower bounds follow .
there is a separate class of results , initiated by the classical work of kovalenko @xcite and balakin @xcite , in which the weight is random with a law which changes as @xmath6 increases ; in the classical case it is @xmath447 for @xmath448 . much more on this model
is given in @xcite , for example .
in this section we work towards proving theorem [ allothm ] and proposition [ multi ] . our null vector problem can be naturally formulated in terms of classical occupancy problems of random allocations of balls into urns .
we shall use the following terminology .
suppose @xmath8 is a random variable taking values in @xmath408 , and @xmath430 , and @xmath449 are numbers in @xmath450 $ ] such that @xmath451 .
( in most of the rest of the paper we assume @xmath452 , but for this section we can allow @xmath8 to take value @xmath26 . )
let us say the random variable @xmath44 has the @xmath453 distribution if for each @xmath454 the conditional distribution of @xmath44 , given that @xmath455 , is binomial with parameters @xmath456 .
let us say that a random vector @xmath457 has the multinomial @xmath458 distribution if for each @xmath454 the conditional distribution of @xmath457 , given that @xmath455 , is multinomial with parameters @xmath459 .
recall from section [ sec : matrixlaw ] that we assume @xmath90 ( having the distribution of row weights for our matrix with @xmath6 columns ) is chosen to converge in distribution to a limiting random variable @xmath8 .
an important special case is the so - called _ binomial _ model . in the binomial scheme
take @xmath351 to be distributed as the number of odd components in a multinomial @xmath460 random vector . by lemma [ wnconv ]
below , @xmath461 as @xmath88 , so this is indeed a special case .
we write @xmath462 for probability associated with the binomial allocation scheme . for the general model
we write @xmath96 as before .
recall from ( [ adef ] ) that @xmath215 denotes the event that @xmath70 is a null ( row ) vector for @xmath0 .
[ lem : nullp ] in the binomial allocation scheme , we have the exact formula @xmath467 = 2^{-n } \sum_{j=0}^n \binom{n}{j } \left ( \rho ( 1 - ( 2j / n ) ) \right)^m .\ ] ] in the general allocation scheme , @xmath468 & = 2^{-n } \sum_{j \subseteq [ n ] } \left ( { { { \mathbb{e}}_{\rho_n } } } [ ( -1)^{s_{1,j } } ] \right)^m \\ \label{phyp } & = 2^{-n } \sum_{j=0}^n \binom{n}{j } ( 2 p^{(n)}_{j } - 1)^m,\end{aligned}\ ] ] where @xmath469 $ ] and @xmath470 is given by @xmath471 event @xmath215 occurs if and only if all the @xmath228 are even , so @xmath472 = { \mathbb{e}}\prod_{j=1}^n \left ( \frac{1 + ( -1)^{y_j } } { 2 } \right ) = 2^{-n } \sum_{j \subseteq [ n ] } { \mathbb{e}}\left [ ( -1)^{\sum_{j \in j } y_j } \right ] , \end{aligned}\ ] ] where the latter sum is over subsets @xmath473 of @xmath78 $ ] , including the empty set . since @xmath474 and @xmath475 are i.i.d . , ( [ panm ] ) follows .
consider the binomial allocation scheme .
in the binomial model , @xmath476 where @xmath477 has a multinomial @xmath478 distribution so that @xmath479 has a @xmath480 distribution . recalling that if @xmath481 then @xmath482=(sp + ( 1-p))^n$ ] , we then obtain ( [ pmulti ] ) from ( [ panm ] ) . in the general scheme ,
conditional on @xmath483 , the distribution of @xmath484 is hypergeometric with parameters @xmath485 .
we do not use the generating function ( see e.g. @xcite ) explicitly , but apply the hypergeometric probability mass function directly to ( [ panm ] ) .
let @xmath486 $ ] denote the set of values of @xmath57 for which @xmath487 .
for @xmath488 $ ] , we write @xmath489 for expectation in the case where @xmath490=1 $ ] . instead of fixing @xmath491 $ ] and choosing @xmath492 as a uniform random @xmath238-subset , we obtain an exact formula for @xmath493 $ ] by fixing @xmath494 and a @xmath238-subset @xmath492 , and selecting @xmath473 uniformly from the @xmath57-subsets of @xmath78 $ ] .
the probability @xmath470 that @xmath495 is even is given by summing probabilities for @xmath496 , giving the expression in ( [ pjk ] ) .
it follows that @xmath493 = 2 p_{j , r } - 1 $ ] , and hence @xmath497 = \sum_{r=1}^n ( 2 p_{j , r } -1 ) { \mathbb{p } } [ w_n = r ] = 2 p_{j}^{(n ) } -1.\ ] ] substitution of this into ( [ panm ] ) gives ( [ phyp ] ) .
the remaining parts of section [ sec : occupancy ] are concerned with asymptotic analysis of the quantities in lemma [ lem : nullp ] .
first we state a result that will enable us to work primarily with even @xmath37 , which has technical advantages .
[ oddlem ] suppose that @xmath87 and @xmath498 > 0 $ ] . then for any @xmath499 , @xmath500 + o ( \log n ) \leq \log { { { \mathbb{p}}_{\rho_n } } } [ a ( n , m ) ] \leq \log { { { \mathbb{p}}_{\rho_n } } } [ a ( n , m+3 ) ] + o
( \log n ) .\ ] ] the fact that @xmath498 > 0 $ ] and @xmath87 implies that there exist @xmath176 and @xmath501 such that @xmath502>{\varepsilon}$ ] for all @xmath6 large enough . for any @xmath503 , suppose that @xmath504 occurs .
then @xmath505 will occur if the @xmath506 additional rows themselves constitute a hypercycle .
with probability at least @xmath507 , these new rows each have @xmath238 units , and given this , there is a probability at least @xmath508 , say , that these units form a hypercycle . in other words , @xmath509 \geq \log { { { \mathbb{p}}_{\rho_n } } } [ a(n , m -3 ) ] + o ( \log n)$ ] . applying this inequality twice , once with @xmath510 in place of @xmath37 ,
gives the result .
recall that in general we assume @xmath87 .
next we give an elementary lemma that confirms the binomial model s place in this framework .
[ wnconv ] for @xmath8 a @xmath408-valued random variable , let @xmath511 be the number of odd components in a multinomial @xmath512 random vector .
then @xmath461 as @xmath88 .
let @xmath8 and @xmath511 be coupled in the natural way .
then for each @xmath430 , by the union bound @xmath513 \leq \binom{k}{2 } \frac{1}{n},\ ] ] which tends to zero as @xmath88 , and the result follows easily from this
. we will prove theorem [ allothm ] ( in section [ sec : alloproof ] ) by first showing that ( [ allolim ] ) holds in the binomial setting , using ( [ pmulti ] ) and the stirling approximation for the binomial coefficients as discussed in section [ sec : appendix ] .
then we will extend this to the general setting using an approximation argument described in section [ sec : approx1 ] .
we start by proving a slightly more general statement than ( [ allolim ] ) in the binomial case , which we will also need later in the proof of theorem [ thm1 ] .
[ binlem ] recall the definition of @xmath514 from ( [ rrhodef ] ) .
then @xmath515 is continuous and nondecreasing as a function of @xmath11 .
suppose that either ( i ) @xmath516 for all @xmath6 ; or ( ii ) @xmath517 > 0 $ ] .
suppose that there exist @xmath518 with @xmath519 such that , for all @xmath6 sufficiently large , @xmath520 .
then , @xmath521 & \leq - r_\rho ( \alpha_1 ) ; \nonumber\\ \liminf_{n \to \infty } n^{-1 } \log { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a ( n , m_n ) ] & \geq - r_\rho ( \alpha_2 ) .\end{aligned}\ ] ] in particular , if @xmath522 , then @xmath523 = - r_\rho ( \alpha ) .\ ] ] suppose that @xmath524 .
first assume that the @xmath218 are even . by ( [ pmulti ] ) and lemma [ pgf](iii ) , @xmath525 & \leq ( n+1 ) 2^{-n } \max_{0 \leq j \leq n } \binom{n}{j } | \rho ( 1 -(2j / n ) ) |^{m } \\
& \leq ( n+1 ) 2^{-n } \sup_{\gamma \in [ 0,1/2 ] } \binom{n}{\gamma n } ( \rho ( 1 -2\gamma ) ) ^{m},\end{aligned}\ ] ] where we set @xmath526 if @xmath59 is a not an integer in @xmath527 . using the upper bound on binomial coefficients from the first inequality in ( [ chooseub2 ] )
, we get @xmath528 \leq ( n+1 ) \sup_{\gamma \in [ 0,1/2 ] } \left ( 2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } \right)^{-n } ( \rho ( 1 -2\gamma ) ) ^{m_n } .\ ] ] by monotonicity , we then obtain @xmath529 \leq n^{-1 } \log(n+1 ) + \log \sup_{\gamma \in [ 0,1/2 ] } g_{m_n / n } ( \gamma),\ ] ] where we have set @xmath530 so with @xmath514 as defined at ( [ rrhodef ] ) , @xmath531 } g_\alpha ( \gamma)$ ] .
note that @xmath532 is continuous in @xmath142 for @xmath162 and @xmath342 $ ] , and @xmath533 is nonincreasing in @xmath11 ; this monotonicity implies , by dini s theorem , that if @xmath534 monotonically then @xmath535 converges uniformly to @xmath536 on the compact interval @xmath346 $ ] .
it follows that @xmath537 } g_\alpha ( \gamma)$ ] is continuous as a function of @xmath225 , and is also nonincreasing in @xmath11 .
in particular , this shows that @xmath515 is continuous and nondecreasing in @xmath11 , as claimed in the lemma .
moreover , we obtain from ( [ eq45 ] ) and the fact that @xmath538 that @xmath539 \leq \log \sup_{\gamma \in [ 0,1/2 ] } g_{\alpha_1 } ( \gamma ) , \ ] ] which gives the first inequality in ( [ infsup ] ) .
for the second inequality , we have from ( [ pmulti ] ) and ( [ chooselb ] ) that for any integer @xmath540 ,
@xmath541 \geq 2^{-n } \binom{n}{i_n } ( \rho ( 1 - ( 2i_n /n ) ) ) ^{m_n } \geq { { \mathrm{e}}}^{-1/6 } \left ( \frac{n}{2\pi i_n ( n - i_n ) } \right)^{1/2 } \!\ ! ( g_{m_n / n } ( i_n /n ) )
^{n } , \end{aligned}\ ] ] using the fact that @xmath218 is even .
then , since @xmath542 , @xmath543 \geq { { \mathrm{e}}}^{-1/6 } \left ( \frac{n}{2\pi i_n ( n - i_n ) } \right)^{1/2 } ( g_{\alpha_2 } ( i_n /n ) ) ^{n } .\ ] ] now use continuity of @xmath536 to choose a sequence of integers @xmath540 , @xmath80 , such that @xmath544 } g_{\alpha_2 } ( \gamma)$ ] , with @xmath545 and @xmath546 as @xmath88 .
the lower bound in ( [ infsup ] ) follows , for @xmath218 even .
the results in ( [ infsup ] ) extend to the case of odd @xmath218 with @xmath498 > 0 $ ] by lemma [ oddlem ] , which is applicable here by lemma [ wnconv ] . finally , ( [ binlim2 ] ) follows from ( [ infsup ] ) on taking @xmath547 and @xmath548 , for arbitrary @xmath176 , and using the continuity of @xmath514 .
the exact formula ( [ pmulti ] ) is simpler to work with than the more complicated exact formula ( [ phyp ] ) , but intuition suggests that the asymptotics of any of the models in the class with @xmath87 should be similar . in this section
we quantify this intuition .
[ approxlem1 ] suppose that @xmath87 and either ( i ) @xmath516 for all @xmath6 ; or ( ii ) @xmath517 > 0 $ ] .
suppose that there exist @xmath518 with @xmath519 and @xmath549 such that @xmath520 for all @xmath550 . then , uniformly over sequences @xmath218 satisfying the given conditions , @xmath551 - \log { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a(n , m_n ) ] \right| = 0 .
\label{approxa}\end{aligned}\ ] ] in particular , if @xmath552 , @xmath553 = \lim_{n \to \infty } n^{-1 } \log { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a(n , m_n ) ] .\ ] ] denote the weight of row 1 by @xmath90 in the general allocation scheme , and by @xmath511 in the binomial scheme
. then @xmath87 and @xmath461 , and we can work in a probability space where @xmath554 \to 0 $ ] . as in section
[ sec : exact ] , set @xmath555 ; on @xmath556 , the ( conditional ) law of @xmath484 is the same as in the binomial model .
thus @xmath557 } \left| { { { \mathbb{e}}_{\rho_n } } } [ ( -1)^{s_{1,j } } ] - { { { \mathbb{e}}^{\mathrm{bin}}_{\rho_n } } } [ ( -1)^{s_{1,j } } ] \right| \leq 2 { \mathbb{p } } [ w_{n } \neq w^{\rm bin}_{n } ] \to 0.\ ] ] first suppose that the @xmath218 are even . by ( [ panm ] ) with ( [ pmulti ] ) and ( [ delta ] )
, there exists a triangular array of numbers @xmath558 \cup\{0\ } , n \in { \mathbb{n}})$ ] satisfying @xmath559 as @xmath88 , and @xmath560 = 2^{-n } \sum_{j=0}^n \binom{n}{j } \left ( \rho ( 1 - ( 2j / n ) ) + \delta_{j , n } \right)^{m_n } .\end{aligned}\ ] ] let @xmath176 and choose @xmath561 large enough so that @xmath562 and @xmath563 . then choose @xmath209 such that @xmath564
. finally assume @xmath6 is large enough so that @xmath565 \cup \{0\ } } |\delta_{j , n}| \leq \delta$ ] and @xmath566 .
it follows from ( [ eq0 ] ) that @xmath587 = \log { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a(n , m_n ) ] + { \varepsilon}_n n + \log \left ( 1 + \frac{\delta_n } { { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a(n , m_n ) ] \exp \ { { \varepsilon}_n n \ } } \right ) , \ ] ] where @xmath588 is the final term in ( [ eq0 ] ) . by lemma [ binlem ]
, we have that @xmath589 \geq \exp \ { - n r_\rho ( \alpha_2 ) -{\varepsilon}n \}$ ] , for all @xmath6 large enough .
so we may take @xmath176 small enough so that @xmath590 \exp \ { { \varepsilon}_n n \ } } \right ) = o ( \exp \ { - n \ } ) , \ ] ] say . hence @xmath591
= n^{-1 } \log { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a(n , m_n ) ] + { \varepsilon}_n + o(1 ) .\ ] ] since @xmath592 and @xmath176 was arbitrary , ( [ approxa ] ) follows in the case of even @xmath218 . in the other case , lemma [ oddlem ]
yields the same conclusion .
the final statement in the lemma then follows from lemma [ binlem ] , which says that @xmath593 $ ] exists in @xmath594 when @xmath552 .
now we can complete the proofs of theorem [ allothm ] and proposition [ multi ] .
the theorem is now a consequence of lemmas [ binlem ] and [ approxlem1 ] .
set @xmath595 in the case @xmath596 , it follows from theorem [ allothm ] that @xmath597 } f_\alpha ( \gamma)$ ] .
proposition [ multi ] will follow once we prove that , setting @xmath598 , @xmath599 } f_\alpha ( \gamma ) = - ( \lambda \tanh \lambda ) ( 1 - \log ( \tanh \lambda ) ) + \log ( \cosh \lambda ) .
\label{sup1}\end{aligned}\ ] ] note that @xmath600 as @xmath601 . differentiating ( [ fdef ] ) gives , for @xmath602 , @xmath603 which is zero at @xmath604 defined implicitly in terms of @xmath11 by @xmath605 one can verify that for @xmath225 ( [ gamma1 ] ) defines a unique stationary value @xmath606 which is a local maximum , since for @xmath606 the right - hand side of ( [ gamma1 ] ) is positive , continuous , and strictly decreasing as a function of @xmath607 , vanishing at @xmath608 ; this local maximum is indeed the maximum of @xmath609 for @xmath342 $ ] since @xmath610 as @xmath328 ( and also @xmath611 ) . setting @xmath612
we see that @xmath613 as given by ( [ gamma1 ] ) , since we get @xmath614 . to verify ( [ sup1 ] ) we need to express @xmath615 in terms of @xmath366 to get the expression on the right - hand side of ( [ sup1 ] ) .
we have @xmath616 where we have used the fact that @xmath617 and @xmath618 .
collecting the terms involving @xmath607 in the last displayed equation , we see that they simplify to @xmath619 as given by ( [ gamma1 ] ) , so we verify ( [ sup1 ] ) .
we give an alternative proof of proposition [ multi ] based on a poissonization device ( as used by kolchin in his proof of theorem 2 in @xcite ) and large deviations arguments of a slightly different flavour from those in the proof above .
the proof in this section is direct , avoiding the general theorem [ allothm ] , but does use instead some relatively deep local limit theory .
the following result can be found for example in @xcite .
[ poissoneven ] suppose that @xmath620 are independent poisson random variables with mean @xmath621 .
let @xmath622 be i.i.d .
, where the law of @xmath623 is the same as the conditional law of @xmath624 , given that @xmath624 is even : @xmath625 = { \mathbb{p } } [ z_1 = k \mid z_1 \in 2 { \mathbb{z}}]$ ] . then @xmath626 } { { \mathbb{p}}\left[z_1 + \cdots + z_n = m\right ] } \left ( \frac{1 + { { \mathrm{e}}}^{-2 \mu } } { 2 } \right)^n.\ ] ] by the well - known relationship between the poisson and multinomial distributions ( see e.g. @xcite or @xcite ) , and bayes theorem , @xmath627 \\ & = \frac { { \mathbb{p } } [ z_1 + \cdots + z_n = m \mid z_1 \in 2{\mathbb{z } } , \ldots , z_n \in 2{\mathbb{z } } ] } { { \mathbb{p } } [ z_1
+ \cdots + z_n = m ] } \cdot { \mathbb{p } } [ z_1 \in 2{\mathbb{z}}]^n , \end{aligned}\ ] ] which with the expression for @xmath628 $ ] in lemma [ lem4](ii ) yields ( [ pi2 ] ) .
[ local ] let @xmath629 be an i.i.d . sequence of @xmath630-valued random variables , @xmath631 , and @xmath632 a sequence of even integers .
suppose that @xmath633 < \infty$ ] for some @xmath634 , @xmath635 \wedge { \mathbb{p}}[x_1 = 2 ] > 0 $ ] , and @xmath636 + o(n)$ ] . then @xmath637 = 0.\ ] ] write @xmath638 $ ] and @xmath639 $ ] , which is finite and positive by the conditions in the lemma . write @xmath640 and @xmath641 , so @xmath642 = 0 $ ] , @xmath643 = \sigma^2 $ ] , and @xmath644 . if @xmath645 , richter s local central limit theorem ( * ? ?
* chapter 7 , 1 and 4 ) tells us that @xmath646 = { \mathbb{p } } [ z_n = y_n ] = \theta \left ( n^{-1/2 } \exp \left\ { - \frac{1}{2 } y_n^2 \left(1 + o ( n^{-1/2 } y_n ) \right ) \right\ } \right ) , \ ] ] which is @xmath647 , since @xmath648 . from lemma [ poissoneven ]
, @xmath649 can be expressed as @xmath650 - n^{-1 } \log { \mathbb{p}}[z_1 + \cdots + z_n = m ] + \log \cosh \mu - \mu .\end{aligned}\ ] ] by assumption , @xmath651 is such that @xmath652 .
the proof proceeds by choosing @xmath653 so that @xmath654 = \alpha$ ] ; then the first term in ( [ eq50 ] ) vanishes in the limit by lemma [ local ] , and the proof of the theorem then reduces to evaluating the other logarithmic rate .
we choose @xmath653 so that @xmath654 = \alpha$ ] ; by lemma [ lem4](ii ) this means @xmath655 so that @xmath656 .
since @xmath657 is poisson with mean @xmath658 , @xmath659 & = { \lim_{n \to \infty } \frac{1}{n } \log}\frac{(n{{\lambda}})^{m_n } { { \mathrm{e}}}^{-n { { \lambda}}}}{m_n ! } \\ & = -{{\lambda}}+ \lim_{n \to \infty } \frac{m_n}{n } \left[\log n{{\lambda}}- \frac{1}{m_n } \log m_n ! \right ] .\end{aligned}\ ] ] stirling s formula implies that @xmath660 , so that @xmath661 & = -{{\lambda}}+ \lim_{n \to \infty } \frac{m_n}{n } \left[\log \left ( \frac{n{{\lambda}}{{\mathrm{e}}}}{m_n}\right ) + o(1)\right ] \nonumber \\ & = -{{\lambda}}+ \alpha \left(\log { { \lambda}}- \log \alpha + 1\right ) .\end{aligned}\ ] ] combining ( [ eq51 ] ) with the @xmath656 case of ( [ eq50 ] ) we complete the proof .
let @xmath662 denote the number of left null vectors of weight @xmath663 , so that @xmath664 the value of @xmath2 is the number of collections of rows of @xmath41 which sum to @xmath25 ( mod @xmath40 ) , and for each set of @xmath663 rows the probability that it sums to @xmath25 is @xmath665 $ ] .
hence @xmath666 & = \binom{m}{\ell } { { { \mathbb{p}}_{\rho_n } } } [ a ( n,\ell ) ] , ~\textrm{and } \\
\label{eq1 } { { { \mathbb{e}}_{\rho_n}}}[{{\mathcal{n}}}(n , m ) ] & = \sum_{\ell=0}^m \binom{m}{\ell } { { { \mathbb{p}}_{\rho_n } } } [ a ( n,\ell ) ] .\end{aligned}\ ] ] we can thus express @xmath667 $ ] using our exact formulae for @xmath668 $ ] given in lemma [ lem : nullp ] . the proofs of our main results , theorems [ thm0 ] , [ thm1 ] , and [ thm3 ] , will be based on an asymptotic analysis of ( [ eq1 ] ) . as in the proof of theorem [ allothm ] ( see section [ sec : allo ] ) it is most convenient to work in the binomial model , for which @xmath511 is the number of odd components in a multinomial @xmath669 vector .
thus a key step in the proof will be showing that , in the general case of @xmath87 , the expression in ( [ eq1 ] ) can be well approximated by the binomial case .
first , in the next section , we make some preliminary computations . before embarking on the main proof
, we study the rate functions that will appear .
define @xmath670 and recall from ( [ alphastar2 ] ) and ( [ fdef2 ] ) that @xmath671 } f_{\rho,\alpha}(\gamma)$ ] and @xmath672 .
note that for @xmath342 $ ] , @xmath673 . by continuity
, @xmath674 attains its supremum over @xmath342 $ ] ; we denote by @xmath675 $ ] the _ smallest _ point at which the supremum is attained .
we collect results on @xmath676 and @xmath106 in the next lemma , which will enable us to complete the proof of proposition [ alphaprop ] .
[ alphalem ] suppose that @xmath677 = 0 $ ] .
for any @xmath162 , @xmath678 , and @xmath111 is continuous and nondecreasing .
the threshold @xmath679 enjoys the following properties .
* @xmath112 $ ] , and @xmath107 for @xmath129 but @xmath109 for @xmath110 .
* if @xmath135 , then for any @xmath176 , @xmath680 } f_{\rho , \alpha } ( \gamma ) < 0 $ ] . * if @xmath110 , then @xmath681 .
* suppose that @xmath119 is another @xmath83-valued random variable , with @xmath120 $ ] , such that @xmath121 for all @xmath122 $ ]
. then @xmath123 .
* @xmath115 if and only if @xmath116 > 0 $ ] . * if @xmath117 = 1 $ ] , then @xmath118 . *
if @xmath114 < \infty$ ] , then @xmath113 . by lemma [ pgf ] , @xmath682=0 $ ] and @xmath683 ; hence @xmath684 and @xmath685 , so that @xmath686 .
since @xmath674 is nondecreasing as a function of @xmath162 , dini s theorem implies continuity of @xmath676 as a function of @xmath162 . for part ( i )
, @xmath687 implies that @xmath688 for @xmath689 , so that @xmath690 . on the other hand , for @xmath338 , @xmath691 $ ] , since by continuity there is some neighbourhood of @xmath26 for which @xmath692 .
since @xmath674 is nondecreasing as a function of @xmath11 , for @xmath693 , @xmath694 , i.e. , @xmath111 is also nondecreasing .
hence @xmath695 for @xmath110 . also , the fact that @xmath696 for @xmath135 is immediate from the definition of @xmath106 and the fact that @xmath697
. then @xmath698 by the continuity of @xmath111 established above .
thus we obtain part ( i ) . for part ( ii ) , suppose that @xmath135 .
suppose for some @xmath699 that @xmath700 .
since @xmath701 for @xmath702 , @xmath703 is strictly increasing in @xmath11 , so there exists @xmath704 for which @xmath705 , contradicting the definition of @xmath106 . this gives ( ii ) . for part ( iii ) , suppose that @xmath110
. then @xmath695 by part ( i ) of the lemma ; since @xmath706 , the supremum is attained in @xmath707 . for part ( iv ) , we have that for any @xmath342 $ ] , @xmath708 , since @xmath709 .
so @xmath710 for all @xmath162 , and hence @xmath711 .
for the remaining parts of the lemma we use more detailed properties of the generating function @xmath712 ( see lemma [ pgf ] ) . for part ( v )
, differentiating in ( [ ffdef ] ) we obtain @xmath713 this is well defined at least for @xmath714 . at @xmath715 this equates to @xmath716 $ ] , since by lemma [ pgf ] @xmath717 $ ] and @xmath718 .
so if @xmath719 > 0 $ ] , @xmath720 is equal to @xmath26 at @xmath721 and its derivative there is negative for any @xmath225 , so that , for any @xmath225 , @xmath722 for some @xmath702 . the `
if ' part of part ( v ) follows .
conversely , suppose that @xmath719 = 0 $ ] .
then the previous argument shows that @xmath723 , while a calculation shows that @xmath724 .
hence by continuity there exists @xmath209 such that for @xmath725 and @xmath726 we have @xmath727 . hence by taylor s theorem , @xmath728 for @xmath725 and @xmath726 . also , @xmath729 as @xmath730 , which is strictly negative apart from at @xmath731 .
thus by dini s theorem , for all @xmath11 small enough we have @xmath728 for @xmath732 .
so all together we have shown that @xmath733 for all @xmath11 sufficiently small .
hence @xmath734 in this case , giving the ` only if ' part of ( v ) . for part ( vi ) , suppose that @xmath117=1 $ ] , i.e. , @xmath735 . in this case ,
( [ ffdiff ] ) has a zero at @xmath736 if @xmath737 where @xmath738 we claim that @xmath739 is decreasing on @xmath707 , with a unique minimum of @xmath740 . to verify this
, we show @xmath741 for @xmath736 , which , after simplification , amounts to @xmath742 setting @xmath743 , it suffices to show that @xmath744 for @xmath745 $ ] , which can be verified by term - by - term comparison of the corresponding power series , namely @xmath746 .
hence @xmath747 has no solution for @xmath748 , in which case the only stationary value of @xmath749 is at @xmath715 , necessarily the maximum .
hence @xmath750 . on the other hand , if @xmath275 then @xmath751 , while @xmath752 , so by taylor s theorem and continuity there exists @xmath702 with @xmath753 .
hence @xmath754 , proving ( vi ) .
finally we prove part ( vii ) .
if @xmath114 < \infty$ ] , lemma [ pgf](ii ) implies that , as @xmath328 , @xmath755 + o(1)$ ] .
thus the final term on the right - hand side of ( [ ffdiff ] ) dominates in the @xmath328 limit , and there exists @xmath209 such that @xmath756 for all @xmath757 $ ] and all @xmath758 $ ]
. then by an application of the mean value theorem , @xmath759 for all @xmath758 $ ] . thus taking @xmath338 close enough to @xmath21
we see that @xmath760 , which implies that @xmath113 .
extract the relevant parts of lemma [ alphalem ] . in section [ sec :
approx1 ] we showed ( in lemma [ approxlem1 ] ) that @xmath761 $ ] can be well approximated by @xmath583 $ ] on the logarithmic scale , provided that @xmath272 .
the following result is an analogous approximation lemma for @xmath762 $ ] .
one could obtain such a result from lemma [ approxlem1 ] applied to ( [ eq1 ] ) , with some work ( including dealing separately with terms with @xmath763 : cf section [ sec : few ] below ) .
however , it is more convenient to proceed directly , albeit using similar ideas to the proof of lemma [ approxlem1 ] ; in this case we are helped by the fact that @xmath764 $ ] possesses monotonicity properties absent for @xmath765 $ ] .
[ approxlem2 ] suppose that @xmath87 and @xmath206 .
then @xmath766 - \log { { { \mathbb{e}}^{\mathrm{bin}}_{\rho_n } } } [ { { \mathcal{n}}}(n , m_n ) ] | = 0.\end{aligned}\ ] ] we use a coupling argument , constructing the general model with row weights distributed as @xmath87 on the same probability space as the binomial model with row weights distributed as @xmath767 ( by lemma [ wnconv ] ) . for any @xmath6 , we can use a probability space in which , for each row , the weight in each model converges almost surely to a copy of @xmath8 .
indeed , let @xmath768 be independent copies of @xmath8 . using the skorokhod representation theorem
, we may take @xmath769 as independent copies of @xmath90 , being the weights of the rows in the general model , such that @xmath770 almost surely . also , take @xmath771 to be the number of odd components in a multinomial @xmath772 distribution , so that @xmath773 are independent copies of @xmath774 and the weights of the rows in the binomial model .
let @xmath775 .
then for any @xmath209 , we may take @xmath6 large enough so that @xmath776 \leq \delta$ ] , uniformly in @xmath56 .
let @xmath777 denote the number of ` bad ' rows
. then @xmath778 is stochastically dominated by a @xmath779 variable . in particular , for any fixed @xmath176 and any @xmath780
, standard binomial tail bounds imply that we may take @xmath781 small enough , and hence @xmath6 sufficiently large , so that @xmath782 \leq { \mathbb{p } } [ { { \rm bin } } ( 2 \alpha n , \delta ) \geq { \varepsilon}n ] \leq \exp \ { - c n \ } .\ ] ] we claim that each row added to a matrix can increase the number of null vectors by at most a factor of 2 ; this follows from ( [ sigmadef ] ) and ( [ ndef ] ) .
hence @xmath783 where @xmath784 is the number of null vectors in the matrix with the @xmath785 and @xmath786 is the number of null vectors in the matrix with the @xmath771 .
in particular , on @xmath787 , the bound in ( [ difference ] ) is @xmath204 .
the statement in the lemma follows from ( [ badbound ] ) and ( [ difference ] ) , since @xmath176 and @xmath788 were arbitrary . in the asymptotics of @xmath789 $ ]
, it turns out that null vectors of _ low weight _
play a distinct and important role . recall ( [ wtsum ] ) .
the main result of this section is the following lemma , which exhibits a polynomial growth rate for null vectors of few rows .
[ lem3 ] suppose that there exist @xmath130 and @xmath131 such that @xmath132 = 1 $ ] for all @xmath6 .
suppose that @xmath522 .
then there exists @xmath790 such that @xmath791 = o ( n^{2-r_0 } ) .\ ] ] the exponent @xmath792 in ( [ lowwt ] ) can not be improved when @xmath793 > 0 $ ] , because @xmath794 $ ] is itself of order @xmath795 .
indeed , there are of order @xmath796 weight-2 candidate vectors , and each is null if each of the two corresponding rows have @xmath797 non - zeros in matching positions , an event of probability of order @xmath798 .
let @xmath799 .
let @xmath800 denote the ` column range ' of the matrix @xmath801 , that is , the number of columns of degree at least 1 .
let us take @xmath802 , to be chosen later . we shall estimate @xmath803 $ ] by considering separately the events @xmath804 and @xmath805 .
we interpret @xmath801 as arising from a random allocation scheme , where for each row we throw balls at random into @xmath6 urns ( columns ) . if @xmath804 then there is some set of @xmath806 columns , such that all the balls land in these @xmath806 columns .
for each ball , the probability that it lands in one of the first @xmath806 columns , given that the other balls cast so far for that row all land in the first @xmath806 columns , is at most @xmath807 .
hence since for each row at least @xmath797 balls are cast , and we consider @xmath808 rows here , @xmath809
\leq \binom{n}{k } \left ( \frac{k}{n } \right)^{\ell r_0 } \leq \frac{n^{k- \ell r_0 } k^{\ell r_0}}{k ! } .
\label{120501a}\ ] ] if @xmath805 then to have @xmath810 occur we need to have each of the columns in the range get hit at least twice ( i.e. , have degree at least 2 ) . thus if @xmath805 and @xmath810 occurs there is a collection of @xmath811 columns such that each column in the collection gets hit at least twice . let @xmath812 be the event that the column @xmath56 gets hit at least twice .
the probability that a particular entry is 1 , given the values of up to @xmath806 other entries in the same row , is at most @xmath813 .
hence the union bound yields for @xmath814 that @xmath815 \leq \binom{\ell}{2 } \left ( \frac{r_1}{n - k } \right)^{2},\ ] ] and hence we have @xmath816 \leq \left ( \binom{\ell}{2 } \left ( \frac{r_1}{n - k } \right)^{2 } \right)^{k+1}\ ] ] so that by the union bound , provided @xmath817 we have @xmath818 \leq \binom{n}{k+1 } \left ( \binom{\ell}{2 } \left ( \frac{2 r_1}{n } \right)^{2 } \right)^{k+1 } \leq \frac { n^{-(k+1 ) } \ell^{2(k+1 ) } c_1^{k+1 } } { ( k+1)!},\end{aligned}\ ] ] where we put @xmath819 .
combined with ( [ 120501a ] ) this gives @xmath820 \leq \frac { n^{k- \ell r_0 } k^{\ell r_0}}{k ! } + \frac { n^{-(k+1 ) } \ell^{2(k+1 ) } c_1^{k+1 } } { ( k+1)!}.\end{aligned}\ ] ] we assume @xmath220 , so for @xmath6 large enough so that @xmath821 we have for all @xmath663 , and for @xmath817 , that @xmath822 & = \binom{m_n}{\ell } { { { \mathbb{p}}_{\rho_n } } } [ a(\ell , n ) ] \nonumber \\ & \leq \left ( \frac { ( ( \alpha + 1 ) n)^{\ell } } { \ell ! } \right ) \left ( \frac { n^{k- \ell r_0 } k^{\ell r_0}}{k ! } + \frac { n^{-(k+1 ) } \ell^{2(k+1 ) } c_1^{k+1 } } { ( k+1 ) ! } \right ) .
\label{120502a}\end{aligned}\ ] ] taking @xmath823 , we obtain for each fixed @xmath663 that for some constant @xmath824 we have @xmath822 & \leq c(\ell ) ( n^{\ell ( 1-r_0 ) + \ell + r_0 -2 } + n^{\ell - k -1 } ) \nonumber \\ & = c(\ell ) ( n^{(\ell -1)(2 -r_0 ) } + n^{1 -r_0 } ) , \label{120501c}\end{aligned}\ ] ] which is @xmath825 for any fixed @xmath826 .
fix an integer @xmath827 , to be chosen later , and consider @xmath828 .
now put @xmath829 .
assume @xmath830 ; then for @xmath831 this choice of @xmath806 satisfies @xmath817 , so that ( [ 120502a ] ) remains valid .
also note that , since @xmath130 , @xmath832 provided @xmath826 . by the bound @xmath833 and similar for @xmath806 , there are constants @xmath834 such that the first term in the right side of ( [ 120502a ] ) ( i.e. the product of the first factor with the first term in the second factor ) is bounded by a constant times @xmath835 where for the inequality we used the fact that @xmath836 to replace @xmath837 by @xmath838 in the denominator
. similarly , there are constants @xmath839 such that the second term in the right side of ( [ 120502a ] ) is bounded by a constant times @xmath840 combining with ( [ 120502b ] ) , since @xmath130 so @xmath841 and @xmath842 , we can find a constant @xmath843 such that for @xmath844 we have @xmath845 \leq c_8 \left(\frac{c_8 \ell}{n } \right)^{\ell/2 } .\ ] ] by calculus we have that @xmath846 is decreasing in @xmath847 , so provided @xmath848 the last bound is maximized , over @xmath828 , at @xmath849 , so that @xmath850 \leq
c_8 \delta n \left ( \frac{c_8 k}{n } \right)^{k/2},\ ] ] which is @xmath851 provided we choose @xmath852 so that @xmath853 .
first we prove theorem [ thm1 ] for the binomial model , i.e. , for @xmath462 on the right - hand side of ( [ eq1 ] ) , and then use lemma [ approxlem2 ] .
specifically , we prove the following result .
[ binthm ] suppose that @xmath84 = 1 $ ] .
suppose that @xmath126 as @xmath88 . then with @xmath676 as defined by ( [ fdef2 ] ) , @xmath854 = f_\rho ( \alpha ) .
\label{suplim}\end{aligned}\ ] ] from ( [ pmulti ] ) , @xmath855 \leq ( m+1 ) ( n+1 ) 2^{-n } \sup_{0 \leq \ell \leq m } \sup_{0 \leq j \leq n } \binom{m}{\ell } \binom{n}{j } | \rho ( 1 - ( 2j / n ) ) |^\ell \nonumber\\ & \phantom{mmmmmmm } \leq ( m+1 ) ( n+1 ) 2^{-n } \sup _ { \beta \in [ 0,1 ] } \sup_{\gamma \in [ 0,1 ] } \binom{m}{\beta m } \binom{n}{\gamma n } | \rho ( 1 - 2\gamma ) |^{\beta m},\end{aligned}\ ] ] setting @xmath526 for @xmath856 . write @xmath857 taking @xmath858 in ( [ ub1 ] ) and using the first inequality in ( [ chooseub2 ] ) , we obtain @xmath859 \leq o ( n^{-1 } \log n ) + \log \sup _ { \beta \in [ 0,1 ] } \sup_{\gamma \in [ 0,1 ] } s_{m_n
/ n } ( \beta , \gamma ) .
\end{aligned}\ ] ] for any @xmath860 , routine calculus ( with a separate argument for @xmath861 ) shows that @xmath862 } \left ( \frac{b^\beta}{\beta^{\beta}(1-\beta)^{1-\beta } } \right ) = b+1,\ ] ] with the supremum attained at @xmath863 , so that from ( [ eq2 ] ) we have @xmath864 \leq o ( n^{-1 } \log n ) + \sup_{\gamma \in [ 0,1 ] } \log \left ( \frac { ( 1 + | \rho ( 1 - 2\gamma ) | ) ^{m_n / n}}{2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right ) .
\end{aligned}\ ] ] considering the transformation @xmath865 , we see that @xmath866 } \ !
\left ( \frac { ( 1 + | \rho ( 1 - 2\gamma ) | ) ^{\alpha}}{2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right ) = \ ! \sup_{\gamma \in [ 0,1/2 ] } \ ! \left ( \frac { ( 1 + | \rho ( 2\gamma- 1 ) | ) ^{\alpha}}{2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right ) \leq \ ! \sup_{\gamma \in [ 0,1/2 ] } \ !
\left ( \frac { ( 1 + \rho ( 1- 2\gamma ) ) ^{\alpha}}{2 \gamma^\gamma ( 1-\gamma)^{1-\gamma } } \right ) , \ ] ] since , for @xmath867 $ ] , @xmath868 , by lemma [ pgf](iii ) .
hence from ( [ eq3 ] ) we have , with @xmath869 as defined at ( [ fdef2 ] ) , @xmath870 \leq o ( n^{-1 } \log n ) + f_\rho ( m_n /n ) $ ] . since @xmath871 and @xmath872 is continuous ( see lemma [ alphalem ] ) , @xmath873 \leq f_\rho ( \alpha ) .\ ] ] for the lower bound , we use the fact that @xmath874 & \geq \sum_{\ell = 0}^{m_n } \binom{m_n}{\ell } \sum_{j=0}^ { \lfloor n/2 \rfloor } 2^{-n } \binom{n}{j } ( \rho ( 1-(2j / n ) ) ) ^\ell \\ & \geq 2^{-n } \sup_{\beta \in [ 0 , 1 ] } \sup_{\gamma \in [ 0,1/2 ] } \binom{m_n}{\beta m_n } \binom{n}{\gamma n } nonnegativity of the appropriate terms for both inequalities . using the lower bound in ( [ chooselb ] ) , similarly to above , we obtain that @xmath875 \geq f_\rho ( \alpha)$ ] . hence combining the upper and lower bounds , we obtain ( [ suplim ] ) .
now we can give the proof of our main result .
lemma [ binthm ] shows that ( [ thm1eq ] ) holds for the case where @xmath351 , and lemma [ approxlem2 ] shows that the result carries over to the general case .
for the final statement of the theorem , suppose that @xmath135 and that @xmath132 = 1 $ ] for some @xmath130 and @xmath131 .
lemma [ lem3 ] shows that , for a suitable @xmath209 , @xmath876 = o ( n^{2-r_0 } ) .\ ] ] for @xmath877 , we first restrict to the binomial model
. choose @xmath878 so that @xmath879 . by a similar argument to ( [ ub1 ] ) , but splitting the supremum over @xmath57 into two parts , @xmath880 \leq ( m_n+1 ) ( n+1 ) 2^{-n } \!\ ! \sup_{\delta
n \leq \ell \leq m_n } \sup_{j : | j - ( n/2 ) | \leq { \varepsilon}n } \ ! \scalebox{0.9}{$\dbinom{m_n}{\ell}$ } \scalebox{0.9}{$\dbinom{n}{j}$ } | \rho ( 1 - ( 2j / n ) ) |^\ell \nonumber\\ & \phantom{mmmmmm } { } + ( m_n+1 ) ( n+1 ) 2^{-n } \sup_{0 \leq \ell \leq m_n } \sup_{j : | j - ( n/2 ) | > { \varepsilon}n } \scalebox{0.9}{$\dbinom{m_n}{\ell}$ } \scalebox{0.9}{$\dbinom{n}{j}$ } | \rho ( 1 - ( 2j / n ) ) |^\ell . \end{aligned}\ ] ] similarly to ( [ eq3 ] ) , the second term on the right - hand side of ( [ ub2 ] ) is bounded above by @xmath881 } f_{\rho , m_n / n } ( \gamma ) \right\ } , \ ] ] which decays to @xmath26 exponentially fast , by lemma [ alphalem](ii ) , since @xmath882 . on the other hand , for @xmath883
, we have from lemma [ pgf](i ) that @xmath884 , for @xmath176 small enough , so that , since @xmath885 , @xmath886 , so by the choice of @xmath887 the first term in the right hand side of ( [ ub2 ] ) tends to @xmath26 exponentially fast .
hence @xmath888 < 0 .\ ] ] we next deduce a version of ( [ 266 ] ) with @xmath96 in place of the special case @xmath462 , using lemma [ approxlem1 ] once more .
to this end , observe first that the @xmath96-analogue of the sum in ( [ 266 ] ) consists of @xmath889 nonnegative terms , so is bounded between the largest term and @xmath889 times that same term , so that @xmath890 = n^{-1 } \log \max _ { \delta n \leq \ell \leq m_n } \binom{m_n}{\ell } { { { \mathbb{p}}_{\rho_n } } } [ a ( n,\ell ) ] + o ( n^{-1 } \log n ) .\ ] ] by lemma [ approxlem1 ] , for any @xmath176 , there exist @xmath891 with @xmath892 , uniformly for @xmath663 with @xmath893 and @xmath6 sufficiently large , such that @xmath894 = \binom{m_n}{\ell } { { { \mathbb{p}}^{\mathrm{bin}}_{\rho_n } } } [ a ( n,\ell ) ] \exp \ { { \varepsilon}_{n,\ell } n \ } .\ ] ] so we obtain @xmath895 < 0 , \ ] ] which , combined with ( [ ub3 ] ) , yields ( [ polybound ] ) .
given a set @xmath896 , whose elements we call _ vertices _ , a non - empty subset of @xmath897 is called a _
hyperedge_. given a collection @xmath898 of @xmath37 hyperedges , we refer to the pair @xmath899 as a _
hypergraph_. this hypergraph may be identified with an @xmath1 matrix @xmath900 with entries in @xmath901 ( the _ incidence matrix _ of the hypergraph ) , having no zero rows , as follows .
the entry @xmath902 of @xmath900 takes the value @xmath21 if and only if @xmath903 , in which case we say row @xmath56 is _ incident _ to column @xmath57 , and that hyperedge @xmath904 is incident to vertex @xmath905 , and refer to @xmath906 as an _ incidence _ of the hypergraph .
the number of hyperedges incident to a vertex @xmath907 is the _ degree _ of @xmath907 .
fix a hypergraph @xmath899 . for @xmath908 , the set @xmath909 of vertices which are incident to at least one of the hyperedges in @xmath910
is called the _ vertex span _ of @xmath910 .
we identify the hypergraph @xmath911 by the edge subset @xmath910 that induces it , and call @xmath908 a _
partial hypergraph_. a partial hypergraph @xmath912 is a _ hypercycle _ if every vertex @xmath907 has even degree with respect to @xmath910 . for an incidence matrix @xmath900 , a left null vector is the indicator of a hypercycle in the corresponding hypergraph . given a hypergraph @xmath899 , the _
2-core _ is defined via the following algorithm : 1 . if there exists no vertex of degree one ,
. 2 . otherwise , select an arbitrary vertex of degree one , and delete the unique incident hyperedge ; then return to step 1 .
the algorithm terminates , because the partial hypergraphs are decreasing ; the terminal partial hypergraph , which does not depend on the arbitrary choices made in step 2 ( see @xcite ) , is called the _
2-core _ of @xmath913 , denoted @xmath914 .
possibly @xmath915 has no hyperedges .
figure [ fig6 ] illustrates a hypergraph with 19 vertices and 12 hyperedges , of which 4 hyperedges are in the 2-core . .
the first 8 hyperedges are not in the 2-core .
the last 4 hyperedges form a linear system of rank 3 over gf[2 ] .
the entire @xmath916 system has rank 11 , and so contains a hypercycle ( hyperedges 9 and 11).,width=302 ] the connection between the @xmath40-core and hypercycles was exploited by cooper @xcite , following an idea that he attributes to molloy ( see @xcite ) .
the connection is demonstrated by the following useful observation .
[ hypercycle-2core ] suppose that the 2-core @xmath917 of a hypergraph @xmath899 has vertex span @xmath918 and size ( number of hyperedges ) @xmath919 .
* any hyperedge @xmath920 can not belong to a hypercycle of @xmath899 . * if @xmath921 , then @xmath899 contains no hypercycle . * if @xmath922 , then @xmath899 contains a hypercycle .
if there are @xmath923 hyperedges not in the 2-core @xmath924 , there exists a labelling of them as @xmath925 with the property that , for every @xmath57 , @xmath926 has some vertex with degree one after hyperedges @xmath927 are removed .
suppose @xmath899 has some hypercycle @xmath912 .
none of @xmath925 can belong to @xmath910 : otherwise , there would be some minimum @xmath57 for which @xmath928 , and this @xmath926 has some vertex @xmath907 of degree one in the partial hypergraph from which @xmath927 have been removed , which contains @xmath910 ; so @xmath907 can not have even degree in @xmath910 , which is a contradiction .
this proves ( i ) , and ( ii ) follows . for ( iii ) , say @xmath929 .
then there are @xmath930 non - empty partial hypergraphs , but only @xmath931 possible indicator vectors for a set of vertices of odd degree . by the pigeonhole principle
, there must be two distinct partial hypergraphs @xmath932 for which the sets of vertices of odd degree are the same .
then @xmath933 is a hypercycle . in this subsection
we consider a certain _ uniform random _ hypergraph model , which is different from ( but related to ) the hypergraph model induced by our random matrix @xmath934 ; in section [ secpoisson ] we will connect the two models .
the incidences ( non - zero entries ) in a random incidence matrix may be viewed as edges of a random bipartite graph , whose left and right nodes are the row labels and column labels , respectively .
a standard probability model for a random bipartite graph is to fix the degrees of all nodes in advance ( subject to a consistency condition ) , and sample uniformly from the bipartite graphs with this set of left and right node degrees . by interpreting an incidence matrix as a hypergraph , as above
, this automatically gives a _ uniform random hypergraph _ model , where the degree of a left node is the _ weight _ of a hyperedge ( meaning its number of incident vertices ) , and the degree of a right node is the vertex degree defined above . darling and norris @xcite analyse the statistical properties of the 2-core for such random hypergraphs under suitable conditions on the hyperedge - weight and vertex - degree distributions .
in unpublished work of the same authors , a generalization to unbounded vertex degrees and row weights is given , under finite third moments assumptions .
for the purposes of the present paper , we require a more modest relaxation of the conditions of @xcite , to cover the case where the row weights remain uniformly bounded but the vertex degrees are approximately poisson distributed . for each @xmath6 , define vectors of nonnegative integers @xmath935 and @xmath936 with @xmath937 and @xmath938 ; we assume that @xmath939 and @xmath940 are compatible in the sense that @xmath941 .
we also assume that @xmath942 .
suppose that for each @xmath943 and @xmath944 , @xmath945 define generating functions @xmath946 and @xmath947 .
we assume that the weights are uniformly bounded , i.e. , @xmath948 for all @xmath949 sufficiently large , and the degree distribution has all moments , i.e. , @xmath950 for all @xmath951 .
under these conditions , @xmath952 and @xmath953 are the ( finite ) means corresponding to these distributions . consider a sequence of random hypergraphs with @xmath6 vertices and @xmath218 hyperedges , selected uniformly from those hypergraphs with edge weight multiplicities @xmath940 and vertex degree multiplicities @xmath939 . to avoid triviality , we assume that @xmath966 ( equivalently , @xmath967 ) , i.e. , there are no 1-edges , and @xmath968 ( otherwise the 2-core is of no interest ) .
define @xmath969 from the conditions @xmath970 and @xmath971 , we deduce that @xmath972\to { \mathbb{r}}$ ] is strictly increasing ; moreover @xmath973 ( since @xmath974 ) and @xmath975 , so @xmath976 takes values in @xmath977 , and there exists a largest solution @xmath185 in @xmath977 of the equation @xmath978 .
that is , @xmath979 in the case where @xmath980 and the curve @xmath981 crosses the curve @xmath982 ( rather than just touching it ) at @xmath983 , we also have @xmath984 now we can state the result on the @xmath40-core that we shall use , which amounts to a variant of theorem 7.1 of @xcite .
[ t:2core ] consider a sequence of uniform random hypergraphs associated with sequences @xmath940 and @xmath939 satisfying ( [ uniflimits ] ) with @xmath985 for all @xmath949 large enough and @xmath986 for all @xmath987 .
suppose that the corresponding pair ( [ e : pgfs ] ) of random - incidence generating functions has @xmath974 , @xmath988 , and is such that @xmath989 , given by ( [ sstar ] ) , has either @xmath990 or @xmath989 satisfying ( [ e : crossing ] )
. then the following hold a.s . in the limit as @xmath991 . *
if @xmath990 , the proportion of hyperedges which survive in the 2-core converges to zero . *
if @xmath992 , then for any @xmath993 with @xmath994 , the proportion of weight-@xmath806 hyperedges which survive in the 2-core is asymptotically @xmath995 ; overall , a proportion @xmath996 of hyperedges survive , and a proportion @xmath997 of incidences . *
if @xmath992 , then for any @xmath998 with @xmath999 and @xmath1000 , the proportion of vertices of degree @xmath806 whose degree in the 2-core is @xmath960 converges to @xmath1001 * if @xmath992 , the 2-core is again a uniform random hypergraph , given its hyperedge weights and vertex degrees , whose distributions are determined by the previous assertions . as mentioned above , in @xcite all but finitely many coefficients of the generating functions ( [ e : pgfs ] ) were taken to be zero , but the methods admit the modest extension of this section , and indeed can be extended to the case where @xmath1002 and @xmath1003 are finite , corresponding to finite third moments for hyperedge weight and vertex degree distributions .
because of its proximity to the result in @xcite , we do not prove theorem [ t:2core ] here . in the previous subsection we assumed the row and column weights of our random matrix
were specified in advance , but now we return to the random matrix model used in the rest of the paper , so our random @xmath1004 incidence matrix @xmath900 will be precisely the matrix @xmath934 , described in section [ sec : matrixlaw ] , i.e. , with i.i.d . rows with weights having the distribution of @xmath90 , and corresponding generating function @xmath1005 having limit @xmath1006 . to justify being able to apply theorem [ t:2core ] in this setting , we give the following strong law of large numbers for the empirical distributions of the row and column weights of @xmath0 . first note that @xmath1014 = { \mathbb{p } } [ w_n = k]$ ] , which converges to @xmath1015 $ ] by assumption . to deduce almost sure convergence from this convergence in means
, we use the azuma hoeffding inequality in a standard way , as follows .
fix @xmath6 and for @xmath1016 let @xmath1017 be the @xmath1018-algebra generated by the rows @xmath1019 of @xmath934 .
define @xmath1020 $ ] , with @xmath1021 $ ] . since resampling a single row changes the number of rows of weight @xmath806 by at most 1 , we have for @xmath1022 that @xmath1023| \leq 1,\ ] ] where @xmath1024 is defined like @xmath1009 but based on a matrix with the @xmath56th row resampled . by the azuma
hoeffding inequality applied to the martingale @xmath1025 we obtain for any @xmath176 that @xmath1026| > { \varepsilon}n ] \leq 2 \exp ( - { \varepsilon}^2 n/2),\ ] ] so by the first borel
cantelli lemma , we have @xmath1027| \leq { \varepsilon}n $ ] for all but finitely many @xmath6 almost surely .
combined with the convergence of the mean , this gives us ( [ emprow ] ) .
the remaining two parts of the lemma use the assumption @xmath1028 = 1 $ ] for @xmath131 . to prove ( [ empcol ] ) note that the weight of the first column ( or any other column ) of @xmath934 is binomially distributed with parameters @xmath1029 ( number of trials ) and @xmath1030/n$ ] ( probability of success ) . hence @xmath1031 = { \mathbb{p}}[{\rm bin}(m_n,{\mathbb{e}}[w_n]/n ) = k]$ ] , and by binomial - poisson convergence this tends to @xmath1032 as @xmath88 .
given this convergence of means , we may prove ( [ empcol ] ) by a similar argument ( based on the azuma hoeffding inequality ) to the one used to prove ( [ emprow ] ) , since resampling a single row changes the number of columns of degree @xmath806 by at most @xmath99 . for the final statement in the lemma
, we have that @xmath1033 , \ \mathrm{a.s . } , \end{aligned}\ ] ] by ( [ emprow ] ) , and then ( [ incidences ] ) follows . [ transfer ] consider the random matrix model @xmath934 with row weight distribution @xmath87 , where the @xmath90 are uniformly bounded .
suppose that @xmath1034 .
then , a.s .
, taken as hypergraph incidence matrices the sequence @xmath934 defines a sequence of uniform random hypergraphs whose row weight and vertex degree distributions satisfy ( [ uniflimits ] ) with @xmath961 and @xmath962 given by @xmath1035 $ ] and @xmath1036 respectively , where @xmath1037 $ ] .
since the distribution of @xmath41 is invariant under permutations of the rows or columns , conditional on the empirical distribution of row and column weights , all possible outcomes with those row and column weight distributions are equally likely , so this conditional distribution is indeed uniform .
moreover by lemma [ emplem ] the limiting proportion of rows of weight @xmath806 is given by @xmath1015 $ ] and the limiting proportion of columns of degree @xmath806 is given by @xmath1038 $ ] where @xmath1039 po@xmath1040 , with @xmath1041 .
hence conditionally on this sequence of empirical distributions , almost surely we have a sequence of random matrices satisfying the hypotheses of section [ sec : fluid ] . in the notation of section [ sec : fluid ] , in this case
@xmath1042 is the generating function of a po@xmath1040 random variable , so , by ( [ gf2 ] ) , the pair ( [ e : pgfs ] ) becomes @xmath1043 in this case , we have from ( [ varphidef ] ) that @xmath1044 and to emphasize the dependence on @xmath11 we will use the notation @xmath1045 for @xmath1046 from now on . recall from ( [ sstar ] ) that @xmath989 was defined as the largest @xmath1047 for which @xmath1048 . in order for the model of this section to fit into the setting discussed in section [ sec : fluid ] , we need to assume that @xmath1049 and @xmath988 . here
@xmath1050}$ ] and @xmath1051}{{\mathbb{e}}[w]}$ ] .
so it suffices to assume that @xmath225 , @xmath124 = 1 $ ] , and @xmath1052 < \infty$ ] ; in this case the argument in section [ sec : fluid ] shows that @xmath989 is well defined .
note that @xmath185 depends both on @xmath102 and on @xmath11 ; in this section we write @xmath1053 to emphasize the dependence on @xmath11 ; we will show ( see lemma [ gstarlem ] ) that the present definition is equivalent to that at ( [ gs0 ] ) given in section [ sec : general ] . for
any solution @xmath1047 to @xmath1054 , so in particular for @xmath1055 , provided @xmath1056 , we have @xmath1057 as given by ( [ hdef ] ) .
we note some facts about @xmath190 ; recall the definition of @xmath202 from ( [ alphasharp ] ) .
[ gstarlem ] suppose that @xmath124 = 1 $ ] and @xmath114 < \infty$ ] . with the convention @xmath1058 , the definition ( [ gs0 ] ) is equivalent to the definition ( [ sstar ] ) of @xmath190 as the largest solution of @xmath1054 .
also , @xmath1059 $ ] , and @xmath1060 for all @xmath1061 , and for @xmath1062 , the function @xmath190 is positive and strictly increasing , with @xmath1063 as @xmath171 .
now assume also that @xmath170 = 1 $ ] and @xmath1064 < \infty$ ] . then the following hold .
* we have @xmath1065 and @xmath1066 . *
the function @xmath185 is right continuous , and there is a finite set @xmath1067 , with @xmath1068 , such that @xmath185 is continuous apart from jumps at points of @xmath1069 . for each @xmath1070
, @xmath1071 is a local minimum for @xmath165 .
* if @xmath1072 , then @xmath190 satisfies the crossing condition ( [ e : crossing ] ) . since @xmath1073=1 $ ] and @xmath114 < \infty$ ] , we have @xmath1074 and @xmath1075
. therefore by continuity we may rewrite ( [ sstar ] ) as @xmath1076 using the convention @xmath164 . by the definition ( [ hdef ] ) of @xmath165 , for @xmath1077 it is easy to check that @xmath1078
if and only if @xmath1079 , and this shows that ( [ gs0 ] ) and ( [ sstar ] ) give equivalent definitions of @xmath190 . by ( [ hdef ] ) and subsequent remarks , @xmath1080 is positive , continuous in @xmath59 , and tends to infinity as @xmath157 . by the definition ( [ alphasharp ] ) , and the subsequent remark , @xmath1059 $ ] . by the definition ( [ gs0 ] ) , it is clear that @xmath1081 for @xmath1082 , and the fact that @xmath1083 is positive and strictly increasing for @xmath1084 is easily deduced from the continuity of @xmath165 .
also , given @xmath1085 we can choose @xmath1086 with @xmath1087 so that @xmath1088 , and together with the monotonicity of @xmath185 this shows @xmath1089 as @xmath171 .
for part ( i ) , under the extra assumption @xmath166=1 $ ] we have @xmath1090 going to infinity at 0 and at 1 , and by continuity @xmath165 attains its infimum on @xmath180 , so using ( [ alphasharp ] ) and ( [ gs0 ] ) we have that @xmath1091 is the supremum of a non - empty compact set contained in @xmath180 , and so lies in ( 0,1 ) .
the last part of ( i ) also follows from the continuity of @xmath165 . for part ( ii ) , under the extra assumption @xmath1064 < \infty$ ] , note first that if @xmath1092 then @xmath1093 .
hence @xmath185 is continuous at @xmath62 for all @xmath1094 .
now let @xmath1095 ; note that by ( [ gs0 ] ) and continuity of @xmath165 , we have @xmath1096 . take a monotonic sequence @xmath1097 tending to @xmath62 ; set @xmath1098 .
suppose first that @xmath1099 .
then the sequence @xmath1100 is nonincreasing ; denoting the limit by @xmath1101 we have @xmath1102 so @xmath1103 by continuity , and therefore @xmath1104 by ( 2.9 ) .
since also @xmath1105 by monotonicity we have @xmath1106 ; hence @xmath1107 is right - continuous at @xmath62 .
now suppose instead that @xmath1108 . set @xmath187 .
if @xmath165 does not have a local minimum at @xmath59 then @xmath1109 , so that @xmath1110 , and hence @xmath185 is left - continuous at @xmath62
. hence , if @xmath185 is discontinuous at @xmath62 then @xmath165 has a local minimum at @xmath1111 .
the function @xmath1112 is analytic and non - constant on @xmath180 so its zeros do not accumulate except possibly at 0 or 1
. however @xmath1113 implies @xmath1114 , so by the assumption @xmath1064 < \infty$ ] there exists @xmath1115 such that @xmath1116 for @xmath1117 and for @xmath1118 ; for the latter case we use the fact that , as @xmath155 , @xmath1119 if @xmath130 is the smallest possible value of @xmath8 .
thus @xmath165 has only finitely many local minima in @xmath180 , and hence @xmath165 has a local minimum at @xmath1111 for at most finitely many @xmath62 .
this completes the proof of ( ii ) . for part ( iii )
note that , for @xmath1077 , @xmath1120 if and only if @xmath1121 , so ( [ e : crossing ] ) gives @xmath1122 , which for @xmath1123 agrees with the definition ( [ gs0 ] ) .
to apply the results in section [ sec : fluid ] to @xmath934 , we need to assume that @xmath190 either is zero or satisfies ( [ e : crossing ] ) .
lemma [ gstarlem ] shows that a sufficient condition for this is that @xmath1124 , @xmath1072 . by theorem [ t:2core](ii ) and ( [ incidences ] ) , @xmath136 times the number of incidences which survive in the 2-core converges a.s . to @xmath1125 by theorem [ t:2core](iii ) and ( [ empcol ] ) , for @xmath1126 the proportion of original vertices whose degree in the 2-core is @xmath960 is asymptotically @xmath1127 the remainder having degree @xmath26 in the @xmath40-core ( some columns in the core have degree zero because in the algorithm of section [ sec : hypermodel ]
, we never delete any columns ) .
in other words , the 2-core vertex degrees have the distribution of a random variable @xmath1128 , where @xmath1129 ; by ( [ 2coreinc ] ) , @xmath1130 . as a check on the previous calculation ( [ 2coreinc ] ) of the number of surviving incidences , @xmath136 times the total number of incidences in the 2-core should converge to the mean of the vertex - degree distribution , which is @xmath1131 , as in ( [ 2coreinc ] ) . the key issue in predicting hypercycles and rank deficiency is whether the number of rows in the 2-core exceeds the number of occupied columns in the 2-core , which is treated in theorem [ c : alphabar ] ; a related result appears in @xcite . recall the definition of @xmath1132 from ( [ psidef ] ) and @xmath146 from ( [ ubardef ] ) .
[ c : alphabar ] suppose @xmath90 are uniformly bounded and @xmath156 = 1 $ ] .
let @xmath1124 .
consider the 2-core of the random incidence matrix @xmath934 where @xmath1133 as @xmath88 .
then if @xmath1134 , the number of rows in the 2-core is @xmath1135 , a.s .
now suppose @xmath1062 , so @xmath1136 , and suppose that @xmath1072 .
then : * @xmath136 times the number of rows in the 2-core converges a.s .
to @xmath1137 .
* @xmath136 times the number of occupied columns in the 2-core converges a.s . to @xmath1138 , where @xmath1139 . * almost surely , for all @xmath6 large enough
, the @xmath40-core has more rows than occupied columns if @xmath1140 but has fewer rows than occupied columns if @xmath1141 .
moreover , there exists @xmath209 such that if @xmath1142 , for all @xmath6 large enough , the @xmath40-core has more rows than columns and so the corresponding hypergraph has a hypercycle .
figure [ fig2 ] shows an example of some of the more exotic behvaiour that can occur in the random weight setting . in the case where @xmath1143 , @xmath1144 changes sign several times , and so theorem [ c : alphabar ] shows that as @xmath11 increases from @xmath26 the @xmath40-core switches from having asymptotically more columns than rows to having more rows than columns not just once ( at @xmath146 ) , but _
twice_. this non - monotone behaviour does not occur in the fixed weight case .
proposition [ psiprop ] below , and the subsequent discussion , explains some of the features in the figure . .
the left plot shows parts of the curves @xmath1145 ( all the line ) and @xmath187 ( solid line ) .
the right plot shows parts of the curves @xmath1146 ( dashed line ) and the locus of @xmath189 ( solid line ) .
again @xmath190 has two discontinuities , one at @xmath1147 and one at @xmath1148 , with the first corresponding to a jump from @xmath193 to @xmath1149 and the second to a jump from @xmath1150 to @xmath1151 .
the right plot shows the three positive roots of @xmath182 .
the two of these roots achieved by @xmath210 are at @xmath1152 and @xmath1153 .
the first corresponds to @xmath1154 and the second to @xmath1155 .
hence as @xmath11 ranges in @xmath180 , @xmath1156 changes sign from positive , to negative , to positive , and finally to negative again.,title="fig:",width=264 ] .
the left plot shows parts of the curves @xmath1145 ( all the line ) and @xmath187 ( solid line ) .
the right plot shows parts of the curves @xmath1146 ( dashed line ) and the locus of @xmath189 ( solid line ) .
again @xmath190 has two discontinuities , one at @xmath1147 and one at @xmath1148 , with the first corresponding to a jump from @xmath193 to @xmath1149 and the second to a jump from @xmath1150 to @xmath1151 .
the right plot shows the three positive roots of @xmath182 .
the two of these roots achieved by @xmath210 are at @xmath1152 and @xmath1153 .
the first corresponds to @xmath1154 and the second to @xmath1155 .
hence as @xmath11 ranges in @xmath180 , @xmath1156 changes sign from positive , to negative , to positive , and finally to negative again.,title="fig:",width=264 ] by corollary [ transfer ] , a.s .
we have a sequence of random matrices satisfying the hypotheses of theorem [ t:2core ] .
if @xmath1134 , then @xmath1157 , and theorem [ t:2core ] shows that the 2-core has @xmath1135 rows .
so from now on suppose that @xmath1062 , so @xmath1136 ( see lemma [ gstarlem ] ) . for the statement ( i )
, note that out of @xmath1158 rows , a proportion @xmath1159 survives , by theorem [ t:2core](ii ) . for ( ii )
, the discussion around ( [ 2coredegs ] ) implies that the proportion of the @xmath6 original vertices whose degree in the @xmath40-core is non - zero is obtained by subtracting from @xmath21 the mass that a po@xmath1160 random variable places on @xmath901 . for ( iii )
, we compare the limits in ( i ) and ( ii ) .
suppose that these limits satisfy @xmath1161 by our assumptions on @xmath11 and @xmath8 , we have @xmath1162 and @xmath1163 , which implies that @xmath1164 and @xmath1165 are both positive
. then we may rewrite ( [ morerows ] ) as @xmath1166 using the definition of @xmath185 . now substituting in @xmath1167 for @xmath11 on the left - hand side of the last display ( given @xmath1168 )
we may rewrite the last inequality as @xmath1169 , where @xmath150 is defined by ( [ psidef ] ) .
similarly , @xmath1170 is equivalent to @xmath1171 if @xmath1172 , then ( [ morerows ] ) holds and the limit in ( i ) in strictly greater than the limit in ( ii ) , which shows that the @xmath40-core eventually has more rows than occupied columns , and vice versa if @xmath1173 ( so that ( [ morecols ] ) holds ) . recall the definition of @xmath146 from ( [ ubardef ] ) .
we know from lemma [ gstarlem](iii ) that @xmath185 has only finitely many discontinuities .
either @xmath146 is a continuity point for @xmath190 ( and hence for @xmath210 ) , or else @xmath1174 with @xmath1175 and no other point of @xmath1069 is in a neighbourhood of @xmath146 . in either case ,
@xmath1140 for @xmath11 in an interval of the form @xmath1176 with @xmath1177 . the final statement in the theorem , about the existence of a hypercycle , follows from lemma [ hypercycle-2core](iii ) .
we next state a result giving an upper bound for @xmath146 .
[ lessthan1 ] suppose that @xmath156=1 $ ] and @xmath1178 < \infty$ ]
. then @xmath1179 . we know from lemma [ gstarlem ] that @xmath161 , so if @xmath1180 there is nothing to prove .
hence we assume @xmath1181 from now on .
first we show that @xmath1182 by the definition ( [ ubardef ] ) of @xmath146 , and the assumption @xmath1181 , if ( [ psicross ] ) fails then there exists @xmath1183 such that @xmath1184 is identically zero on the interval @xmath1185 , and by taking @xmath781 small enough we may assume the interval @xmath1186 contains no discontinuities of @xmath185 .
but then the image @xmath1187 is also an open interval because @xmath185 is continuous and strictly increasing on @xmath1186 .
so we would then have @xmath150 identically zero on @xmath473 , which would contradict the fact that @xmath150 is analytic and non - constant on @xmath180 .
thus ( [ psicross ] ) must hold as asserted .
observe next that every time the 2-core algorithm deletes a row , it has to create at least one column of degree zero , and possibly more .
so the aspect ratio ( i.e. , number of rows divided by number of occupied columns ) is nondecreasing at each step of the algorithm , provided the initial aspect ratio is at least @xmath21 .
hence the aspect ratio of a non - empty 2-core is at least as large as the aspect ratio of the original incidence matrix to which the algorithm is applied , provided the latter is at least @xmath21 .
so if @xmath1188 , then the aspect ratio of the original matrix exceeds @xmath21 for all @xmath6 large enough , and hence so does the aspect ratio of the @xmath40-core , assuming it exists .
suppose that @xmath1189 .
then by ( [ psicross ] ) and the finiteness of @xmath1069 , there exists @xmath1190 such that @xmath1191 .
then , by theorem [ c : alphabar](iii ) , with @xmath1192 , the @xmath40-core has aspect ratio less than @xmath21 for all @xmath6 large enough , which contradicts the previous conclusion that @xmath689 implied the @xmath40-core having limiting aspect ratio greater than @xmath21 .
hence @xmath1179 .
next we give more information on the key functions @xmath165 and @xmath150 , which should clarify the situation in theorem [ c : alphabar](iii ) . by a _ root _ of @xmath150 , we mean any number @xmath1193 with @xmath182 .
[ psiprop ] suppose that @xmath156=1 $ ] and @xmath1178 < \infty$ ]
. then @xmath1194 .
the function @xmath150 has at least one root in @xmath180 , and @xmath165 has at least one local minimum in @xmath180 .
suppose that the following condition holds : * @xmath165 has a single local minimum @xmath1195 in @xmath180 , with @xmath1196 .
then @xmath1195 is the location of the unique local maximum of @xmath150 in @xmath180 , @xmath1197 , and the interval @xmath1198 contains exactly one root of @xmath150 , denoted @xmath181 , which satisfies @xmath1199 .
moreover , @xmath1200 , and @xmath1201 finally , in the fixed row - weight case where where @xmath201 a.s . , condition ( a ) holds , and the unique positive root of @xmath150 is @xmath1202 . an important observation that helps to explain the close connection between the functions @xmath165 and @xmath150 ( apparent in figure [ fig1 ] , for example ) and will also form an ingredient in the proof of proposition [ psiprop ] is the following result .
[ lem : hpsi ] for all @xmath147 , @xmath1203 has the same sign as @xmath1204 , so , in particular , the locations of the local minima of @xmath165 correspond exactly to the locations of the local maxima of @xmath150 in @xmath180 . finally , if @xmath1064<\infty$ ] , then as @xmath1205 we have @xmath1206}{2{\mathbb{e}}[w ] } ( 1 + o(1 ) ) x^2 \log x \nonumber\\ & = 1 - h(1-x ) - x + o(x ) .
\label{hpsi}\end{aligned}\ ] ] differentiating ( [ psidef ] ) , we obtain @xmath1207 on the other hand , from ( [ hdef ] ) , we have that , for @xmath147 , @xmath1208 by comparison with ( [ psidiff1 ] ) . finally , ( [ hpsi ] ) follows from a routine calculation . before completing the proof of proposition [ psiprop ]
, we make some further remarks and present some examples .
the main complication in the interpretation of theorem [ c : alphabar](iii ) is due to the fact that @xmath185 has discontinuities , so @xmath1209 is only a subset of @xmath1210 .
let @xmath1211 by lemma [ gstarlem](ii ) , @xmath1212 is a union of finitely many intervals @xmath1213 where @xmath1214 , and , for each @xmath806 , @xmath1215 for @xmath1070 , and @xmath1216 is a local minimum . recall that @xmath1217 and @xmath184 is increasing for @xmath1062 ( see lemma [ gstarlem ] ) , so @xmath1218 must be increasing on @xmath1212 .
so in fact @xmath1219 .
the ` curve ' @xmath1220 , @xmath1221 is then a ( discontinuous ) trace of @xmath1222 , where @xmath59 runs over @xmath1212 , piecewise continuously on intervals starting at @xmath1223 which , by lemma [ lem : hpsi ] , correspond to local _ maxima _ of @xmath150 .
figures [ fig1 ] and [ fig2 ] give some illustrations of possible behaviour . observe that @xmath1222 is not necessarily decreasing for all @xmath1224 . note that condition ( a ) in proposition [ psiprop ] is not necessary for the sharp transition property ( [ sharp ] ) to hold .
two other relevant conditions are : * @xmath150 has a single root in @xmath180 ; * the global minimum of @xmath165 on @xmath180 is the rightmost local minimum . if @xmath156 = 1 $ ] and @xmath114<\infty$ ] , then @xmath154 as @xmath1225 and as @xmath1226 , so ( a ) @xmath1227 ( c ) , while in the course of the proof of proposition [ psiprop ] below , we show that ( a ) @xmath1227 ( b ) as well .
we mention 3 illustrative examples .
* an example for which conditions ( a ) and ( b ) do not hold but ( c ) does is provided by @xmath1228 , for which @xmath150 has 3 positive roots ( see figure [ fig8 ] ) .
* an example for which condition ( b ) holds but conditions ( a ) and ( c ) do not is @xmath1229 , for which @xmath185 has two discontinuities ( see figure [ fig1 ] ) .
* an example in which none of ( a ) , ( b ) or ( c ) holds and where ( [ sharp ] ) fails is provided by @xmath1143 ( see figure [ fig2 ] ) . for @xmath1230 ( _ left _ ) and @xmath1231 ( _ right _ ) . in the first case the only positive root is @xmath1232 , while in the second case the @xmath506 positive roots are @xmath1233 , @xmath1234 , and @xmath1235 . for the case on the right @xmath1236 and @xmath1237 , and only the root @xmath1238 exceeds this value .
proposition [ psiprop ] gives @xmath1239 for the case on the left and @xmath1240 for the case on the right.,title="fig:",width=264 ] for @xmath1230 ( _ left _ ) and @xmath1231 ( _ right _ ) .
in the first case the only positive root is @xmath1232 , while in the second case the @xmath506 positive roots are @xmath1233 , @xmath1234 , and @xmath1235 . for the case on the right @xmath1236 and @xmath1237 , and only the root @xmath1238 exceeds this value .
proposition [ psiprop ] gives @xmath1239 for the case on the left and @xmath1240 for the case on the right.,title="fig:",width=264 ] first we show that if @xmath170 = 1 $ ] and @xmath114<\infty$ ] , then @xmath150 has at least one root in @xmath180 .
so suppose that there exists an integer @xmath51 for which @xmath1241 = 1 $ ] and @xmath1242 = p > 0 $ ] .
then @xmath1243 as @xmath153
. from ( [ psidiff1 ] ) we have @xmath1244 taking @xmath155 in ( [ psidiff1 ] ) and ( [ psidiff2 ] ) , using @xmath1245 for @xmath1246 , we obtain @xmath1247 since @xmath51 . hence @xmath1248 is a local minimum , and @xmath1249 for @xmath72 small enough . but @xmath158 as @xmath157 , so continuity implies that @xmath150 has at least one root in @xmath180 .
consider the condition ( a ) in the proposition .
suppose that @xmath165 has a unique local minimum located at @xmath1250 , so @xmath1251 .
then by lemma [ lem : hpsi ] , @xmath150 has a unique local maximum at @xmath1195 , and necessarily @xmath1252 . by continuity ( and
rolle s theorem ) it follows that @xmath150 has exactly one root @xmath1253 .
so ( a ) @xmath1227 ( b ) .
moreover , it follows that @xmath1249 for @xmath1254 and @xmath1255 for @xmath1256 . hence in this case @xmath1257 , and the claim ( [ sharp ] ) follows . finally , we show that if @xmath1258 for some @xmath51 , then ( a ) holds . in the case
@xmath1258 we obtain ( cf ( [ psidiff2 ] ) ) @xmath1259 hence @xmath1260 and for @xmath147 we have @xmath1261 if and only if @xmath1262 ( an inflexion point ) .
so , by rolle s theorem , @xmath1263 for at most one @xmath147 , necessarily @xmath1264 , and this must be a local maximum for @xmath150 since @xmath158 as @xmath157 .
another application of rolle s theorem shows that @xmath150 has a single positive root , @xmath251 say , which must be in @xmath1265 . by lemma [ lem : hpsi ]
, the fact that @xmath150 has a single local maximum also shows that @xmath165 has a single local minimum in @xmath180 , which is lies in @xmath1266 .
now we can complete the proof of theorem [ thm3 ] .
the expected number @xmath667 $ ] of null vectors is at least one , and may be large even when @xmath1267 $ ] is small .
nevertheless we can derive bounds on @xmath1268 by studying the asymptotics of @xmath667 $ ] because @xmath1269 = { { { \mathbb{p}}_{\rho_n } } } [ { { \mathcal{n } } } ( n , m ) \geq 2 ] \leq { { { \mathbb{e}}_{\rho_n } } } [ { { \mathcal{n}}}(n , m ) ] -1,\ ] ] by markov s inequality applied to the nonnegative random variable @xmath1270 .
suppose that @xmath1271 . then by ( [ lowbound ] ) with ( [ polybound ] ) , @xmath1272 = o ( n^{-1 } ) $ ] .
it follows that , for any @xmath176 , @xmath1273 \to 0 $ ] .
on the other hand , theorem [ c : alphabar](iii ) implies that there exists @xmath1177 such that for any @xmath1274 , @xmath1275 \to 1 $ ] .
moreover , these results together show that @xmath1276 , and @xmath1179 by proposition [ lessthan1 ] .
to conclude this section , we give the proof of proposition [ threshasym ] .
take @xmath1258 for @xmath51 .
as already mentioned , the asymptotic for @xmath240 is in @xcite .
fix @xmath225 . then , @xmath1277 as @xmath281 .
hence @xmath1278 for all @xmath238 sufficiently large , which by ( [ alphasharp ] ) shows that @xmath1279 .
since @xmath225 was arbitrary , it follows that @xmath1280 .
finally , by proposition [ psiprop ] , ( [ fixedbar ] ) holds . then with ( [ grsim ] ) and repeated taylor expansions we obtain @xmath1281 completing the proof of ( [ starsim ] ) .
the following simple result , on the probability that certain integer - valued random variables take even values , will be used several times .
the formula in lemma [ lem4](i ) for the probability that a binomial variable is even may be found for example in @xcite .
[ lem4 ] let @xmath44 be a @xmath408-valued random variable with probability generating function @xmath1282 $ ]
. then @xmath1283 = \frac{1}{2 } ( 1 + \phi(-1))$ ] .
in particular ( i ) if @xmath1284 , then @xmath1283 = ( 1 + ( 1 - 2p)^n)/2 $ ] ; and ( ii ) if @xmath1285 , then @xmath1286 = { { \mathrm{e}}}^{-\mu } \cosh \mu$ ] . for @xmath1287 , @xmath1288 , yielding the first statement in the lemma . for the @xmath1289 case , @xmath1290 , giving part ( i ) , while in the @xmath1291 case , @xmath1292 , which gives @xmath1286 = ( 1 + { { \mathrm{e}}}^{-2\mu})/2 = { { \mathrm{e}}}^{-\mu } \cosh \mu$ ] .
we saw in lemma [ lem4 ] that the probability that a @xmath1289 random variable is even is @xmath1293 . in this section
we extend this formula to a more complicated multinomial setting and more general congruence conditions modulo @xmath238 .
here is our probabilistic model .
we perform a sequence of @xmath6 independent trials .
each trial is probabilistically identical , and we are interested in the outcome of a trial described in terms of an arbitrary collection of @xmath806 events @xmath1294 .
probabilities @xmath1295 are specified for each of the ` elementary ' events @xmath1296 defined as @xmath1297 \setminus i } a_j\right).\ ] ] for each @xmath1298 $ ] ( here @xmath1299:= \{1,\ldots , k\}$ ] ) .
we assume that @xmath1300 for each @xmath1186 and that @xmath1301 } p(e_i ) = 1 $ ] .
use @xmath1302 to denote the probability measure associated with the model consisting of @xmath6 trials as just described .
let @xmath1303 be the number of occurrences of event @xmath1304 . for @xmath1305 $ ] set @xmath1306 and set @xmath1307,j}$ ]
. also , set @xmath1308 to be the probability for a single trial that an odd number of outcomes @xmath1309 occur , and note that @xmath1310 where @xmath1311 denotes the number of elements of @xmath959 .
the next result gives a general formula for the probability in @xmath6 trials that for each @xmath56 the @xmath1303 falls into a particular congruence class modulo @xmath238 , and a specialization to the event that @xmath1303 is even for each @xmath56 in a specified subset @xmath1186 of @xmath1299 $ ] and @xmath1303 is odd for each @xmath1312 \setminus i$ ] .
for positive integer @xmath238 , define the complex number @xmath1313 ( a complex @xmath238th root of unity ) .
[ parity ] ( i ) let @xmath1314 be an integer and @xmath1315 . then @xmath1316 = r^{-k } \sum_{{{\bf h}}\in \{0,1,\ldots , r-1\}^k } \omega^{- { { \bf t}}\cdot { { \bf h } } } \left ( \sum_{{{\bf g}}\in \{0,1\}^k } \omega^{{{\bf g}}\cdot { { \bf h } } } p_{{{\bf g } } } \right)^n , \label{maineq}\end{aligned}\ ] ] where for @xmath1317 we set @xmath1318 with the event @xmath1319 defined in terms of a single trial , and @xmath1320 and @xmath1321 .
\(ii ) for any @xmath1298 $ ] , @xmath1322 \setminus i } \
{ n_j \notin 2{\mathbb{z}}\ } \right ) \right ] = 2^{-k } \sum_{j \subseteq [ k ] } ( -1)^{|j \cap i| } ( 1 - 2 p_o(j))^n .\end{aligned}\ ] ] part ( ii ) follows from the @xmath50 case of part ( i ) on setting @xmath1323 to be @xmath21 on @xmath1186 and @xmath26 otherwise .
thus we need to prove part ( i ) .
we obtain ( [ maineq ] ) by induction on @xmath6 .
first consider the case @xmath1324 . in this case
the left side of ( [ maineq ] ) is equal to 1 if @xmath1325 for all @xmath56 and is equal to zero otherwise , and the right side is equal to @xmath1326 if @xmath1327 then this expression is 1 .
otherwise , it is zero since if @xmath1328 for some @xmath56 then @xmath1329 thus the inductive hypothesis holds for @xmath1324 .
suppose it holds for some @xmath6 .
in the case of @xmath1330 trials , conditioning on the outcome of the first trial we obtain @xmath1331 = \sum_{{{\bf g}}\in \{0,1\}^k } p_{{{\bf g } } } { \mathbb{p}}_{n } \left [ \cap_{i=1}^k \ { n_i \equiv t_i ~ ( { \mathrm{mod}~}r ) \ } | e_{{\bf g}}\right ] \\ = \sum_{{{\bf g}}\in \{0,1\}^k } p_{{{\bf g } } } { \mathbb{p}}_{n+1 } \left [ \cap_{i=1}^k \ { n_i \equiv t_i - g_i ~ ( { \mathrm{mod}~}r ) \ } \right ] \\ = r^{-k } \sum_{{{\bf g}}\in \{0,1\}^k } p_{{{\bf g } } } \sum_{{{\bf h}}\in \{0,1,\ldots , r-1\}^k } \omega^{- ( { { \bf t}}- { { \bf g } } ) \cdot { { \bf h } } } \left ( \sum_{{{\bf f}}\in \{0,1\}^k } \omega^{{{\bf f}}\cdot { { \bf h } } } p_{{{\bf f } } } \right)^n \\ = r^{-k } \sum_{{{\bf h}}\in \{0,1,\ldots , r-1\}^k } \omega^{- { { \bf t}}\cdot { { \bf h } } } \left ( \sum_{{{\bf g}}\in \{0,1\}^k } \omega^{{{\bf g}}\cdot { { \bf h } } } p_{{{\bf g } } } \right ) \left ( \sum_{{{\bf f}}\in \{0,1\}^k } \omega^{{{\bf f}}\cdot { { \bf h } } } p_{{{\bf f } } } \right)^n \\ = r^{-k } \sum_{{{\bf h}}\in \{0,1,\ldots , r-1\}^k } \omega^{- { { \bf t}}\cdot { { \bf h } } } \left ( \sum_{{{\bf g}}\in \{0,1\}^k } \omega^{{{\bf g}}\cdot
{ { \bf h } } } p_{{{\bf g } } } \right)^{n+1},\end{aligned}\ ] ] which completes the induction .
the next result collects some elementary properties of probability generation functions .
[ pgf ] let @xmath1332 $ ] , @xmath1333 $ ] , for a @xmath408-valued random variable @xmath44 .
then @xmath1334 $ ] , @xmath1335 , and @xmath1336 is infinitely differentiable at least for @xmath1337 ; if @xmath1338 < \infty$ ] then @xmath1339 is continuous in the closed interval @xmath1340 $ ] .
moreover , * suppose that @xmath1341=0 $ ] .
then as @xmath153 , @xmath1342 + o(s^2 ) , ~\textrm{and}~ \phi'(s ) = { \mathbb{p } } [ x=1 ] + o(s ) .\ ] ] * if @xmath1343 < \infty$ ] , then as @xmath153 , @xmath1344 + o(s ) , ~\textrm{and}~ \phi'(1-s ) = { \mathbb{e } } [ x ] + o(1 ) .\ ] ] * for any @xmath122 $ ] , @xmath1345 .
apart perhaps from part ( iii ) , all of the properties stated in the lemma are well known : see for example @xcite . for part ( iii ) , let @xmath122 $ ] .
then @xmath1346 = \phi(s).\end{aligned}\ ] ] we shall use the following bounds on the binomial coefficient @xmath1347 .
let @xmath80 and @xmath1348 .
then @xmath1349 on the other hand , if @xmath1350 , @xmath1351 we apply robbins s refinement of stirling s formula ( see e.g. @xcite ) , which says that for any @xmath1352 , @xmath1353 where @xmath1354 .
this yields the upper bound , for @xmath1352 and @xmath1355 , @xmath1356 where we have used the fact that @xmath1357 since @xmath1358 . by considering separately the cases ( i ) @xmath1359 , and
( ii ) @xmath1360 , using ( [ chooseub ] ) in case ( ii ) , we obtain the first inequality in ( [ chooseub2 ] ) .
the second inequality in ( [ chooseub2 ] ) follows from the fact that @xmath1361 for the lower bound , another application of robbins s bounds yields ( [ chooselb ] ) , where for the @xmath1362 term we have used the fact that @xmath1363 .
the authors thank julian west for pointing out the elementary application of the pigeonhole principle in the proof of lemma [ hypercycle-2core ] , thus avoiding the use of linear algebra .
m. dietzfelbinger , a. goerdt , m. mitzenmacher , a. montanari , r. pagh , and m. rink , tight thresholds for cuckoo hashing via xorsat , _ proc .
37th international colloquium on automata , languages and programming _ , lecture notes in computer science , 2010 , volume 6198/2010 , springer , pp .
213225 .
kovalenko and a.a .
levitskaya , stochastic properties of systems of random linear equations over finite algebraic structures , pp .
6470 in : probabilistic methods in discrete mathematics , v.f .
( eds . ) , tvp / vsp , 1993 . | let @xmath0 be a random @xmath1 matrix with binary entries and i.i.d .
rows . the weight ( i.e. , number of ones ) of a row has a specified probability distribution , with the row chosen uniformly at random given its weight .
let @xmath2 denote the number of left null vectors in @xmath3 for @xmath0 ( including the zero vector ) , where addition is mod 2 .
we take @xmath4 , with @xmath5 , while the weight distribution may vary with @xmath6 but converges weakly to a limiting distribution on @xmath7 ; let @xmath8 denote a variable with this limiting distribution . identifying @xmath0 with a hypergraph on @xmath6 vertices , we define the _ 2-core _ of @xmath0 as the terminal state of an iterative algorithm that deletes every row incident to a column of degree 1 .
we identify two thresholds @xmath9 and @xmath10 , and describe them analytically in terms of the distribution of @xmath8 .
threshold @xmath9 marks the infimum of values of @xmath11 at which @xmath12 $ ] converges to a positive limit , while @xmath10 marks the infimum of values of @xmath11 at which there is a 2-core of non - negligible size compared to @xmath6 having more rows than non - empty columns .
we have @xmath13 , and typically these inequalities are strict ; for example when @xmath14 almost surely , numerics give @xmath15 and @xmath16 ( previous work on this model has mainly been concerned with such cases where @xmath8 is non - random ) .
the threshold of values of @xmath11 for which @xmath17 in probability lies in @xmath18 $ ] and is conjectured to equal @xmath10 .
the random row weight setting gives rise to interesting new phenomena not present in the non - random case that has been the focus of previous work .
_ key words : _ random sparse matrix , null vector , hypercycle , random allocation , xorsat , phase transition , hypergraph core , random equations , large deviations , ehrenfest model _ ams subject classification : _
60c05 ( primary ) 05c65 ; 05c80 ; 15b52 ; 60b20 ; 60f10 ( secondary ) |
the dynamo model proposed by roberts 1972 @xcite has been chosen as the starting point for an experimental demonstration of homogeneous fluid dynamo at the forschungszentrum karlsruhe @xcite . the fluid velocity field considered by roberts which is of particular interest in this context
is given by @xmath0 here a cartesian coordinate system @xmath1 is used .
the flow pattern is sketched in fig .
[ fig : roberts ] .
@xmath2 is the length of the diagonal of a cell in the @xmath3plane and the parameter @xmath4 , which is a constant , determines the @xmath5component of the flow and so the helicity of the velocity field .
roberts has demonstrated that a flow of this kind is capable of dynamo action .
he investigated , however , only magnetic fields which show the same periodicity in @xmath6 and @xmath7 as the flow pattern .
these fields , which we call here harmonic fields " , possess parts which do not depend on @xmath6 and @xmath7 but only on @xmath5 or , in other words , they have infinite wave lengths in the @xmath6 and @xmath7 directions .
as roberts himself pointed out the considered flow allows also non decaying magnetic fields with finite wave lengths in all directions . for a particular case
such fields were investigated by tilgner and busse @xcite , who called them subharmonic " , and in a more general frame by plunian and rdler @xcite . despite the finite dimensions of the karlsruhe experimental device many estimates concerning excitation conditions etc .
have been made on the basis of findings about harmonic magnetic fields .
it is , however , of high interest to compare these results with such derived from results on subharmonic fields . in this paper
we start with the basic equations of the roberts dynamo problem and some general consequences ( section [ robdyn ] ) , present some findings on its harmonic solutions ( section [ harm ] ) and explain a mean field approach to the dynamo problem on that level ( section [ mfappr ] ) .
after that we turn to subharmonic solutions and give some results for them ( section [ sec : sub model ] ) .
we then deal with the karlsruhe experiment , derive in the framework of a mean field approach and under simplifying assumptions on the dynamo module an excitation condition and compare it with a corresponding result of the subharmonic analysis ( section [ sec : applkar ] ) .
finally summarize the main consequences of our findings ( section [ sec : conclusion ] ) .
to discuss the roberts dynamo problem in some detail we consider the induction equation governing the magnetic field @xmath8 , assuming that it applies in all infinite space .
we use its dimensionless form @xmath9 with @xmath10 instead of @xmath11 we have introduced here dimensionless coordinates @xmath12 defined by @xmath13 , instead of @xmath14 the dimensionless velocity @xmath15 defined by @xmath16 , and we measure the time in units of @xmath17 . further @xmath18 is the magnetic reynolds number defined by @xmath19 with @xmath20 being the magnetic diffusivity of the fluid . for a steady flow
as envisaged here we may expect solutions @xmath8 varying like @xmath21 in time , where the real part of @xmath22 is the dimensionless growth rate . in this case an eigenvalue problem for @xmath8 with the eigenvalue parameter @xmath22 occurs .
furthermore , since the flow is @xmath23-independent , @xmath8 can be assumed to possess the form @xmath24 where @xmath25 is a complex vector field independent of @xmath23 , and @xmath26 a dimensionless wave number with respect to the @xmath23-direction .
when inserting ( [ eq : bb ] ) into ( [ inductiondim ] ) we find @xmath27 @xmath28 the @xmath29 and @xmath30 components of ( [ eq : bharm ] ) are equations for @xmath31 and @xmath32 which do not contain @xmath33 .
they constitute the mentioned eigenvalue problem . after solving it
, we can calculate @xmath33 from ( [ eq : divbharm ] ) without any integration .
we may easily conclude from ( [ eq : bharm ] ) that the results for @xmath22 for any @xmath4 can be inferred from those for @xmath34 with the help of the relation @xmath35 as long as we deal with direct solutions of the roberts dynamo problem ( up to section [ sec : sub model ] ) we therefore restrict our attention to the solutions with @xmath34 .
in the discussion of the karlsruhe dynamo experiment ( section [ sec : applkar ] ) we admit also other values of @xmath4 .
as mentioned above , roberts solved the relevant equations only for magnetic fields with the same periods in @xmath29 and @xmath30 as the flow pattern .
we consider first this case only , in which we speak of harmonic solutions "
. then @xmath25 must have the same periodicity in @xmath29 and @xmath30 as the flow pattern . solving the eigenvalue problem defined by ( [ eq : bharm ] ) roberts found growth rates , that is real parts of @xmath22 , in its dependence on @xmath26 for values of @xmath18 up to 64 as shown in fig .
[ fig : p ] .
@c@c@ + & the imaginary parts of @xmath22 proved to be equal to zero ( numerically always close to zero ) , attesting that the dynamo instability is an absolute one .
this implies that the magnetic field geometry is stationary while the intensity in general varies in time .
soward @xcite has shown that in the limit of large @xmath18 the order of the maximum of the dimensionless growth rate , @xmath36 , is given by @xmath37 .
it occurs at a wave number @xmath38 for which @xmath39 .
that is , @xmath40 as @xmath41 .
thus the roberts dynamo proves to be a slow one .
this applies not only in the context of harmonic solutions , for it turns out that other solutions never grow faster than the fastest of the harmonic ones . results for finite @xmath18 obtained in our calculations are given in fig . [ fig : p ] and in table [ tab : maxpkrm ] ; see also @xcite .
note that the quantities @xmath42 and @xmath43 given in table [ tab : maxpkrm ] approach constant values as @xmath18 grows and thus illustrate the mentioned asymptotic laws .
.maximum growth rates @xmath44 , the corresponding wave numbers @xmath45 and the quantities @xmath46 and @xmath47 for various @xmath18 and @xmath48 . [ cols="<,^,^,^,^,^,^,^,^,^,^",options="header " , ]
in order to check a simple mean - field theory of the karlsruhe dynamo experiment we are interested in solutions @xmath8 of the induction equation ( [ inductiondim ] ) with period lengths exceeding those of the flow pattern , which we call subharmonic solutions " .
we consider the case in which the period lengths of @xmath8 are larger by an integer factor @xmath49 than those of the flow pattern .
as mentioned above , this problem has already been investigated by tilgner & busse @xcite for a few special values of @xmath49 and later by plunian and rdler @xcite .
we focus our attention again on the induction equation ( [ inductiondim ] ) governing the magnetic field @xmath8 in all space .
we use again ( [ eq : bb ] ) but consider @xmath50 no longer as a field with the same periodicity in @xmath29 and @xmath30 as the flow pattern .
instead we put @xmath51 , where @xmath52 has now the same periodicity as the flow pattern and @xmath53 and @xmath54 are subharmonic wave numbers in the @xmath29 and @xmath30 directions . in that sense
we look for solutions of the induction equation of the form @xmath55 with @xmath56 being a complex periodic vector field with the same period length in @xmath29 and @xmath30 direction as the flow pattern , @xmath57 the real vector @xmath58 , and @xmath22 again a complex quantity ; for more details see @xcite .
we restrict our attention here to the case @xmath59 .
then the period lengths of the magnetic field @xmath8 are just @xmath49 times that of the flow pattern
. the harmonic solutions @xmath8 discussed above correspond to the limit @xmath60 . inserting ( [ eq : bdef ] ) in ( [ inductiondim ] ) we find @xmath61 @xmath62 the system ( [ eq : b2 ] )
defines an eigenvalue problem with @xmath22 being the eigenvalue parameter in his numerical calculations . ] .
it has been solved numerically .
marginal values of @xmath18 versus @xmath26 for @xmath63 are shown in fig .
[ fig : marginal ] for different values of @xmath64 and again @xmath34 . for @xmath65
there are both a critical @xmath18 and a critical @xmath26 below which dynamo action is not possible . in general @xmath22
is no longer real , that is , we have no longer stationary but moving field structures .
we point out that relation ( [ ki ] ) allows us the calculation of @xmath22 for arbitrary @xmath4 from that for @xmath48 . @c@c@ + &
the essential piece of the karlsruhe dynamo experiment is the dynamo module " , a cylindrical container with both radius and height somewhat less than 1 m , through which liquid sodium is driven by external pumps @xcite . by means of a system of channels with conducting walls , constituting 52 spin generators " , a helical motion is organized .
the flow pattern is similar to that defined by ( [ new1 ] ) .
the 52 spin generators correspond to 26 periodic units of the flow pattern such as the one shown in fig .
[ fig : roberts ] . the arrangement of the pumps allows to vary the parameters @xmath66 , or @xmath18 , and @xmath4 independently from each other . in order to give an estimate for the self excitation condition of
the experimental device a simple mean
field theory has been developed @xcite .
the mean magnetic field @xmath67 defined as above is assumed to satisfy the equations ( [ inducmean ] ) and ( [ eq : electromotive force ] ) inside the dynamo module and to continue in some way in outer space .
of course , in this context @xmath67 can no longer be independent on @xmath29 and @xmath30 , and @xmath68 can no longer have the simple forms ( [ kh09 ] ) , ( [ kh11 ] ) or ( [ kh17 ] ) . relying on some traditional concept
it was assumed that the variations of @xmath67 in space are sufficiently weak so that @xmath68 in a given point can be represented by @xmath67 and its first spatial derivatives in this point .
together with the symmetry properties of the roberts flow this leads to @xmath69 where @xmath70 and @xmath71 are constants depending on @xmath18 and @xmath4 , and @xmath72 is again the unit vector in @xmath23-direction @xcite . as in ( [ kh17 ] )
the term with @xmath73 describes the anisotropic @xmath74effect acting in the @xmath75plane only .
the terms with @xmath76 and @xmath77 can be interpreted by introducing an anisotropic mean - field diffusivity different from the molecular magnetic diffusivity .
finally the term with @xmath71 describes a part of @xmath68 depending on derivatives of @xmath67 which can not be expressed by @xmath78 and therefore not be interpreted as a contribution to a modified diffusivity .
several results have been derived on the dependence of @xmath73 on the fluid flow @xcite , and also such on @xmath79 , @xmath80 and @xmath71 @xcite . for a field @xmath67
not depending on @xmath29 and @xmath30 and having no @xmath23component we have @xmath81 , and the last three terms on the right hand side of ( [ meanemf ] ) can be written in the form @xmath82 . as to be expected , in this special case the structures of ( [ kh17 ] ) and ( [ meanemf ] ) coincide , and we have @xmath83 .
our above remark on the @xmath71term in ( [ meanemf ] ) explains why the interpretation of the @xmath84term in ( [ kh17 ] ) as a contribution to a mean - field diffusivity is not compelling .
the assumption on small variations of @xmath67 in space means in particular that @xmath67 does not change markedly across a spin generator . in that sense
the usage of ( [ meanemf ] ) in a theory of the dynamo module can only be justified for a very large number of spin generators within the module .
quite a few solutions of the equation ( [ inducmean ] ) for @xmath67 , applied to the dynamo module , with @xmath68 according to ( [ meanemf ] ) and various boundary conditions have been calculated @xcite . in most cases ,
however , no other contribution to @xmath68 than the @xmath74effect , that is , only the first term on the right hand side of ( [ meanemf ] ) was taken into account .
contributions with higher than first derivatives of @xmath67 have never been considered . by these and other reasons
a check of the results of the simple mean field theory on a way that avoids the mentioned shortcomings seems very desirable .
for this purpose we deal now with a very simple model of the dynamo module .
we consider no longer a cylindrical module but instead a rectangular dynamo box " with a quadratic basis area in the @xmath75plane and denote the edge lengths of the box in this plane by @xmath85 and its hight by @xmath86 .
thinking of the shape of the real dynamo module we put @xmath87 .
we will study the excitation condition for a mean magnetic field @xmath67 which satisfies the equation ( [ inducmean ] ) and the relation ( [ eq : electromotive force ] ) in all space and is periodic in @xmath29 and @xmath30 with the period length @xmath88 and in @xmath23 with the period length @xmath89 .
this periodicity means that the dynamo box contains just a half wave " of the field @xmath67 . for the sake of simplicity we use ( [ eq : electromotive force ] ) in its reduced form containing no other induction effect than the @xmath74effect
we will then compare this excitation condition with that for a subharmonic @xmath8field whose longest wave lengths show the same periodicity , that is , which fits in the same sense to the dynamo box . in this context
we put @xmath90 so that an area of @xmath91 in the @xmath75plane contains just 100 period units , that is , 200 cells of the flow pattern , consequently the basis area of the dynamo box 50 cells , which have to be compared with the 52 spin generators in the real dynamo module .
@xmath90 means @xmath92 , and with @xmath87 we arrive at @xmath93 . instead of a realistic boundary condition for the dynamo module we use here in fact the condition of periodic continuation of the magnetic fields both on the mean field and the subharmonic level
. such a condition might be in general problematic but seems acceptable for the comparison which we have in mind . from fig .
[ fig : alpha ] we see that the value of @xmath94 for @xmath95 can , except for small @xmath18 , not be inferred from @xmath96 and @xmath97 only .
this suggests that there will be discrepancies between the excitation conditions obtained with a mean field theory which ignores contributions to @xmath68 with higher than first - order spatial derivatives and those derived from the subharmonic analysis .
as already explained we assume for our mean field consideration that equation ( [ inducmean ] ) and the reduced form of ( [ meanemf ] ) , that is , @xmath98 apply in all space with constant @xmath99 .
we may represent @xmath67 as a sum of a poloidal and a toroidal part , @xmath100 with two defining scalars @xmath101 and @xmath102 . inserting this in ( [ inducmean2 ] ) and dropping unimportant constants we find @xmath103 the special periodic solution of ( [ inducmean2 ] ) which we are looking for
is obtained with the ansatz @xmath104 where @xmath105 and @xmath106 are constants , @xmath64 and @xmath26 the parameters specified above and @xmath22 , which will prove to be real , is again the growth rate .
when inserting this in ( [ system ] ) we arrive at two linear homogeneous equations for @xmath105 and @xmath106 .
the requirement that they allow non - trivial solutions leads to @xmath107 growing @xmath67 are possible in the case of the upper sign of the last term if @xmath108 is sufficiently large .
the excitation condition reads @xmath109 in the representations of results on @xmath73 on which we now rely @xcite the latter is given in its original dimension so that it corresponds @xmath110 with our dimensionless @xmath73 .
furthermore these results are given in terms of the two magnetic reynolds numbers @xmath111 and @xmath112 for the flow in the @xmath75-plane and in @xmath23-direction , respectively .
these are connected with our @xmath18 and @xmath4 by @xmath113 the marginal states of the dynamo , in which @xmath67 neither grows nor decays , are given by pairs of @xmath112 and @xmath111 , or by the corresponding neutral curve in the @xmath114diagram .
we may represent the result for arbitrary @xmath64 and @xmath26 by using the modified magnetic reynolds number @xmath115 defined by @xmath116 instead of @xmath111 .
note that @xmath115 is no longer determined by @xmath14 alone but also by @xmath8 .
[ fig : f / k const ] shows a @xmath117diagram in which curve ( a ) gives just the result of our mean field calculation .
clearly dynamo action requires that @xmath115 exceeds a critical value .
it appears , however , to be possible for any @xmath112 if only @xmath115 is sufficiently large .
let us now compare this result with that for a corresponding subharmonic solution @xmath8 of the induction equation . in fig .
[ fig : f / k const ] the curve ( b ) is the neutral one for the subharmonic solution with the values of @xmath64 and @xmath26 specified above . clearly dynamo action requires now not only that @xmath115 exceeds a critical value but also that @xmath112 lies above such a value .
in addition for each given @xmath112 allowing dynamo action the marginal value of @xmath115 derived in the subharmonic approach is higher than that concluded from the mean field approach . in the range of @xmath112 between 1.2 and 2 , which corresponds to the actual situation in the karlsruhe experiment ,
the deviation is larger than 20 % . of course it will become smaller in a comparison with a mean field model which involves also the induction effects connected with first derivatives of @xmath67 indicated in ( [ meanemf ] ) ;
see @xcite . but
even then the mean field approach underestimates the requirements for self excitation .
we have dealt with several aspects of the roberts dynamo problem and derived some results which are of interest for the karlsruhe dynamo experiment .
although a rectangular dynamo box was considered , there are good reasons to assume that the main conclusions apply as well to the real experimental device with a cylindrical dynamo module . in the framework of
the simple mean field theory of the experiment self
excitation seems possible for arbitrary values of the magnetic reynolds number @xmath118 describing the flow perpendicular to the axes of the spin
generators if only the magnetic reynolds number @xmath119 for the axial flow is sufficiently large .
an analysis based on subharmonic solutions revealed that a dynamo is only possible if both @xmath118 and @xmath119 exceed critical values .
apart from this it was found that the simple mean - field theory underestimates the excitation condition of the dynamo .
this discrepancy of the mean - field results with those obtained with subharmonic solutions can not be completely removed by taking into account the effect of the mean field diffusivity .
busse , f.h . ,
mller , u. , stieglitz , r. & tilgner , a. , a two - scale homogeneous dynamo : an extended analytical model and an experimental demonstration under development .
_ magnetohydrodynamics _ * 32 * , 235248 ( 1996 ) .
rdler , k .- h . ,
apstein , e. , rheinhardt , m. & schler , m. , contributions to the theory of the planned karlsruhe dynamo experiment - supplements and corrections .
_ report astrophysical institute potsdam _ ( 1997 ) .
rdler , k .- h . ,
apstein , e. & schler , m. , the @xmath74effect in the karlsruhe dynamo experiment .
3rd int . conf . on `` transfer phenomena in magnetohydrodynamics and electro - conducting flows '' ,
aussois , france , vol .
, 9 - 14 ( 1997 ) .
rdler , k .- h . ,
rheinhardt , m. , apstein , e. , & fuchs , h. , on the mean field theory of the karlsruhe dynamo experiment .
ii . back reaction of the magnetic field on the fluid flow .
_ magnetohydrodynamics , this volume_. | two different approaches to the roberts dynamo problem are considered .
firstly , the equations governing the magnetic field are specified to both harmonic and subharmonic solutions and reduced to matrix eigenvalue problems , which are solved numerically .
secondly , a mean magnetic field is defined by averaging over proper areas , corresponding equations are derived , in which the induction effect of the flow occurs essentially as an anisotropic alpha - effect , and they are solved analytically . in order to check the reliability of the statements on the karlsruhe experiment which have been made on the basis of a mean - field theory , analogous statements are derived for a rectangular dynamo box containing 50 roberts cells , and they are compared with the direct solutions of the eigenvalue problem mentioned .
some shortcomings of the simple mean
field theory are revealed . |
a century of cosmology has led us to a new understanding of the universe . today we know the model at large scales .
_ per aspera
ad astra_. after many years of hypotheses and markets of models we now have the standard cosmological model , yet separated from what we have at small scales the standard model of elementary particles .
both models progressively converge and interact with each other leading us to a joint physical model of the world we are a part of .
the progress in cosmology is ensured by observations .
this creates identification problem .
it is a specific feature of astronomy .
astronomers _ see _ structures unknown to physicists . they can not touch or test them
, they can learn only general properties of observed matters assuming some theoretical extrapolations ( general relativity , atomic physics , etc . ) . on the contrary ,
physicists need experiment to judge things . to understand what astronomers see
, physicists are looking in labs for what is unknown to them , since there is not enough information about the target . in this way
the problem of identification arises . what do astronomers see
they observe structures made of invisible matter , the _ dark matter_. dm does not interact with light , generally - with luminous matter , or baryons .
how is dm observed then ?
through its gravitational influence on visible matter .
fortunately , light is there where dm concentrations are .
1 shows a region of sky in the direction of one of dm _ halos _ , the non - linear dm concentration gravitationally bound in all three directions with total mass @xmath0 .
we see optical galaxies captured by gravitational field of this concentration , x - ray gas residing at the bottom of the gravitational well , and a multi - image of one of the background galaxies that happened to be on the line of sight of the dm halo and was distorted by its gravitational field . .
] we study spatial distribution of dm halos analyzing galaxy catalogs and quasar absorption lines . besides
, the dm surface mass density can be reconstructed via its gravitational week lensing action on numerous background galaxies .
so , there is more than enough independent probes of dark mass inside and beyond dm halos .
we can state that the mean contrast of dm density field is larger than unity at small scale ( @xmath1mpc ) still remaining less than unity at large scale ( @xmath2 mpc ) .
accordingly , we do not find dm halos exceeding @xmath3 .
this is the current dm density field .
we are lucky to have a map of much younger matter density field using cmb anisotropy . that time ( @xmath4 ) the mean density contrast was @xmath5 , and no halos had formed
having these two pictures of cosmic matter distribution at different epochs of its evolution and assuming that only gravity is responsible for such evolution , we can obtain the dm energy - momentum tensor .
what are dm properties ? actually , they are very simple .
dm is non - relativistic _ weakly interacting massive particles _ with cosmological density five times higher than that of baryons .
wimps should be cold ( non - relativistic ) long before the equality epoch to be able to form the structure that we observe today .
owing to such simple properties , dm has straightforwardly affected the development of the universe gravitational potential .
the dm density contrast was increasing in time due to gravitational instability .
baryons , after they decoupled from radiation , were captured into gravitational wells of dm concentrations .
that is why light is there where dm is , although dm particles do not interact with light .
thanks to this remarkable feature of gravitational instability it is possible to study amount , state and distribution of dm in observations ranging from radio to x - ray bands .
contemporary physics does not know particles with dm properties .
it is necessary to go beyond the standard model . but
how and in which direction ?
what should we look for ?
the analysis of large scale structure in the universe has revealed that the amount of non - relativistic dm entering structure is small .
the overall mass density of all particles which have been involved in the process of gravitational clustering , can not exceed @xmath6 of the critical density .
massive neutrino contribute less than few percent to the matter budget . at the same time the characteristics of cmb anisotropy
have evidenced the flat spatial geometry of our universe .
it means that the rest @xmath7 of the critical density should be in the form that takes no part in gravitational clustering .
what are the properties of such a stable medium which is not perturbed by gravitational potential of the structure and remains essentially unclustered ?
theory gives a clear answer to this question the pressure - to - energy ratio of this medium , @xmath8 , should satisfy the following condition : @xmath9 only under this inequality the medium remains lorentz - invariant and invariable both in space and time .
_ dark energy_. this is all we know about de . it is crucial that the process of gravitational instability could be launched in the friedmann universe only if the seed density perturbations were present since the very beginning .
the existence of primordial cosmological perturbations has nothing to do with dm or any other particles .
these are the total density perturbations that were produced by the big bang physics .
thus , another important problem arises , the problem of origin of the seed density perturbations which have developed dynamically into dm structures . these hot topics searching for unknown matter and determining the initial conditions for structure formation
display new physics and are expected to be solved in near future . in this review we dwell upon them .
the observed structure of the universe is a product of start conditions and evolution of matter density field .
up - to - date observational data made it possible to determine characteristics of the density field at different epochs of its development .
it allowed us to separate information about the initial conditions and development conditions , thus giving rise to independent investigations of the early and late universe physics . in modern cosmology the term `` early universe ''
stands for the final period of the inflationary _ big bang stage _ with subsequent transition to hot period of cosmological expansion .
currently we have no model of the early universe as we do not know bbs parameters ( there are only upper bounds , see eq .
( [ eq-12 ] ) ) .
however , we have a well - developed theory of quantum - gravitational generation of the cosmological perturbations . using this theory
, we can derive the spectra of primordial density perturbations and cosmic gravitational waves as functions of cosmological parameters , and constrain the latter if the spectra are known .
the reason why we still have no generally - accepted model of the early universe stems from stable predictions of bbs inflationary paradigm , that are obtained in a wide class of model parameters .
namely , the generated spectra are almost flat , the amplitude of the cosmological gravitational waves is relatively small , etc .
detection of the cosmological gravitational waves would give crucial information about the early universe .
this discovery may come in case the planck experiment succeeds .
our knowledge of the late universe is quite opposite .
we have rather precise model we know the main matter components and cosmological parameters , the evolution of the universe and theory of structure formation .
but we do not understand how the matter components have originated .
the known properties of the visible universe allow us to describe the geometry of both , late and early universe , in the framework of the perturbation theory .
there is a small parameter @xmath10 , the amplitude of cosmological perturbations .
the main tool of geometry is metric tensor . to zeroth order
the universe is friedmannian and described with only one time function the scale factor @xmath11 .
the first order is a bit more complicated .
the metrics perturbations are the sum of three independent modes the scalar one @xmath12 , the vector one @xmath13 , and the tensor one @xmath14 , each of them being described by its spectrum , the function of the wave number @xmath15 .
the scalar mode describes the cosmological density perturbations , the vector mode is responsible for vortical matter motions and the tensor mode presents the gravitational waves . if the first order fields are gaussian then the entire geometry of our universe is described with only four positively defined functions , @xmath11 , @xmath12 , @xmath14 and @xmath13 . currently we know the first two functions in some ranges of their definition .
bbs was a catastrophic process of rapid expansion accompanied by intensive time varying gravitational field . under this gravitational action
the small - scale cosmological perturbations of metric and density were being parametrically born from vacuum fluctuations .
it is very general and fundamental effect of creation of any massless degree of freedom in external coupled non - stationary field .
available observational data confirm the quantum - gravitational origin of seed density perturbations responsible for structure formation in the universe .
it is a good example of the solution of measurability problem in quantum field theory .
the basic properties of the perturbation fields generated according to this mechanism are the following : the gaussian statistics ( random distribution in space ) , the preferred time phase ( `` growing '' branch of evolution ) , the absence of characteristic scales in a wide range of wavelengths , a non - zero amplitude of the gravitational waves .
the latter is crucial for building - up the bbs model as gravitational waves couple the simplest way to the background scale factor .
the development of @xmath16-mode has resulted in formation of galaxies and other astronomical objects .
the cmb anisotropy and polarization have emerged long before under the joint action of all three perturbation modes ( @xmath17 and @xmath18 ) on the photon distribution .
joint analysis of the observational data on galaxy distribution and the cmb anisotropy allowed us to relate @xmath16 and @xmath19 modes . making use of the fact that the sum @xmath20 is known from the cmb anisotropy
, we obtain the upper bound for the vortical and tensor perturbation modes in the visible universe : @xmath21 in case the latter inequality were violated the density perturbation value would not be sufficient to form the observed structure .
the detection of @xmath22 and/or @xmath18 ( e.g. cosmological magnetic field ) will become possible only with increase of observational precision .
let us consider zero order geometry more detailed . table [ tab : table1 ] presents average values of the cosmological parameters obtained from astronomical observations ( with @xmath23 accuracy ) . with these parameters ,
we obtain from the friedmann equations the hubble function , @xmath24 , and its time derivative , @xmath25 : = 10 ^ 61= ( + + 0.7)^1/2 , + = = = , where @xmath26 is the inverse hubble constant , @xmath27gev@xmath28 cm is the planck mass ( or inverse planck scale , hereafter @xmath29 ) .
@xmath30-function relates the hubble size of the universe with its expansion factor .
.basic cosmological parameters .
[ cols= " < , < " , ] one of the versions on agenda the neutralino hypothesis rises from minimal supersymmetry .
this hypothesis can be verified in cern at lhc that will run in 2008 .
the expected mass of these particles is @xmath31 , and their density in our galaxy is a particle per cup of coffee .
dm particles are being searched in many experiments all over the world .
interestingly , the neutralino hypothesis can be independently verified both in underground experiments on elastic scattering and by indirect data on neutralino annihilation in galaxy .
so far the positive signal has been found only in one of the underground detectors ( dama ) , where a season signal of unknown origin has been observed for several years now .
but the range of masses and cross - sections associated with this experiment has not been confirmed in other experiments , which makes reliability and meaning of the results quite questionable .
neutralino give an important possibility of indirect detection by their annihilation gamma - ray flux . during the process of hierarchic clustering these particles
could form mini - halos of the earth masses and characteristic sizes comparable to that of the solar system .
some of these mini - halos could stay intact till now . with high probability
the earth itself is inside one of these halos where the particle density is as much as tens of times higher than the mean halo density .
hence , the probability of both direct and indirect detection of dm in our galaxy gets higher .
availability of so different search techniques gives a solid hope that the physical nature of at one version of dm will soon be verified .
let us consider the first order geometry in more detail .
the effect of the quantum - gravitational generation of massless fields is well - studied .
matter particles can be created with this effect ( see @xcite etc . ) ( although the background radiation photons emerged as a result of the bbs proto - matter decay in the early universe ) .
the gravitational waves @xcite and the density perturbations @xcite are generated in the same way since they are massless fields and their creation is not suppressed by the threshold energy condition . the problem of the vortical perturbation creation is waiting for its researchers .
the theory of the @xmath16 and @xmath22 perturbation modes in the friedmann universe reduces to a quantum - mechanical problem of independent oscillators @xmath32 in the external parametrical field @xmath33 in minkovski space with time coordinate @xmath34 .
the action and the lagrangian of the elementary oscillators depend on their spatial frequency @xmath35 : [ eq-2 ] s_k & = & , + l_k & = & ( q^2-^2q^2 ) .
a prime denotes derivative with respect to time @xmath36 , @xmath37 is the oscillator frequency , @xmath38 is the speed of the perturbation propagation in the vacuum - speed - of - light units ( henceforth , the sub - index @xmath15 for @xmath39 is omitted ) . in the case of the @xmath22 mode
@xmath40 is a transversal and traceless component of the metric tensor , @xmath41 in the case of the @xmath16 mode @xmath42 is a linear superposition of the longitudinal gravitational potential ( the scale factor perturbation ) and the potential of the 3-velocity of medium times the hubble parameter @xcite : @xmath43 where @xmath44 , @xmath45 . as it is seen from eq .
( [ eq-3 ] ) , the field @xmath46 is fundamental , because it is minimally coupled with background metrics and does not depend on matter properties ( in general relativity the speed of gravitational waves is equal to the speed of light ) . on the contrary ,
the relation between @xmath47 and the external field ( [ eq-4 ] ) is more complicated : it includes both derivatives of the scale factor and some matter characteristics ( e.g. the speed of perturbation propagation in the medium ) .
we do not know anything about the proto - matter in the early universe .
there are only general suggestions concerning this problem .
commonly , ideal medium is considered with the energy - momentum tensor depending on the energy density @xmath48 , the pressure @xmath49 , and the 4-velocity @xmath50 . for the @xmath16 mode ,
the 4-velocity is potential and represented as a gradient of the 4-scalar @xmath51 : @xmath52 where a comma denotes the coordinate derivative , and @xmath53 is a normalizing function .
the speed of sound is given by `` equation of state '' and relates comoving perturbations of the pressure and energy density : @xmath54 where @xmath55 , and @xmath56 is the potential of the 3-velocity of medium . in the linear order of the perturbation theory
the ideal medium concept is equivalent to the field concept where the lagrangian density @xmath57 is ascribed to the material field @xmath51 @xcite- @xcite . in the field
approach the speed of the perturbation propagation is found from equation : latexmath:[\[\label{eq-7 } \beta^{-2}=\frac{\partial \ln|\partial l/\partial w order , @xmath38 is a function time .
however , in most models of the early universe one usually assumes @xmath59 ( e.g. at the radiation - dominated stage @xmath60 ) .
the evolution of the elementary oscillators is given by klein - gordon equation : @xmath61 where @xmath62 the solution of eq.(10 ) has two asymptotics : an adiabatic one ( @xmath63 ) when the oscillator freely oscillates with the excitation amplitude being adiabatically damped ( @xmath64 ) , and a parametric one ( @xmath65 ) when the @xmath39 field freezes out ( @xmath66 ) .
the latter conditions in respect to quantum field theory implies a parametrical generation of a pair of particles from the state with an elementary excitation ( see fig .
2 ) . ) . ]
quantitatively , the spectra of the generated perturbations depend on the initial state of the oscillators : @xmath67 where the field operators are given in the parametrical zone ( @xmath68 ) .
the factor @xmath69 in the tensor mode expression is due to two polarizations of gravitational waves .
the state @xmath70 is considered to be a ground state , i.e. it corresponds to the minimal level of the initial oscillator excitation .
this is the basic hypothesis of the big bang theory . in case
the adiabatic zone is there , the ground ( vacuum ) state of the elementary oscillators is unique @xcite .
thus , assuming that the function @xmath71 grows from zero with time ( i.e. the initial adiabatic zone is followed by the parametric one ) and @xmath59 , we obtain a universal and general result for the @xmath14 and @xmath12 spectra : @xmath72 where @xmath73 specifies the moment of creation ( @xmath74 ) . as it is seen from eq .
( [ eq-11 ] ) , the theory does not discriminate the @xmath22 from @xmath16 mode .
it is the value of the factor @xmath30 in the creation period that matters when we relate @xmath22 and @xmath16 . from the observed fact that the @xmath22 mode is small in our universe ( see eq .
( [ eq-1 ] ) ) we obtain the upper bound on the energetic scale of the big bang and on the parameter @xmath30 in the early universe : @xmath75 the latter condition implies that bbs was just inflation ( @xmath76 ) .
we have important information on phases : the fields are generated in certain phase , only the growing evolution branch is parametrically amplified .
let us illustrate it for a scattering problem , with @xmath77 at the _ initial _
( adiabatic ) and _ final _ ( radiation - dominated , @xmath78 ) evolution stages ( see fig .
2 ) . for either of the two above - mentioned stages
the general solution is @xmath79 where the constant operators @xmath80 yield the amplitudes of the `` growing '' and `` decaying '' solutions . in the vacuum state
the initial time phase is arbitrary : @xmath81 .
however , the solution of the evolution equations yields that only the growing branch of the sound perturbations takes advantage at the radiation - dominated stage ) . ] : @xmath82 . by the moment of matter - radiation decoupling at the recombination era
, the radiation spectrum appears modulated with typical phase scales @xmath83 , where @xmath84 is a natural number .
it is these acoustic oscillations that are observed in the spectra of the cmb anisotropy ( see fig .
3 , the highest peak corresponds to @xmath85 ) and the density perturbations , which confirms the quantum - gravitational origin of the @xmath16 mode .
we see , the standard cosmological model can begin as follows .
`` in the beginning was sound .
and the sound was of the big bang '' .
it differs a bit from the scenario described in the bible .
the sound modulation in the density perturbation spectrum is suppressed by the small factor of the baryon fraction in the entire budget of matter density .
this allows one to determine this fraction independently of other cosmological tests .
the oscillation scale itself is an example of the standard ruler that is used to determine cosmological parameters of the universe .
the problem of degeneracy of cosmological parameters reconstructed from observational data hindered scientists from building - up the model of the universe for many years .
but now the acuteness of this problem is looser thanks to many independent and complementary observational tests . to summarize
we can say that in principle the problem of the generation of the primordial cosmological perturbations and the large scale structure of the universe is solved today .
the theory of the quantum - gravitational creation of perturbations in the early universe will be finally confirmed as soon as the t mode is discovered , which is anticipated in the nearest future .
for example , the simplest bbs ( power - law inflation on massive field ) predicts the t mode amplitude to be only as much as 5 times smaller than that of the s mode ( which corresponds to @xmath86 ) @xcite .
modern devices and technologies are quite able to solve the problem of registering such small signals analyzing observational data on the cmb anisotropy and polarization .
nowadays it became possible to separately determine properties of the early and late universe from observational astronomical data .
we understand how the primordial cosmological density perturbations that formed the structure of universe emerged . we know
crucial cosmological parameters on which the standard model of the universe is based , and the latter has no viable rivals .
however , some fundamental questions of the origin of the big bang and of main matter constituents remain unsolved .
observational discovery of the tensor mode of the cosmological perturbations is a key to building - up the model of the early universe . in this domain of our knowledge
we have a clear - cut theory prediction that is already verified in the case of the @xmath16 mode and can be experimentally verified for the @xmath22 mode in the nearest future . giving a long list of hypothetical possibilities where and how to look for dm particles and dm
physics theory has exhausted itself .
now it is experiment s turn .
the current situation reminds great moments in history of science when quarks , @xmath87- and @xmath88-bosons , neutrino oscillations , the cmb anisotropy and polarization were discovered . | observational cosmology is on the verge of new discoveries that will change the essence of our world - view .
the matter concerns origin of initial conditions and physics of dark matter .
[ 1999/12/01 v1.4c il nuovo cimento ] |
it has now been firmly established that the photon can interact strongly , as though it were a hadron . indeed , two of the first physics results to be published by the hera experiments h1 and zeus were the measurements of the total photoproduction cross section @xcite .
these confirmed the cross section of the photon to be of order 100 @xmath0b , close to the area of a typical hadron . of course the total cross section is dominated by peripheral collisions which can not be described in perturbative qcd . in the presence of a hard scale the photon proton cross section factorizes into terms describing the photon and proton structures , and a hard qcd subprocess .
then a perturbative expansion may be applied to determine the subprocess cross section ( for an introduction to the theory of hard photoproduction see michael klasen s contribution to this proceedings ) . with subsequent data sets h1 and zeus measured the photoproduction cross section for events in the perturbatively calculable regime , i.e. with jets of at least @xmath1 gev @xcite .
this cross section is naturally much smaller , of the order of 10 nb , however the hadronic nature of the photon is still prevalent .
the hera experiments clearly established two classes of contribution to the photoproduction of jets : the direct process in which the photon itself participates in the hard subprocess and the resolved process in which the photon fluctuates into a hadronic object and one of its partonic constituents participates in the hard subprocess .
for instance , h1 found an excess of energy in the rear direction , over what was expected for direct photoproduction processes @xcite .
this energy could be attributed to the presence of a photon remnant jet in the resolved photon process .
an unambiguous distinction between resolved and direct processes exists only at leading order ( lo ) . in figures [ fig : dirres](a ) and
( b ) examples of direct diagrams in lo and next - to - leading order ( nlo ) , respectively , are shown .
figure [ fig : dirres](c ) shows an example of a resolved diagram at leading order . clearly , if the outgoing quark line in the nlo direct diagram [ fig : dirres](b ) has small transverse momentum then this process could as well be represented by the lo resolved diagram [ fig : dirres](c ) .
therefore some prescription must be introduced in order to make a well - defined distinction between direct and resolved processes . to this end
the observable quantity @xmath2 has been defined : @xmath3 where @xmath4 and @xmath5 are the jet transverse energy and pseudorapidity respectively , @xmath6 is the fraction of the electron s energy carried by the incoming photon , @xmath7 is the incoming electron energy , and the sum runs over the two highest @xmath4 jets within the @xmath5 acceptance . at leading order @xmath8 for direct processes and @xmath9 for resolved processes .
however , @xmath2 is well defined theoretically to any order of perturbation theory ( provided , of course , that the jet - finding algorithm is well - behaved ) .
therefore , a quantitative confrontation of a measurement of a `` direct '' or `` resolved '' photoproduction cross section with a pqcd calculation can be made provided the separation between the direct and resolved regions is made in terms of @xmath2 . in figure
[ fig : zeus_xgam ] a measured @xmath2 distribution is shown for photoproduction events containing two jets of @xmath10 gev within @xmath11 in the hera lab frame @xcite .
spectrum for events with @xmath12 gev , compared to the herwig 5.9 and the pythia 5.7 monte carlo predictions .
the direct component from the herwig monte carlo is shown separately as the shaded histogram .
only statistical uncertainties are plotted [ fig : zeus_xgam ] ] the data show a peak at high @xmath2 values with a broad tail extending to @xmath13 .
the data are compared with the predictions of two parton shower monte carlo programs , herwig 5.9 @xcite and pythia 5.7 @xcite .
the monte carlo calculations implement the qcd matrix elements at leading order only .
the effect of higher order processes is approximated through initial and final - state parton showers .
the programs differ in the choice of evolution variable for the parton shower calculation and also in the technique chosen to convert the final partonic configuration into colourless hadrons .
both are able to provide a good description of the @xmath2 distribution .
the distribution of the herwig events with lo direct photoproduction subprocesses is shown separately as the shaded histogram . the parton showering and hadronizaton phases smear the @xmath2 values such that this distribution is peaked just below one and can extend to the lowest available @xmath2 values .
nevertheless it is plain to see that a sample of events with @xmath14 is essentially of the direct photoproduction type .
zeus and h1 have published several papers in which direct and resolved photoproduction regions are defined based on the @xmath2 observable @xcite and @xcite to @xcite an important confirmation of the presence of the direct and resolved contributions to photoproduction was provided by the zeus measurement of dijet scattering angles @xcite .
a two jet final state can be completely specified in its centre - of - mass frame ( up to an arbitrary azimuthal rotation ) by the dijet invariant mass , @xmath15 , and the dijet scattering angle , @xmath16 .
direct photoproduction processes should proceed predominantly through quark exchange ( the diagram shown at leading order in figure [ fig : dirres](a ) ) . as
the exchanged parton is of spin @xmath17 , at leading order the dijet scattering angle is distributed according to @xmath18 .
in contrast , resolved processes are dominated by the exchange of the integer spin gluon which gives rise to the steeper angular dependence , @xmath19 .
zeus observed a steeper angular dependence for the events with @xmath20 than for those with @xmath21 , as shown in figure [ fig : zeus_angle ] .
normalized to one at @xmath22 for @xmath23 ( _ black dots _ ) and @xmath24 ( _ open circles _ ) photoproduction . in (
* a * ) , the zeus data are compared to nlo predictions ( _ solid line _ ) and lo predictions ( _ broken line _ ) . in ( * b * ) , the broken line is the pythia distribution and solid line is the herwig distribution .
the inner error bars are the statistical errors , the outer error bars are the sum in quadrature of the statistical and systematic uncertainties ] this is a compelling observation .
an invariant mass cut of @xmath25 gev has been applied to ensure that the @xmath26 gev requirement does not bias the angular distribution .
moreover , the dijet scattering angle is defined in the centre - of - mass frame of the two jets so the fact that the lower @xmath2 events are boosted more in the proton direction is not responsible for the differences observed in the @xmath27 distributions .
the different @xmath28 distributions for the high and low @xmath2 samples are therefore an unambiguous demonstration of the differing underlying qcd subprocess dynamics .
there is a greater contribution from gluon exchange processes contributing to the @xmath29 sample , than there is contributing to the @xmath21 sample .
this is consistent with the expectation that more resolved photon processes contribute to the @xmath29 sample .
thus it has been established that both direct and resolved processes contribute to photoproduction at hera .
the study of these processes is now providing a fruitful forum for the investigation of strong interactions .
photoproduction processes access the physics involved in the structures of the photon and proton , in the dynamics of hard subprocesses , and in the fragmentation and hadronization of the final state partons .
the structure of the photon is probed in deep inelastic @xmath30 experiments at @xmath31 colliders in just the same way as the structure of the proton is probed in @xmath32 interactions at hera .
these measurements of the photon s structure function , @xmath33 , constrain the quark densities at intermediate probing energy scales . however , the gluon density is not directly constrained in these experiments , nor do the quark density constraints extend to the high energy scale region accessible at hera .
it is in these two areas that the hera experiments have concentrated their efforts .
phenomenological anstze for the parton densities of the photon exist .
these generally involve the assumption that at low probing virtualities , the photon undergoes a quantum fluctuation into a vector meson or an unbound @xmath34 state and this provides the photon s structure . a dglap evolution
is then invoked in order to evolve the parton densities to arbitrary scale , using the @xmath30 @xmath35 data as a constraint .
thus , studies of photoproduction sensitive to the photon s structure test fundamental physical assumptions , in addition to providing a means to investigate the universality of photon structure data obtained through different processes . in sections
[ sec : real ] and [ sec : virtual ] the latest results from h1 and zeus pertaining to the structure of the photon are presented .
perturbative qcd governs the behaviour of the partons emerging from the hard subprocess . as the distribution of the jets of hadrons in the final
state bears a close correspondence to the distribution of the underlying partons , measurements may be designed which are sensitive to the subprocess dynamics .
the differing dijet angular distributions for direct and resolved photoproduction have already been discussed .
as the available luminosity delivered by hera has increased , it has become possible to look for unusual dynamical signatures in high @xmath4 dijet processes , and also to investigate the underlying mechanism of three jet production .
studies of the matrix element dynamics are presented in section [ sec : matrix ] .
our knowledge is limited about the physics of hadronization , whereby the partons resulting from a collision are converted to colourless hadrons by the inexorable confinement force of qcd .
the hadronization occurs at low momentum transfers where the qcd coupling is much too strong for a perturbative expansion to be relevant .
however , there are experimental results which have provided some information about this interesting area of physics . for instance
, a universality of the hadronization of quarks is supported by the measurements of jet shapes in @xmath32 and @xmath31 collisions @xcite .
recently , a procedure for measuring jet structure in hadronic collisions has been proposed which is valid for an all - orders calculation in perturbative qcd @xcite . in this way
, a well - defined comparison of theory and data can be undertaken which begins well within the regime where a perturbative approach should be valid and then approaches the mysterious realm of the very strong hadron producing force .
a measurement of jet substructure using this algorithm is presented in section [ sec : structure ] .
a measurement which is sensitive to the gluon density of the photon has been made by the h1 collaboration @xcite .
the gluonic component of the photon dominates at low @xmath36 .
therefore the events must be selected keeping @xmath4 as low as possible and allowing @xmath5 to extend as far forward into the incoming proton direction as possible .
this is a difficult kinematic region experimentally , as the energy - scale uncertainty and angular resolution of the calorimeter are worst here .
there are also theoretical limitations as the contribution from events in which there is a secondary scatter , and the smearing between a partonic and hadronic distribution , increase as @xmath4 lowers .
nevertheless it has been possible to make the measurement with sufficient precision to illustrate the sensitivity of the data to the distribution of gluons in the photon .
figure [ fig : h1_xgam ] shows @xmath37 is essentially the same quantity as has been called @xmath38 . ] for events containing two jets of @xmath26 gev within @xmath39 .
the data are compared with the predictions of the leading order plus parton shower monte carlo program phojet @xcite where the predictions including and neglecting the gluons in the photon are shown separately .
although there is a large systematic .
the data are compared to the prediction of phojet .
the direct photon prediction , and the resolved photon predictions for either quarks are gluons , are shown separately ] uncertainty affecting the measurement , within the phojet model a significant gluonic contribution at low @xmath40 is required by the data assuming the grv - lo @xcite parton densities for the photon and proton .
h1 have gone on to subtract the influence of secondary parton scattering and unfold the data to a leading order effective parton density , using the monte carlo model . from this
the modelled quark density has been subtracted to yield the gluon density as a function of @xmath41 where now @xmath41 refers to the fraction of the momentum of the photon which enters the leading order hard subprocess .
this is necessarily a model dependent result ; however it can serve to provide insight into photon structure .
h1 finds that the gluon density rises as @xmath41 decreases .
the zeus analysis of photon structure has concentrated on limiting the data to that kinematic regime in which perturbative qcd should be applicable , without the need for additional model parameters .
dijet events have been selected with @xmath10 gev within @xmath11 @xcite . the @xmath2 distribution for this selection is shown in figure [ fig : zeus_xgam ] .
the pythia and herwig predictions which are compared with the data in this figure contain no simulation of soft underlying events or of secondary parton scatterings and provide a good description of the data .
also , studies have indicated that hadronization effects in this kinematic regime should be small , at most around 10% .
therefore a strong interpretation of this data within perturbative qcd can be made . in figure
[ fig : zeus_highy ] the cross section for the @xmath10 gev selection , @xmath42 , is shown in bins of @xmath43 .
the kinematic region is restricted to a narrow @xmath6 range , in order to improve the sensitivity to the photon s structure . in bins of @xmath44 .
the filled circles correspond to the entire @xmath38 range while the open circles correspond to events with @xmath45 .
the shaded band indicates the uncertainty related to the energy scale .
the thick error bar indicates the statistical uncertainty and the thin error bar indicates the systematic and statistical uncertainties added in quadrature .
the full , dotted and dashed curves correspond to nlo - qcd calculations , using the grv - ho , gs96-ho and the afg - ho parameterizations for the photon structure , respectively . in ( * d * ) the nlo - qcd results for the cross section when @xmath46 and for a particular parameterization of the photon structure are compared ] the cross section is shown separately for @xmath47 , indicating that the direct process dominates when the jets tend toward the incoming photon direction .
nlo perturbative qcd predictions with the three available photon parton parametrizations @xcite are compared with the data in this figure .
( note that the calculations have been checked by several different groups of theorists as reported in @xcite . )
the predictions underestimate the data in the central rapidity region where experimental and theoretical uncertainties are expected to be particularly small .
as previously mentioned , the parton density of the photon is not well constrained by @xmath30 data at these high energy scales .
therefore it is expected that the parton density of the photon may be underestimated in this kinematic regime in the currently available parton density functions of the photon . in another interesting analysis by h1
the photon remnant jet has been tagged in low @xmath48 events by running a clustering algorithm , requiring exactly four jets in the event and defining the photon remnant jet as that jet closest to the incoming photon direction @xcite .
the @xmath4 of the remnant jet was found to be correlated with the @xmath4 of the highest transverse energy jets .
this points to the presence of the anomalous component of the photon s structure whereby the struck parton arises from a @xmath34 splitting of the photon rather than as a constituent of a fluctuation into a vector meson .
the `` remnant jet '' can then , as in figure [ fig : dirres](b ) , be viewed instead as one of the outgoing partons from a next - to - leading order hard subprocess .
another promising process for constraining the parton density of the photon is prompt photon production , in which an outgoing quark from the hard subprocess is balanced not by a gluon , but by a photon ( see @xcite for a complete discussion of the contributing diagrams ) .
zeus has published a measurement of the cross section for prompt photon production in association with a jet @xcite .
a more inclusive measurement , whereby only the prompt photon is tagged without the jet requirement , is free of complications due to the matching of the jet definition in theory and experiment and relatively free of hadronization corrections .
this then , like high @xmath4 production , is an area in which a strong interpretation can be made from the comparision of the data with the predictions of perturbative qcd .
of course the cross section for prompt photon production is suppressed with respect to jet production due to the smallness of the electromagnetic coupling constant and currently the statistics are limited .
nevertheless the technique of prompt photon identification has been refined and a first comparison with the theory indicates a rough agreement @xcite .
it is expected that as the virtuality of the incoming photon increases , less time will be available for it to develop a complex hadronic structure . to test this assumption ,
h1 has measured the dijet triple differential cross section , @xmath49 where @xmath50 is the average jet transverse energy and @xmath51 is the negative of the square of the momentum transfer at the scattered lepton vertex @xcite . the cross section is presented in figure [ fig : h1_virt ] as a function shown as a function of @xmath52 for different regions of @xmath50 and @xmath51 .
the scale factors applied to the cross sections are indicated .
the error bar shows the quadratic sum of systematic and statistical errors .
also shown is the herwig model where the direct component is shown as the shaded histogram ] of @xmath53 in bins of @xmath54 in the range @xmath55 gev@xmath56 and in bins of @xmath51 in the range @xmath57 gev@xmath56 .
the data are concentrated near @xmath58 with a small tail to lower values .
compared with the data are predictions from the herwig model , where the events with a leading order direct subprocess are shown separately by the shaded histogram . looking at fixed @xmath54 , there is clear evidence for the expected suppression of the resolved processes as @xmath59 increases .
however , wherever @xmath60 , the direct processes alone are insufficient to account for the low @xmath53 events .
thus there is evidence for resolved photon processes , even well into the deep inelastic scattering regime , @xmath61 gev@xmath56 .
h1 have extended the analysis in their publication @xcite and unfolded the data to a leading order effective parton density , @xmath62 , relying on the monte carlo simulation to correct the data for hadronization and higher order effects .
the behaviour of this parton density is consistent with a logarithmic rise with the probing resolution , @xmath63 , where @xmath64 is the transverse momentum of the two outgoing partons at leading order .
this is in contrast with the approximate scaling behaviour which has been observed for hadrons and reflects the presence of the anomalous , or @xmath34 splitting term , unique to the structure of the photon .
the effective parton density is also found to exhibit a dependence on the photon s virtuality @xmath51 which is consistent with the expected logarithmic suppression implemented in the existing virtual photon parton density functions . incorporating an earlier measurement of @xmath65 at @xmath66 with these data , a drop in @xmath65 with @xmath51
is indicated .
( to compare the earlier measurement with this one the data have been evolved using the grv lo parton densities to the same @xmath63 and @xmath36 range . )
the zeus collaboration has measured the dijet cross section @xmath67 for jets of @xmath68 gev in the three different photon virtuality ranges @xmath69 ( @xmath70 gev@xmath56 ) , @xmath71 gev@xmath56 and @xmath72 gev@xmath56 @xcite .
this measurement is complementary to the h1 analysis in that the @xmath69 data are analyzed together with the deep inelastic data at @xmath72 gev@xmath56 in a consistent way .
also , making use of an auxiliary small angle electron tagger , the zeus measurement includes data in the important transition region between photoproduction and deep inelastic scattering , @xmath71 gev@xmath56 . from the @xmath67 measurements , similar conclusions can be drawn as in the h1 analysis :
the population of the @xmath2 distribution is suppressed at low values as @xmath51 increases yet there is evidence for a resolved component of the photon even in the deep inelastic scattering regime , @xmath73 gev@xmath56 . in order to make a precise statement concerning the evolution of the virtual photon parton densities with @xmath51 ,
zeus has measured the ratio of the dijet cross section for @xmath74 to that for @xmath47 in bins of @xmath51 @xcite .
the result is shown in figure [ fig : zeus_virt ] .
the ratio of resolved to direct cross sections is found to fall with @xmath51 .
, as a function of photon virtuality , @xmath51 .
the inner error bar represents the statistical error and the outer the statistical and systematic uncertainties added in quadrature .
the band represents the systematic uncertainty due to the jet energy scale .
also shown are the predictions of herwig for two different choices for photon parton densities : grv for real photons ( _ full line _ ) and sas 1d ( _ dashed line _ ) which includes a suppression of the photon parton density with increasing photon virtuality .
the lepto predictions are shown for @xmath75 gev@xmath56 ( _ dot - dashed line _ ) ] in comparison , this ratio within the herwig model is flat for a photon parton density which does not fall with @xmath51 ( grv lo ) and falling for a photon parton density which does ( sas 1d @xcite ) .
therefore the data indicate that the photon parton density is suppressed with @xmath51 in this lo description . furthermore ,
as @xmath51 increases the data tend toward a leading order prediction which does not include any resolved photon contribution ( lepto 6.5.1 @xcite ) .
the influence of the dominant qcd subprocesses on dijet angular distributions for events with @xmath76 gev has been observed , as already discussed . as the integrated luminosity delivered by hera continues to increase
, it is important to re - measure these distributions in the newly accessible kinematic regimes , in order to check for new contributing processes to dijet production . using the 1995 to 1997 integrated data set zeus has measured the @xmath15 and @xmath28 distributions in the kinematic regime @xmath77 gev and @xmath78 @xcite .
the measurement extends to a dijet invariant mass of @xmath79 gev , and the mass and angular distributions are well described by the predictions of perturbative qcd . for three massless jets , five quantities
are necessary to define the system .
these are defined in terms of the energies , @xmath80 , and momentum three - vectors , @xmath81 , of the jets in the three - jet centre - of - mass frame and @xmath82 , the beam direction .
they are the three - jet invariant mass , @xmath83 ; the energy - sharing quantities @xmath84 and @xmath85 , @xmath86 the cosine of the scattering angle of the highest energy jet with respect to the beam , @xmath87 and @xmath88 , the angle between the plane containing the highest energy jet and the beam and the plane containing the three jets , @xmath89 where the jets are numbered , 3 , 4 and 5 in order of decreasing energy .
zeus has measured the three - jet cross section in the kinematic regime defined by @xmath90 gev , @xmath91 and @xmath92 @xcite .
( in fact , the events have been further required to have at least two jets with @xmath26 gev and a third jet with @xmath1 gev , however these cuts are largely irrelevant because high @xmath4 values are forced by the energy and angular cuts . )
the measured cross section is well described by perturbative qcd calculations both in normalization and in the shapes of the @xmath83 , @xmath93 , @xmath94 and @xmath88 distributions .
the measured angular distributions are of particular interest since they differ markedly from the expectation for three jets distributed evenly over the available phase space , and are therefore sensitive to the underlying physical dynamics .
the @xmath94 distribution , since it is primarily determined by the distribution of the highest energy jet , is similar in three jet production to the distribution of @xmath28 in two jet production .
however , @xmath88 is determined by the orientation of the third , or softest , jet . for orientations
in which the third jet is radiated close to the plane defined by the highest energy jet and the beam , @xmath95 or @xmath96 .
the data are shown in figure [ fig : zeus_psi ] . .
the inner error bar shows the statistical error and the outer error bar shows the quadratic sum of the statistical and systematic uncertainties .
the solid histogram shows the default pythia prediction .
the dashed and dot - dashed histograms show the predictions of herwig and of pythia with colour coherence switched off ] they indicate that configurations in which the third jet is far from the plane containing the highest energy jet and the beam ( @xmath97 ) are suppressed .
the dip at @xmath98 and @xmath96 is caused by a loss of phase space at low angles due to the @xmath99 gev requirement .
taking this into consideration , the data indicate a strong tendency for the three - jet plane to lie close to the plane containing the highest energy jet and the beam . to aid in the development of a mental picture for three jet production ,
the data have been compared to the predictions of the pythia and herwig models .
the hard subprocess is included only at leading order in these models so three - jet events arise from the parton shower phase of the simulation .
parton showers do a remarkably good job of simulating three - jet production as is evident from the agreement of the models with the data in the @xmath88 distribution . within the pythia model it is possible to switch off the simulation of qcd colour coherence . with no simulation of coherence ,
pythia predicts a relatively uniform population of the @xmath88 distribution , as shown in figure [ fig : zeus_psi ] .
colour coherence in the parton shower model is required to describe the observed suppression of large angle radiation .
the @xmath100 jet - finding algorithm @xcite clusters objects into jets based on a distance parameter which is essentially their relative transverse momentum .
subjets may be defined within jets by applying the @xmath100 algorithm to the particles of the jet and counting the subjets as a function of the resolution parameter , @xmath101 . for large values of @xmath101
there is only one subjet , the jet itself , but as @xmath101 decreases more and more subjets are resolved until the subjet multiplicity equals the multiplicity of particles within the jet . the dependence of the average number of subjets , @xmath102 , on @xmath101 has been measured by the zeus collaboration for an inclusive sample of jets with @xmath103 gev @xcite . both the herwig and pythia models provide a good description of the evolution of @xmath102 with @xmath101 .
measurements of @xmath102 for @xmath104 have been performed in four different jet pseudorapidity regions , as presented in figure [ fig : zeus_subjet ] . as a function of @xmath105 .
the error bars show the statistical and systematic uncertainties added in quadrature . for comparison ,
the predictions of pythia including resolved plus direct processes for quark jets ( _ thick dashed line _ ) , gluon jets ( _ thick dotted line _ ) , and all jets ( _ thick solid line _ ) are shown .
the predictions of herwig are displayed with thin lines ] the average number of subjets increases as the jets move toward the incoming proton direction .
this behaviour is well described by the pythia and herwig models . in the models the dominant leading order direct process is @xmath106 while the dominant resolved process is @xmath107 .
therefore , relatively more gluon jets are expected for the more forward boosted resolved photon processes . moreover
, the gluon in the @xmath107 subprocess has a tendency to be the more forward parton , further increasing the gluonic content of forward jets .
the fundamental expectation of qcd that gluons , which have a higher colour charge than quarks , should yield a higher multiplicity of hadrons , is borne out in the models by a higher average subjet multiplicity for gluon jets than for quark jets .
thus the increase of @xmath102 with @xmath5 may be understood to arise from an increasing admixture of gluon jets in the forward direction .
we look forward to an eventual comparison of this data with perturbative qcd calculations .
hard photoproduction events have been used in a variety of analyses in order to further the understanding of the physics of strong interactions . from studies of the structure of the real photon
it has been shown that in a leading order interpretation of the data the gluon density rises as the momentum fraction of the photon accessed in the hard subprocess decreases .
there is also an indication that the quark densities of the real photon may be underestimated at high momentum fractions and for high values of the probing energy scale . a global fit of @xmath30 and @xmath108 measurements should be undertaken to discover whether a parton density can be found which describes all the data .
studies of the structure of the virtual photon have illustrated the expected suppression in the photon s structure as its lifetime decreases .
measurements sensitive to the underlying qcd dynamics have shown that perturbative qcd matrix elements , and models with matrix elements together with parton showers , are successful in explaining the mechanisms of high mass dijet and three - jet production .
a measurement of jet substructure has shown the sensitivity of jets produced in hard photoproduction processes to quark and gluon jet differences and revealed their potential to provide insight into the physics of hadronization .
results from photoproduction at hera have progressed from simple manifestations of the photon s hadronic structure , to detailed investigations of that structure and of qcd in general . as the yearly luminosity deliverable by hera continues to increase , and as the dialogue between theorists and experimentalists is continuously strengthened , one may look forward to an even greater variety and quality of physics result to emerge from the investigation of hard photoproduction at hera . | a pedagogical introduction to the experimental results on hard photoproduction at hera is provided . then the latest results in this field from zeus and h1 are reviewed .
glas - ppe/199918november 1999 |
let @xmath0 be a graph on @xmath4 vertices and @xmath5 be a _ size function _ that assigns to each vertex of @xmath2 a list size .
@xmath1-assignment _ @xmath6 be an assignment of lists of colors to the vertices of @xmath2 such that @xmath7 for all @xmath8 . the graph @xmath2 is said to be _
@xmath1-choosable _ if every @xmath1-assignment is @xmath9-colorable .
a choosable size function is called a _ choice function_. for a choice function @xmath1 , define @xmath10 .
the _ sum choice number _
@xmath11 is the minimum of @xmath12 over all choice functions for @xmath2 .
it is not hard to see that for any graph @xmath2 , @xmath13 .
this upper bound is provided by a greedy coloring , see lemma [ gb ] for a proof of this result . when equality holds in the previous inequality , @xmath2 is said to be _ sc - greedy_. the sum @xmath14 is called the _ greedy bound _ and denoted by @xmath15 , or @xmath16 when @xmath2 is implied .
[ gb ] for any graph @xmath2 , @xmath17 let @xmath18 be an ordering of the vertices .
let @xmath19 . using this ordering , a greedy coloring from arbitrary lists of the prescribed sizes will give a proper coloring for any such list assignment
. observe that list coloring and the list chromatic number , or choice number , @xmath20 are related to sum list coloring and the sum choice number . in sum list coloring ,
the list sizes are allowed to vary and one seeks to minimize the sum of list sizes over all vertices . with list coloring ,
one tries to minimize the sizes of lists @xmath9 of colors assigned to the vertices of a graph such that the graph is @xmath9-colorable .
the choice number @xmath20 is the minimum integer @xmath21 for which a graph @xmath2 is @xmath9-colorable given @xmath22 for all @xmath23 .
it is not hard to see that @xmath24 .
moreover , for some graphs @xmath2 , it is the case that @xmath25 is significantly smaller than @xmath20 . in @xcite , fredi and
kantor showed that there is an infinite family of complete bipartite graphs such that @xmath25 can be bounded while the minimum degree increases without bound .
this is not the case for the list chromatic number @xmath20 .
specifically , it is observed that a choosable function can be found for complete bipartite graphs such that the average list size does not necessarily grow with the average degree . to show that @xmath26 , one must provide a choice function @xmath1 of size @xmath27 for @xmath2 and show that for each size function @xmath28 of size @xmath29 , there is a @xmath28-assignment @xmath9 for which @xmath2 is not @xmath9-colorable .
in other words , if @xmath30 , then @xmath28 is not a choice function for @xmath2 .
this paper is devoted to showing that certain graphs made up of cycles are sc - greedy and computing the sum choice number of all graphs on five vertices .
additionally , we provide examples of certain graphs that illustrate some general results and determine information about the sum choice number of other graphs .
sum list coloring was introduced by isaak @xcite in 2002 .
it is a fairly new topic in graph theory , so there is much to be discovered . for more on sum list coloring see also @xcite . in particular
, @xcite is a survey of all sum list coloring results up to 2007 .
we note here that some of the results in this paper can also be found in @xcite . in the next section
, we will further discuss heinold s survey and present some known results about sum list coloring .
first , we define some of the graphs that will be used throughout this paper .
let @xmath31 denote the path on @xmath4 vertices and @xmath32 denote the cycle on @xmath4 vertices .
the graph @xmath33 is the _ wheel _ on @xmath34 vertices @xmath35 formed by taking the cycle @xmath36 and adding the vertex @xmath37 so that it is adjacent to each of the vertices in the @xmath21-cycle .
the graph @xmath38 is the _ broken wheel _ on @xmath34 vertices @xmath35 which is isomorphic to the graph @xmath39 , the wheel on @xmath40 vertices without an edge between @xmath41 and @xmath42 .
note that this graph is often referred to as the fan graph @xmath43 on @xmath40 vertices .
see figure [ fig : bwandw ] for examples of @xmath33 and @xmath38 .
[ def : prod ] the * cartesian product * @xmath44 is the graph with vertex set @xmath45 where any two vertices @xmath46 and @xmath47 are adjacent in @xmath44 if and only if either ( 1 ) @xmath48 and @xmath49 is adjacent to @xmath50 in @xmath51 , or ( 2 ) @xmath52 and @xmath53 is adjacent to @xmath54 in @xmath2 . the * theta graph * @xmath55 is the union of three internally disjoint paths with @xmath56 internal vertices , respectively , that have the same two distinct end vertices .
the graph @xmath57 , the * @xmath21th power * of a graph @xmath2 , is the graph with the same vertex set as @xmath2 and an edge between vertices @xmath53 and @xmath54 if and only if there is a path of length at most @xmath21 between @xmath53 and @xmath54 in @xmath2 .
we define the following graph that is obtained by laying cycles of arbitrary and varying lengths greater than three end to end so that they share an edge .
a graph @xmath2 is called a * path of @xmath21 cycles * , or * path of cycles * , if @xmath58 where 1 .
each @xmath59 is a cycle of length @xmath60 for @xmath61 , 2 .
@xmath62 for all @xmath63 , 3 .
@xmath64 for all @xmath63 , 4 .
@xmath65 for all @xmath66 5 . if @xmath67 , then @xmath68 for @xmath69 .
@xmath2 can be drawn in the plane so that the weak dual of @xmath2 is a path .
see figure [ fig : cyclepath ] for an example of a path of cycles .
we also define a graph that is obtained by laying cycles of arbitrary and varying lengths greater than three along a special tree - like structure so that they share an edge .
a graph @xmath2 is called a * tree of @xmath21 cycles * , or * tree of cycles * , if @xmath58 where 1 .
each @xmath59 is a cycle of length @xmath60 for @xmath61 2 . for all pairs
@xmath70 , it must be that @xmath71 or @xmath72 for two adjacent vertices @xmath73 3 . if @xmath72 , then @xmath74 for all @xmath75 .
4 . if @xmath76 , then @xmath68 for all @xmath69 unless @xmath72 and @xmath77 . 5 .
@xmath2 can be drawn in the plane so that the weak dual of @xmath2 is a tree .
@xmath2 can be drawn in the plane so that in the dual of @xmath2 , the vertex corresponding to the unbounded face of @xmath2 is adjacent to all other vertices .
see figure [ fig : cycletree ] for an example of a tree of cycles .
note that a path of cycles is a special case of a tree of cycles that occurs when the underlying tree - like structure is a path .
a survey by heinold @xcite , and a more recent unpublished version from 2011 , compiles results about sum list coloring , the sum choice number , and states open problems in the area .
specifically , it contains a summary of graphs that are known to be sc - greedy , as well as other graphs whose sum choice number is known .
these graphs , as well as previous results included in the survey that will be used in this paper , are stated here for convenience .
note that the number of graphs known to be sc - greedy is not large , nor is the number of graphs whose sum choice number is known .
the following graphs are known to be sc - greedy : complete graphs @xcite , paths , cycles , trees , and cycles with pendant paths . also , the petersen graph is sc - greedy , as well as @xmath78 and the theta graph @xmath55 unless @xmath79 and @xmath80 is odd @xcite .
paths of cycles are sc - greedy .
while this result is stated without proof in @xcite and elsewhere , this paper contains a proof of the more general result that certain trees of cycles are sc - greedy . the graph @xmath81 is also sc - greedy @xcite .
additional graphs whose sum choice number is known are @xmath82 @xcite , @xmath83 @xcite , @xmath84 @xcite , @xmath85 @xcite , and @xmath86 @xcite . see table [ tab : sc ] for more information . in most cases , these graphs are not sc - greedy .
the graph @xmath87 is the smallest graph which is not sc - greedy .
note that @xmath87 is a graph on five vertices and all graphs on at most four vertices are sc - greedy .
there exist other graphs on five vertices that are not sc - greedy .
these will be presented in section [ sec : five ] .
.graphs that are not generally sc - greedy whose sum choice number is known @xcite . [ cols="<,^,^,<",options="header " , ] we note here that in the case analyses that follow , specific size functions of size @xmath88 will be considered .
these size functions were determined using sage and are the only such such functions that need to be considered .
this is because for all other size functions @xmath28 of size @xmath88 , @xmath28-assignments that are not list - colorable were found using sage .
the number of distinct colors used in the list assignments found using sage is at most two more than the maximum list size .
also , any vertex labeling referred to in this discussion is as in figure [ fig : g5 ] . +
* the graph @xmath89 is not sc - greedy . *
this graph has greedy bound @xmath90 , we show that @xmath91 .
it will be shown that the size function @xmath1 as shown in figure [ fig : g5_4f ] is a choice function of size @xmath92 for @xmath89 . fix an arbitrary @xmath1-assignment @xmath9 .
we determine information about @xmath9 .
if @xmath93 or @xmath94 , then @xmath89 will be @xmath9-colorable .
this is because , without loss of generality , @xmath95 has a cut - vertex so it is sc - greedy .
so assume @xmath96 and @xmath97 .
+ case 1 : there is an element @xmath98 such that @xmath99 .
assign color @xmath100 to @xmath41 , remove it from @xmath101 , then delete @xmath41 .
figure [ fig : g5_4f1 ] illustrates the list sizes on the remaining vertices .
this graph can be colored from lists of these sizes unless the lists are as illustrated in figure [ fig : g5_4l1 _ ] .
this implies @xmath41 should not be assigned @xmath100 .
it also provides information about all of the lists .
there are two possible distinct list assignments obtained from this situation , see figures [ fig : g5_4l1 ] and [ fig : g5_4l2 ] .
the graph @xmath89 can be colored from both of these list assignments .
+ case 2 : there is an element @xmath102 .
assign color @xmath100 to @xmath103 and @xmath104 , remove it from @xmath105 , @xmath106 and @xmath107 , then delete @xmath103 and @xmath104 .
the remaining vertices can be colored in the following order : @xmath108 .
this shows that @xmath1 is a choice function of size @xmath92 for @xmath89 .
+ it is not hard to verify that a size function of size @xmath109 will never be a choice function for @xmath89 .
thus , @xmath91 .
+ * the graph @xmath110 is not sc - greedy . *
this graph is @xmath111 and has greedy bound @xmath112 , but we show that @xmath113 .
it will be shown that the size function @xmath1 , as illustrated in figure [ fig : g5_8f ] , is a choice function for @xmath110 .
if there is an element @xmath114 , then assign @xmath100 to vertices @xmath103 and @xmath115 .
the remaining vertices can then be colored .
thus , assume @xmath116 .
without loss of generality , let @xmath117 , @xmath118 , and assume @xmath119 .
assign @xmath120 to @xmath115 , then remove @xmath120 from the lists of adjacent vertices , and delete @xmath115 from the graph .
figure [ fig : g5_8f _ ] illustrates the list sizes on the remaining vertices .
this graph can be colored from lists of the indicated sizes , unless the lists are as illustrated in figure [ fig : g5_8l _ ] .
this implies @xmath115 should not be assigned @xmath120 .
however , this indicates what all of the lists will be , and figure [ fig : g5_8l ] shows these lists .
this graph can be colored from the provided @xmath9-assignment .
it is not hard to verify that a size function of size @xmath92 will never be a choice function for @xmath110 .
thus , @xmath121 .
in this section , we present new general results about sum list coloring and provide corresponding examples for illustration .
[ sub ] let @xmath0 be a graph . if there is a @xmath23 for which @xmath122 is not sc - greedy , then @xmath2 is not sc - greedy .
assume @xmath123 and @xmath124 .
let @xmath125 be a choice function of minimum size for the graph @xmath122 which is not sc - greedy . then @xmath126 .
now let @xmath1 be a size function for @xmath2 such that @xmath127 for all @xmath128 and @xmath129 .
then @xmath1 is a choice function for @xmath2 such that @xmath130 and it follows that @xmath2 is not sc - greedy .
this claim generalizes in the form of the following corollary .
[ cor : sub ] if a graph @xmath2 contains an induced subgraph that is not sc - greedy , then @xmath2 is not sc - greedy
. let @xmath51 be an induced subgraph of @xmath2 that is not sc - greedy .
the proof is by induction on @xmath131 .
if @xmath132 , then the result follows from claim [ sub ] . now assume the result holds for all induced subgraphs @xmath133 for which @xmath134 . specifically , it holds for @xmath135 for any @xmath136 . by claim [ sub ] , @xmath135 is not sc - greedy . since @xmath137
, it follows by induction that @xmath2 is not sc - greedy .
one might ask whether all triangulations are sc - greedy . corollary
[ cor : sub ] implies that not all triangulations are sc - greedy because it is possible to add three vertices of degree @xmath138 to @xmath87 to obtain a triangulation , see figure [ fig : indsub ] .
this graph is a triangulation with induced subgraph @xmath87 , which is not sc - greedy .
we may also look at what can happen to the sum choice number of a graph upon addition of an edge .
one might predict that adding an edge to a graph would increase the sum choice number by at most one . however , this is not the case .
the following two facts can be observed : 1 . there exist graphs that differ by an edge that have the same sum choice number .
2 . there exist graphs that differ by an edge , but whose sum choice numbers differ by two .
the first fact can be observed in figure [ fig : edgesame ] which displays two graphs that differ by an edge .
the graph in figure [ fig : g1 ] is a @xmath138-cycle with a pendant path , hence sc - greedy , and the graph in figure [ fig : g2 ] is @xmath87 .
both of these graphs have sum choice number @xmath109 . the second fact can be observed in figure [ fig : edgeplus2 ] which displays two graphs that differ by an edge .
the graph in figure [ fig : g3 ] is @xmath87 , whose sum choice number is @xmath109 , while the graph in figure [ fig : g4 ] is a graph we call @xmath139 whose sum choice number is @xmath90 , see section [ sec : five ] .
some things can be said about edge subdivision and sc - greedy graphs .
there exist graphs that are sc - greedy for which it is possible to subdivide an edge and obtain a graph that is not sc - greedy .
see figure [ fig : subnosc ] for an example .
an edge of @xmath140 , which is sc - greedy , can be subdivided to obtain @xmath87 , which is not sc - greedy .
there exists graphs that are not sc - greedy for which it is possible to subdivide an edge and obtain a graph that is sc - greedy
. see figure [ fig : subsc ] for an example . an edge of @xmath87 , which is not sc - greedy ,
can be subdivided to obtain @xmath141 , which is sc - greedy .
the above two claims imply the following : 1 .
there exist graphs that are not sc - greedy with minors that are sc - greedy .
2 . there exist graphs that are sc - greedy with minors that are not sc - greedy .
a graph @xmath2 is * minimally not sc - greedy * if for all @xmath142 , @xmath143 $ ] is sc - greedy , but @xmath2 is not sc - greedy .
two examples of minimally not sc - greedy graphs are @xmath87 and @xmath144 .
two more examples are the graphs @xmath89 ( see figure [ fig : g5_4 ] ) and @xmath110 ( see figure [ fig : g5_8 ] ) , which will be presented later .
[ quest : minnscg ] if @xmath2 is minimally not sc - greedy , does it follow that @xmath11 is equal to @xmath145 ?
note that for the examples of minimally not sc - greedy graphs mentioned above , the answer to this question is yes .
+ assume @xmath2 is a minimally not sc - greedy graph for which @xmath146 .
then , there is a non - simple choice function @xmath28 for @xmath2 of size @xmath147 by lemma [ greedy ] .
additionally , @xmath148 for all @xmath23 .
this implies the following : @xmath149 this must be true for all @xmath23 , which implies that @xmath150 .
however , the graphs @xmath87 , @xmath144 , and @xmath89 are all minimally not sc - greedy graphs of minimum degree @xmath151 .
thus , if @xmath2 is a minimally not sc - greedy graph for which @xmath152 , then @xmath153 . note that @xmath110 is a minimally not sc - greedy graph of minimum degree @xmath154 .
lemma [ greedy ] implies that if a graph @xmath2 is minimally not sc - greedy , then ( 1 ) @xmath122 is sc - greedy for all @xmath23 and ( 2 ) there is a non - simple choice function @xmath1 for @xmath2 such that @xmath155 .
hence , the question remains if there is a non - simple choice function @xmath1 for @xmath2 such that @xmath156 .
does such a graph @xmath2 exist ? this would provide an answer to question [ quest : minnscg ] .
heinold @xcite explored the broken wheel @xmath38 with respect to sum choosability .
in particular , he showed that @xmath157 is not sc - greedy .
this provides some very useful information .
first , it implies that not all outerplanar graphs are sc - greedy .
it also implies that @xmath38 will not be sc - greedy for all @xmath158 since such graphs will have @xmath157 as an induced subgraph .
it was also shown that @xmath38 is sc - greedy for all @xmath159 .
it may be observed here that @xmath157 is minimally not sc - greedy .
heinold also showed that there exist @xmath21 for which the gap between @xmath160 and @xmath161 is arbitrarily large .
while it is known which broken wheels are sc - greedy , it is not known what @xmath161 is in general .
heinold guessed that @xmath162 .
see @xcite for more details .
his dissertation also establishes many techniques that could be used to determine @xmath161 .
as far as we know , the determination of @xmath161 remains an open problem .
however , the results mentioned above can be used to obtain information about the sum choosability of wheels @xmath33 .
since @xmath157 is not sc - greedy , it follows that @xmath33 is not sc - greedy for all @xmath163 since all of these graphs will have @xmath157 as an induced subgraph .
thus , in determining which wheels are sc - greedy , it remains to examine @xmath33 for @xmath164 . for small values of @xmath21 ,
the result follows quickly .
the graph @xmath165 is isomorphic to @xmath166 , and thus is sc - greedy .
it was shown earlier that @xmath111 is not sc - greedy .
the classification would be completed by looking at @xmath33 for @xmath167 . determine whether or not @xmath33 is sc - greedy for @xmath167 .
we summarize this information in the following theorem .
let @xmath38 be a broken wheel and @xmath33 be a wheel .
the following is known : 1 .
if @xmath159 , then @xmath38 is sc - greedy @xcite , 2 . if @xmath158 , then @xmath38 is not sc - greedy @xcite , 3 .
if @xmath168 , then @xmath33 is sc - greedy , and 4 .
if @xmath169 or @xmath163 , then @xmath33 is not sc - greedy .
some general observations may be made about what remains to be done .
first , for any @xmath21 , removing an arbitrary vertex of @xmath33 will yield either a broken wheel or a cycle .
thus , for @xmath167 , the graph @xmath170 is sc - greedy for all @xmath171 . to determine whether or not @xmath33 is sc - greedy , it must be determined whether or not there exists a non - simple choice function for @xmath33 of size @xmath172 .
in this section we show that trees of cycles are sc - greedy . more specifically , we prove that paths of cycles are sc - greedy and the result for trees of cycles follows as a corollary .
let @xmath2 be a path of @xmath21 cycles . recall that @xmath2 can be embedded in the plane so that the weak dual of @xmath2 is a path of length @xmath21 .
for @xmath63 , let @xmath173 be the subgraph of @xmath2 induced by the vertices of @xmath174 , let @xmath175 be the subgraph of @xmath2 induced by the vertices of @xmath176 , and let @xmath177 $ ] .
note that the greedy bound for @xmath2 is equal to @xmath178 .
[ thm : cyclepaths ] paths of cycles are sc - greedy .
the proof is by induction on @xmath21 .
the result clearly holds for @xmath179 , as cycles are known to be sc - greedy .
now assume paths of at most @xmath27 cycles are sc - greedy for all @xmath180 .
it will be shown that the result holds for @xmath21 .
assume @xmath1 is a minimal choice function for a graph @xmath2 which is a path of @xmath21 cycles .
the proof requires a case analysis on @xmath181 for @xmath182 .
observe here that @xmath183 as @xmath184 is isomorphic to @xmath185 and @xmath186 .
assume first that @xmath187 .
now , in @xmath188 the vertex @xmath189 will be a cut - vertex that splits @xmath188 into two shorter paths of cycles , perhaps with pendant paths attached to them . by induction ,
shorter paths of cycles are sc - greedy , hence attaching any pendant paths will also yield an sc - greedy graph .
thus , @xmath188 is sc - greedy .
+ furthermore , assume next there is a @xmath190 for which there exists @xmath191 such that @xmath192 , then in @xmath193 there is a cut - vertex that splits @xmath193 into two shorter paths of cycles , perhaps with pendant paths attached to them . by induction , shorter paths of cycles are sc - greedy , hence attaching any pendant paths will also yield an sc - greedy graph .
thus , @xmath193 is sc - greedy .
+ it follows that @xmath122 is sc - greedy for all @xmath23 .
thus , to show that @xmath2 is sc - greedy , it remains to show there does not exist a non - simple size function of size @xmath88 for @xmath2 .
so we assume that @xmath194 for all @xmath23 for the remainder of the proof .
+ case 1 : @xmath195 for all @xmath63 .
+ recall that @xmath196 for all @xmath197 .
so @xmath198 and the result follows .
+ case 2 : there is a @xmath190 such that @xmath199 .
+ assume @xmath200 . by induction
, @xmath201 so it follows that @xmath202 in fact , the above inequalities must be equalities so that @xmath203 holds .
this allows for an @xmath1-assignment @xmath9 for which @xmath2 is not @xmath9-colorable to be defined as follows : let @xmath204 be a size function for @xmath205 such that @xmath206 and let @xmath207 be a size function for @xmath208 such that @xmath209 thus @xmath210 and @xmath211 , implying @xmath205 and @xmath208 are not @xmath204- and @xmath207-choosable , respectively .
there are @xmath204- and @xmath207-assignments @xmath212 and @xmath213 , respectively , for which @xmath205 and @xmath208 can not be list - colored .
let @xmath9 be an @xmath1-assignment for @xmath2 defined as follows : @xmath214 where @xmath215 and @xmath216 do not appear in any of the lists @xmath212 and @xmath213 , @xmath217 on @xmath205 and @xmath218 on @xmath208 , except append @xmath219 and @xmath215 to the lists of neighbors of @xmath220 and @xmath221 in @xmath222 and @xmath223 , respectively .
if @xmath224 is a proper @xmath9-coloring of @xmath2 , then either @xmath225 and @xmath226 or @xmath227 and @xmath228 . by the construction of @xmath9 , this coloring will not provide a proper coloring of either @xmath222 or @xmath223 , a contradiction .
+ therefore , paths of cycles are sc - greedy .
we now observe that theorem [ thm : cyclepaths ] extends to trees of cycles .
let @xmath2 be a tree of cycles . recall that @xmath2 can be embedded in the plane so that the weak dual ,
call it @xmath229 , of @xmath2 is a tree on @xmath21 vertices .
. then @xmath231 . in particular , the number of pairs of cycles in @xmath2 that share vertices is equal to the number of edges in the weal dual of @xmath2 .
this allows us to compute the greedy bound of @xmath2 as @xmath232 it then follows as a corollary to theorem [ thm : cyclepaths ] that trees of cycles are sc - greedy : [ cor : cycletrees ] trees of cycles are sc - greedy .
this result follows from theorem [ thm : cyclepaths ] and its proof because the argument and case analysis is applied to the intersection of @xmath59 with @xmath233 and the same properties will hold .
this leads us to ask the following questions : are cycles of cycles sc - greedy ?
are paths of cliques sc - greedy ?
if two sc - greedy graphs are joined by an edge , is the resulting graph sc - greedy ?
the authors wish to thank steve butler for his contributions , including the work he did in sage to help with the case analysis for graphs on five vertices .
erds , p. , rubin , a. l. , taylor , h. : choosability in graphs .
proceeding of the west coast conference on combinatorics , graph theory and computing , congress numer .
27 , utiliitas math . , 125157 ( 1980 ) fredi , z. , kantor , i. : list coloring with distinct list sizes , the case of complete bipartite graphs .
european conference on combinatorics , graph theory and applications , electron .
notes discret . math .
34 , 323327 ( 2009 ) | let @xmath0 be a graph and let @xmath1 be a function that assigns list sizes to the vertices of @xmath2 .
it is said that @xmath2 is @xmath1-choosable if for every assignment of lists of colors to the vertices of @xmath2 for which the list sizes agree with @xmath1 , there exists a proper coloring of @xmath2 from the lists .
the sum choice number is the minimum of the sum of list sizes for @xmath1 over all choosable functions @xmath1 for @xmath2 .
the sum choice number of a graph is always at most the sum @xmath3 .
when the sum choice number of @xmath2 is equal to this upper bound , @xmath2 is said to be sc - greedy . in this paper , we determine the sum choice number of all graphs on five vertices , show that trees of cycles are sc - greedy , and present some new general results about sum list coloring . |
the standard presentation of artin s
@xmath8 strand braid group @xmath0 is @xmath28 we denote by @xmath29 the monoid with the above presentation .
an @xmath8 strand _ braid word _ is a word on the @xmath30 letters @xmath31 .
we say that a braid word @xmath4 is _ positive _ ( _ resp .
_ _ negative _ ) if no letter @xmath32 ( _ resp . _
@xmath33 ) occurs in @xmath4 .
we say that @xmath4 is _ positive negative _ if @xmath4 consists of positive letters followed by negative letters , _
i.e. _ , if @xmath4 can be expressed as @xmath34 with @xmath35 positive .
the operations we study here are the following transformations on braid words : let @xmath36 be braid words .
@xmath15 we say that @xmath4 is _ right reversible _ to @xmath37 , denoted @xmath38 , if one can transform @xmath4 to @xmath37 by ( iteratively ) replacing some subword @xmath39 with @xmath40 ( case @xmath41 ) , or with @xmath42 ( case @xmath43 ) , or with @xmath44 ( the empty word , case @xmath45 ) .
@xmath17 symmetrically , we say that @xmath4 is _ left reversible _ to @xmath37 , denoted @xmath46 , if @xmath37 is obtained by ( iteratively ) replacing some subword @xmath47 with @xmath48 ( case @xmath41 ) , with @xmath49 ( case @xmath43 ) , or with @xmath44 ( case @xmath45 ) .
@xmath16 we say that @xmath4 and @xmath37 are _ monotonously equivalent _ , denoted @xmath50 , if @xmath37 is obtained from @xmath4 by ( iteratively ) replacing some subword @xmath51 with @xmath52 ( case @xmath53 ) , or some subword @xmath54 with @xmath55 ( case @xmath56 ) .
it is clear that reversing and monotone equivalence transforms a braid word into an equivalent word , _
i.e. _ , one that represents the same element of the braid group .
observe that the above transformations never introduce trivial pairs of the form @xmath57 or @xmath58 .
so , typically , for a braid word @xmath4 to be reversible to the empty word @xmath44 is _ a priori _ a stronger condition than just being equivalent to @xmath44 , as one is allowed to introduce no @xmath59 or @xmath58 in order to transform @xmath4 into @xmath44 . clearly , the words that can not be transformed using right reversing are the positive negative words . the key result about braid word reversing is as follows : [ p : basic ] @xcite
let @xmath4 be an @xmath8 strand braid word of length @xmath2 .
then there exists a unique positive negative word @xmath37 such that @xmath4 is right reversible to @xmath37 .
moreover , the length of @xmath37 is at most @xmath9 , with @xmath60 .
proposition [ p : basic ] is a consequence of garside s result that common right multiples exist in braid monoids @xcite and of general properties of word reversing @xcite guaranteeing that , for all positive words @xmath61 , the existence of positive words @xmath62 satisfying @xmath63 is equivalent to the existence of a common right multiple for the elements represented by @xmath64 and @xmath65 . in the current paper , we shall only use the following result : [ l : lcm ] @xcite assume that @xmath66 are equivalent positive braid words and , similarly , that @xmath67 are equivalent positive braid words .
let @xmath68 be the positive words satisfying @xmath63 and @xmath69 .
then @xmath70 and @xmath71 are equivalent , and so are @xmath72 and @xmath73 .
the previous results imply that right reversing solves the word problem of the braid monoid and of the braid group , in one and two passes respectively .
indeed lemma [ l : lcm ] implies that two positive braid words @xmath35 represent the same element of the braid monoid if and only if @xmath74 is right reversible to the empty word , and that an arbitrary braid word @xmath4 represents @xmath75 in the braid group if and only if it is right reversible to some positive negative word @xmath76 such that @xmath74 is right reversible to the empty word .
the last step is equivalent to @xmath76 being left reversible to the empty word .
so a braid word @xmath4 represents @xmath75 if and only if the empty word can be obtained from @xmath4 using left and right reversing .
proposition [ p : basic ] says nothing about the words one obtains using both left and right reversing .
the trivial example @xmath77 shows that , starting from @xmath78 , we can produce words of arbitrary length using left and right reversing , since the initial word is a proper factor of the final word .
hence , whenever both left and reversing are involved , restricting to freely reduced words , _
i.e. _ , containing no pattern @xmath59 or @xmath58 , is a minimal requirement if one is to expect bounded length .
we define _ reduced right reversing _ , denoted @xmath79 , to be the variant of right reversing in which a free reduction is performed after each reversing step . reduced left reverving and monotone equivalence are defined similarly .
question [ q : main ] asks in particular whether the words obtained from a given word using reduced reversing and monotone equivalence have a bounded length .
we now establish proposition [ p : lrm ] , which provides a negative answer . to improve readability ,
we adopt a convention of @xcite , using @xmath80 for @xmath81 , and @xmath82 for @xmath83 .
for instance , becomes @xmath84 .
[ proof of proposition [ p : lrm ] ] ( figure [ f : spheres ] ) we find ( the underlined subwords are those we transform ) : @xmath85 and , inductively , @xmath86 transforms into @xmath87 for each @xmath88 as the the words above never finish with the letter @xmath89 .
[ htb ] ( 126,75)(0 , 0 ) ( 0,0 ) using reversing and monotone equivalence ; all words are traced on the fragment of the cayley graph corresponding to the divisors of @xmath90 , _
i.e. _ , on the @xmath5-permutohedron , which , topologically , is a sphere ; the initial path is pushed around the sphere so as to make a loop around the grey facet on the rear ; each other facet is crossed once.,title="fig : " ] note that in the previous counter - example not only the final words , but even all intermediate words are freely reduced .
now we see that these words still involve the commuting pattern @xmath91 , _
i.e. _ , @xmath92 . we shall show now that even such semi - trivial patterns can be avoided .
we say that a braid word is _ strongly reduced _ if it is freely reduced and , in addition , contains no subword of the form @xmath93 with @xmath94 and @xmath41 .
we define _ strongly reduced _ right reversing to be the variant in which a full reduction is performed after each reversing step . in the above definition , strongly reducing a word means iteratively replacing each subword of the form @xmath93 with the corresponding letter @xmath95 .
this is easily seen to lead in finitely many steps to a strongly reduced word . the latter need not be unique , but the various words so obtained are equivalent via commutation relations .
[ p : lrm+ ] starting from @xmath96 , one can derive using strongly reduced left and right reversing and monotone equivalence arbitrary long ( strongly reduced ) words . using @xmath97 , @xmath98 and @xmath99 for the strongly reduced versions of @xmath100 , @xmath101 , and @xmath102 ,
we find @xmath103 @xmath104 @xmath105 @xmath106 @xmath107 @xmath108 @xmath109 @xmath110 @xmath111 @xmath112 @xmath113 @xmath114 @xmath115 @xmath116 @xmath117 @xmath118 @xmath119 @xmath120 @xmath121 @xmath122 @xmath123 @xmath124 @xmath125 @xmath126 @xmath127 @xmath128 @xmath129 @xmath130 @xmath131 .
+ the latter word is @xmath132 .
now we turn to positive results , and establish some upper bounds for the length of the words that can be constructed using reversing and monotone equivalence . in this section ,
we consider the case of right reversing alone .
proposition [ p : basic ] provides an upper bound on the length of the final , _ i.e. _ , positive negative , word that can be obtained from a word @xmath4 , but it gives no bound for the intermediate words .
this is what proposition [ p : bound]@xmath15 does . in order to prove the result
, we need some auxiliary notions . first , as usual , we associate with each @xmath8 strand braid word @xmath4 the braid diagram obtained by concatenating the elementary diagrams for the successive letters of @xmath4 , and the diagram for @xmath33 is ( 55,9)(0 , 0 ) ( 9.5,0 ) ( 9,7)@xmath75 ( 29,7)@xmath133 ( 33,7)@xmath134 ( 54,7)@xmath8 ( 0,3)@xmath33 : ( 16,3)@xmath135 ( 46,3)@xmath135 an @xmath8 strand braid diagram can be seen as the projection on @xmath136 of a 3d - figure consisting of @xmath8 non - intersecting curves . ( figure [ f : simple ] ) a braid word @xmath4 is said to be _ layered _ if the associated diagram can be realized as the projection of a 3d - figure in which each strand lives in some vertical plane .
[ htb ] ( 67,33)(0 , 0 ) ( 0,0 ) is layered : the strands of the associated diagram live in parallel vertical planes , title="fig : " ] ( 19,0)@xmath137 ( 29,4)@xmath138 ( 39,8)@xmath139 ( 59,16)@xmath140 it is well - known that , if @xmath64 is an @xmath8 strand positive word , then @xmath64 is layered if and only if @xmath64 is _ simple _ , _ i.e. _ , it represents a divisor of garside s fundamental braid @xmath141 in the monoid @xmath29 .
[ l : simpleclosed ] ( i ) if @xmath64 and @xmath65 are layered positive words , then @xmath74 is layered .
\(ii ) if @xmath4 is a layered word , then every word obtained from @xmath4 using reversing or monotone equivalence is still layered .
@xmath15 if @xmath64 is a positive layered word , then the diagram of @xmath64 can be realized so that the @xmath133-th strand ,
_ i.e. _ , the strand that starts at position @xmath133 , lives in the plane @xmath142 .
thus @xmath143 can be realized so that the strand finishing at position @xmath133 lives in @xmath144 , and @xmath65 can be realized so that the strand starting at position @xmath133 lives in the same plane .
hence the two diagrams can be concatenated without contradicting layeredness . for @xmath17 , it suffices to check that each elementary transformation introduces no obstruction to the hypothesis that the strands live in a vertical plane .
the case of commutation relations is trivial .
the case of right reversing is illustrated in figure [ f : closure ] ; the cases of left reversing and monotone equivalence are similar .
[ htb ] ( 53,24)(0 , 0 ) ( 0,0 ) occurs in a layered word , then , necessarily , the strand @xmath145 lies in the front plane , while @xmath146 lies in the back plane ; then the pattern @xmath147 can be realised in the same planes without level obstruction .
the case of @xmath148 is similar.,title="fig : " ] ( 0,17)@xmath149 ( 8,17)@xmath146 ( 16,17)@xmath145 ( 36,22)@xmath149 ( 44,22)@xmath146 ( 52,22)@xmath145 ( 25,10)@xmath100 in the braid diagram associated with a layered word @xmath4 , there is a well - defined rear strand , _
i.e. _ , the strand that lives in the plane @xmath150 with minimal @xmath145 : to avoid ambiguity , we choose the leftmost strand in that plane if there are several ones this makes sense as the strands living in a given plane may not intersect . for @xmath4 a layered braid word ,
we denote by @xmath151 the braid word that encodes the diagram obtained from the diagram of @xmath4 by deleting the rear strand .
[ l : del ] if @xmath4 is a layered braid word , then @xmath38 implies @xmath152 . once again , it is sufficient to consider the possible elementary transformations . up to a translation of indices ,
the only non - trivial cases are @xmath153 and @xmath154 , which both reduce to @xmath155 when the rear strand is deleted .
observe that the assumption that the removed strand is the rear ( or the front ) one is necessary : if we remove the middle strand in @xmath156 , we obtain @xmath44 on the left , and @xmath157 on the right .
however , @xmath158 fails .
a braid word is said to be _ elementary _ if it is a subword of a word obtained by right reversing from a word of the form @xmath74 , with @xmath35 layered positive ( or , equivalently , simple ) words . by lemma [ l : simpleclosed ] ,
every elementary word is layered , but the converse is not true : @xmath157 is layered , but not elementary .
indeed , when @xmath74 is right reversible to @xmath159 , then the braids represented by @xmath70 and @xmath72 have no common right divisor in the braid monoid .
[ lengthelem ] the length of an @xmath8 strand elementary braid word is at most @xmath160 .
assume that @xmath4 is right reversible to @xmath37 in one step .
then the crossings in the diagram encoded by @xmath37 are not exactly the same as the crossings in the diagram encoded by @xmath4 , but we can define a notion of inheritance : for instance , in figure [ f : closure ] , we say that the crossing of the strands @xmath149 and @xmath146 in the right figure is the heir of the crossing of these strands in the left figure .
then it is easy to check that , in each case , the crossings in @xmath37 are the heirs of the crossings of @xmath4 , except that two new crossings may appear ( _ e.g. _ , crossings of @xmath146 and @xmath145 in figure [ f : closure ] ) , or two crossings may vanish ( when a free reduction is performed ) .
let us consider a right reversing sequence @xmath161 , , @xmath162 , _ i.e. _ , we assume that @xmath163 is right reversible to @xmath164 in one step for each @xmath88 .
we define the _ total number of crossings _
@xmath165 in this sequence as follows : each crossing in @xmath161 contributes @xmath75 to @xmath165 , and so does every new crossing that appears in some @xmath163 , even if it subsequently vanishes ; on the other hand , the contribution to @xmath165 of a crossing that is the heir of a previously existing crossing is @xmath166 .
so @xmath165 is the sum of the number of crossings of all the terms in the sequence , up to inheritance .
we claim that @xmath167 yields an upper bound for the total number of crossings in a right reversing sequence starting with a ( layered ) word of the form @xmath74 with @xmath35 simple @xmath8 strand braid words .
then , in particular , @xmath168 is an upper bound for the length of each braid word occurring in such a reversing sequence , and , therefore , for the length of every @xmath8 strand elementary braid word . for @xmath169 ,
the only sequence to consider is @xmath170 , so @xmath171 is indeed a valid upper bound .
assume @xmath172 , and let @xmath4 be an @xmath8 strand elementary word .
by hypothesis there is a finite sequence of words @xmath173 , @xmath174 , , @xmath175 such that @xmath35 are positive layered words and each word @xmath176 is right reversible to @xmath177 in one step . by lemma [ l : simpleclosed ] , all the words @xmath163 are layered . by lemma [ l : del ] , the words @xmath178 , , @xmath179 also form a right reversing sequence .
moreover , @xmath178 , _
i.e. _ , @xmath180 , is a word of the form @xmath181 , where @xmath182 are layered positive @xmath183 strand words .
so each word @xmath184 is elementary , and , by induction hypothesis , the total number of crossings ( up to inheritance ) in the sequence @xmath178 , , @xmath184 is bounded above by @xmath185 .
now let us reintroduce the rear strand in the initial word and count how many crossings it can create in ( the diagram associated with ) @xmath176 .
first , in @xmath64 and @xmath65 , which are positive , the rear strand may cross each other strand at most once , so it creates at most @xmath186 crossings .
then , the reversing steps may create new crossings between the front stands and the rear strand .
however , we claim that at most @xmath187 such crossings can be created during the sequence of right reversings .
indeed , during each such reversing , the rear strand moves behind one crossing of the remaining strands , from left to right , and in the process it creates two new crossings ( figure [ f : crossing ] ) . as the total number of crossings not involving the rear strand is at most
@xmath185 , this puts the desired bound on the number of new reversings . in summary , we obtain @xmath188 .
now we calculate @xmath189 this completes the proof of the lemma .
[ htb ] ( 75,21)(0 , 0 ) ( 0,0 ) ( 13,10)@xmath100 ( 58.5,10)@xmath100 we remark that the proof of lemma [ lengthelem ] would go through even if we allowed not only right reversings , but also commutation relations like @xmath190 or even @xmath191 .
now it remains to decompose arbitrary words into products of elementary words .
we recall that a layered positive word is the same as a simple word , in the sense of garside , namely a positive word representing a divisor of @xmath192 . from now on
, we shall be dealing with positive words only , and therefore use the word `` simple '' rather than `` layered '' .
[ l : decomposition ] assume @xmath193 , where @xmath194 are simple positive braid words , and @xmath195 .
then every word @xmath37 obtained from @xmath4 using right reversing can be written as the product of at most @xmath2 elementary words .
first we associate with every right reversing sequence @xmath196 , @xmath197 , a planar oriented graph whose edges are labeled by @xmath33 s .
this graph , which will be called a _ reversing diagram _
, is analogous to a van kampen diagram , and it is constructed inductively as follows ( figure [ f : reversingdiagram ] ) .
first we associate with @xmath196 a path shaped like an ascending staircase by reading @xmath196 from left to right and iteratively appending a horizontal right - oriented edge labeled @xmath33 for each letter @xmath33 , and a vertical down - oriented edge labeled @xmath33 for each letter @xmath32 .
assume that the fragment corresponding to @xmath198 has been constructed and its right side is a path labeled @xmath199 .
by definition , the word @xmath200 is obtained from @xmath199 by replacing some subword @xmath201 with the unique word @xmath202 such that @xmath203 is a relation of the considered presentation .
the involved subword @xmath201 corresponds to some top - left oriented corner in the diagram , and we complete the diagram and transform this corner into a square by adding horizontal edges labelled @xmath64 and vertical edges labelled @xmath65 , following the scheme : ( 56,13)(0,0 ) ( 4,2.5 ) ( 0,6.5)@xmath33 ( 7,12.5)@xmath204 ( 40,6.5)@xmath33 ( 47,12.5)@xmath204 ( 47,1)@xmath65 ( 52,6.5)@xmath64 ( 14,6.5)completed into . [ htb ] ( 108,38)(0 , 0 ) ( 4,2.5 ) : one draws a zigzag path labelled by the initial word , and , then , one iteratively fills the open top - left corners using the braid relations.,title="fig : " ] ( 7,0)@xmath205 ( 15,0)@xmath22 ( 23,0)@xmath206 ( 31,0)@xmath205 ( 70,0)@xmath207 ( 86,0)@xmath208 ( 28,13)@xmath206 ( 11,21.5)@xmath205 ( 28,21.5)@xmath206 ( 71,21.5)@xmath205 ( 88,21.5)@xmath206 ( 28,37.5)@xmath206 ( 88,37.5)@xmath206 ( 0,11)@xmath22 ( 22,7)@xmath205 ( 22,15)@xmath22 ( 22,27)@xmath22 ( 38,9)@xmath205 ( 38,5)@xmath206 ( 38,15)@xmath22 ( 38,27)@xmath22 ( 60,11)@xmath22 ( 82,11)@xmath209 ( 82,27)@xmath22 ( 98,11)@xmath210 ( 98,27)@xmath22 the next step is to observe that each right reversing graph starting with @xmath4 , in particular the maximal one , _
i.e. _ , the one that finishes with a positive negative word , admits a rectangular spine , which will be called the _ right reversing grid _ of @xmath4 .
assume that @xmath4 contains @xmath211 positive letters and @xmath212 negative ones .
we first assume in addition that @xmath4 is a negative positive word .
we define two sequences of simple words @xmath213 for @xmath214 and @xmath215 by setting @xmath216 to be the @xmath133th letter in the inverse of the negative part of @xmath4 for @xmath217 , and to be the empty word @xmath44 for @xmath218 , and by defining @xmath219 to be the @xmath220th letter in the positive part of @xmath4 for @xmath221 , and to be @xmath44 for @xmath222 .
then we inductively define @xmath223 by @xmath224 .
we notice that these words are indeed simple . in this way
, we obtain a grid , which is a fragment of the complete reversing diagram associated with @xmath4 ( figure [ f : grid ] ) .
if @xmath4 is not negative positive , then the construction is similar , except that the word @xmath4 need not correspond to a top left corner , and the top
left corner of the rectangular grid may be missing ( as in figure [ f : reversingdiagram ] ) .
[ htb ] ( 66,44)(0 , 0 ) ( 6,2.2 ) into a product of elementary words ( here delimited by white dots ) ; in the case of the word @xmath225 of figure [ f : reversingdiagram ] , we have @xmath226 , @xmath227 , @xmath228 , and , for instance , @xmath229 and @xmath230.,title="fig : " ] ( 0,9)@xmath231 ( 0,35)@xmath232 ( 21,35)@xmath233 ( 61,9)@xmath234 ( 61,35)@xmath235 ( 11,0)@xmath236 ( 24,0)@xmath237 ( 50,0)@xmath238 ( 11,44)@xmath239 ( 24,44)@xmath240 ( 50,44)@xmath241 ( 11,26.7)@xmath242 ( 24,26.7)@xmath243 ( 50,26.7)@xmath244 ( 9,4.5)@xmath245 ( 14,9)@xmath246 ( 25,8)@xmath247 ( 50,33)@xmath248 ( 61,40)@xmath249 the point now is that every word @xmath37 obtained from @xmath4 using right reversing labels a path from the bottom - left corner to the top - right corner in the right reversing diagram of @xmath4 . as the reversing grid partitions this diagram into squares , we can attribute a square of the grid to each letter in @xmath37 ; we take the convention that , when a letter @xmath33 corresponds to the vertical common edge between two squares of the grid , it is attached to the rightmost square , and that , similarly , a horizontal edge belongs , in case of doubt , to the bottom square .
as the path labelled @xmath4 contains only right - oriented edges and top - oriented edges actually bottom - oriented edges that are crossed in the wrong direction the only possibility after a letter attached to the @xmath250-square is a letter attached to @xmath251 with @xmath252 or @xmath75 .
it follows that the path associated with @xmath37 visits at most @xmath2 squares , and that @xmath37 can be decomposed into a product @xmath253 , where @xmath254 consists of all letters attached to the @xmath88th visited square . it remains to see that each word @xmath254 is elementary .
now an easy induction shows that , if @xmath255 is a fragment of @xmath37 lying in the @xmath256-square , then there exist positive words @xmath257 such that @xmath258 is right reversible to @xmath259 , where @xmath260 and @xmath261 are @xmath262 according to the sides of the square through which the path enters and exits the square .
since the words @xmath263 and @xmath264 are simple , the words @xmath259 and , therefore , @xmath255 are elementary by definition .
thus @xmath37 is the product of @xmath2 elementary words .
we can now conclude as for the length of the words obtained using right reversing .
[ proof of proposition [ p : bound]@xmath15 ] let @xmath4 be an @xmath8 strand braid word of length @xmath2 .
we can write it @xmath266 where each @xmath163 is a single letter @xmath33 and @xmath267 is @xmath262 . by lemma [ l : decomposition ] ,
every word @xmath37 obtained from @xmath4 by right reversing is a product of at most @xmath2 elementary @xmath8 strand words . by lemma [ lengthelem
] , each of these words has length @xmath268 at most .
when monotone equivalence enters the picture , the previous argument fails : simple factors may be changed , and lemma [ l : decomposition ] does not extend .
for instance , the word @xmath269 is a product of two simple words , namely @xmath270 , but a monotone equivalence transforms it to @xmath271 , which can not be decomposed better than @xmath272 , a product of three simple words .
so new arguments are needed . by proposition [ p : basic ] , every braid word @xmath4 is right reversible to a unique positive negative word @xmath76 .
we shall denote @xmath65 by @xmath273 ( the right numerator ) and @xmath64 by @xmath274 ( the right denominator ) .
so @xmath275 always holds .
we first consider the case of positive negative words .
[ proof of proposition [ p : bound]@xmath17 ] let @xmath4 be an @xmath8 strand braid word of length @xmath2 . by proposition [ p : basic ] , @xmath4 is right reversible to @xmath277 , and the latter word has length at most @xmath278 . in order to prove the expected result ,
it is enough to prove the following : if @xmath37 is any positive negative word ( _ i.e. _ , a word satisfying @xmath279 ) which can be obtained from @xmath4 by right reversing and monotone equivalence , then @xmath280 is equivalent to @xmath273 , and @xmath281 is equivalent to @xmath274 ; so , in particular , they have the same length . for an induction , it is enough to assume that only one monotone equivalence is used in the transformation of @xmath4 into @xmath37 .
let us display this monotone equivalence .
the hypothesis is that there exist words @xmath282 and equivalent positive words @xmath283 satisfying @xmath284 , and @xmath285the case when @xmath286 and @xmath287 are equivalent negative words would be treated similarly . proposition [ p : basic ] states in particular that the order of reversing steps does not matter for the positive negative word finally obtained , so @xmath288 holds .
let us compare the reversing processes from @xmath289 to @xmath277 and that from @xmath290 to @xmath37 .
( 84,23)(0 , 0 ) ( 0,2.5 ) is replaced with the equivalent word @xmath287 , the sequel of the reversing changes , but the words are replaced with equivalent words , title="fig : " ] ( 4,0)@xmath72 ( 14,0)@xmath291 ( 24,0)@xmath292 ( 5.5,6)@xmath174 ( 11,7)@xmath70 ( 21,7)@xmath293 ( 31,17)@xmath294 ( 31,7)@xmath295 ( 25.5,16)@xmath296 ( 14,15)@xmath286 ( 24,11)@xmath297 ( 5,19)@xmath4 ( 54,0)@xmath72 ( 64,0)@xmath298 ( 74,0)@xmath299 ( 55.5,6)@xmath174 ( 61,7)@xmath70 ( 71,7)@xmath300 ( 81,17)@xmath294 ( 81,7)@xmath301 ( 75.5,16)@xmath296 ( 64,15)@xmath287 ( 74,11)@xmath297 ( 55,19)@xmath4 let us introduce positive words @xmath72 , @xmath70 , @xmath297 , @xmath294 , @xmath291 , @xmath293 , @xmath292 , @xmath295 satisfying @xmath302 , @xmath303 , @xmath304 , and , finally , @xmath305 . then , by construction , we have @xmath306 and @xmath307 ( figure
[ f : comparison ] ) .
when we replace @xmath286 with @xmath287 , we obtain new positive words @xmath308 satisfying @xmath309 , and lemma [ l : lcm ] guarantees that @xmath300 is equivalent to @xmath293 and @xmath298 is equivalent to @xmath291 .
then , we have @xmath310 for some @xmath311 , and lemma [ l : lcm ] guarantees that @xmath301 is equivalent to @xmath295 and @xmath299 is equivalent to @xmath292 .
we conclude that @xmath280 , which is @xmath312 , is equivalent to @xmath313 , _ i.e. _ , to @xmath273 , and that @xmath281 , which is @xmath314 , is equivalent to @xmath315 , _ i.e. _ , to @xmath274 .
once again , the previous argument only deals with the final , positive negative words obtained using right reversing and monotone equivalence , and it says nothing about the length of the intermediate words . in order to prove proposition [ p : bound ] @xmath16
, we need a new argument . in the sequel , we use @xmath316 and @xmath317 for the unique positive words satisfying @xmath318 ( the left numerator and denominator ) .
we recall that a word is called simple if it is positive and represents a divisor of @xmath141 .
[ l : decomposition2 ] let @xmath4 be a word containing @xmath212 negative letters and @xmath211 positive letters
. then @xmath273 and @xmath316 are the products of at most @xmath211 simple words , and @xmath274 and @xmath317 are the products of at most @xmath212 simple words . as in the proof of lemma
[ l : decomposition ] , consider the right reversing grid of @xmath4 .
it has height @xmath212 and width @xmath211 and all arrows wear simple labels .
so do in particular the bottom and right sides .
this means that @xmath273 is the product of at most @xmath211 simple words , and , similarly , @xmath274 is the product of at most @xmath212 simple words .
the argument is symmetric for left reversing .
[ proof of proposition [ p : bound]@xmath16 ] let @xmath4 be an arbitrary @xmath8 strand braid word of length @xmath2 .
then @xmath4 is right reversible to @xmath277 , and , symmetrically , it is left reversible to @xmath320 . for @xmath64 a braid word , we shall denote by @xmath321 the braid represented by @xmath64 .
now we define @xmath322 to be the restriction of the cayley graph of @xmath0 to the divisors of @xmath323 in @xmath29 , _ i.e. _ , @xmath322 is a finite graph , containing precisely those vertices that lie on some geodesic path from @xmath75 to @xmath323these paths all have the same length , since they correspond to positive words equivalent to @xmath324 . by lemma [ l : decomposition2 ]
, the word @xmath325 is the product of at most @xmath2 simple words , hence its length is at most @xmath326 .
let @xmath327 be a vertex of the graph @xmath322 .
we say that a braid word @xmath64 is _ traced _ in @xmath322 from @xmath327 if there exists a path labelled @xmath64 starting at @xmath327 in @xmath322 , _ i.e. _ , we can read all letters of @xmath64 successively without leaving @xmath322
. then it is proved in @xcite that the word @xmath4 itself is traced from @xmath328 in @xmath322 , and that the family of all words traced from a fixed vertex in @xmath322 is closed under right and left reversing , and under monotone equivalence .
therefore , every word @xmath37 that can be derived from @xmath4 using reversing and monotone equivalence is traced from @xmath328 in @xmath322 .
now , we attribute a weight to every edge @xmath260 in @xmath322 , namely the integer @xmath329 , where @xmath261 is the distance from the source vertex of @xmath260 to the final vertex of @xmath322 , and @xmath329 is the @xmath261th fibonacci number : @xmath330 , and @xmath331 for @xmath332 ( figure [ f : weight ] ) .
finally we define the weight of a path in @xmath322 to be the sum of the weight of its edges .
( 90,28)(0 , 0 ) ( 0,0),title="fig : " ] ( 2.5,26)@xmath333 ( 15,26)@xmath135 ( 55,26)@xmath135 ( 27.5,26)@xmath334 ( 33.5,26)@xmath335 ( 41.5,26)@xmath336 ( 67,26)@xmath337 ( -3,14)@xmath75 ( 74,14)@xmath323 then we observe that the weight can not increase when right reversing or monotone equivalence is performed . indeed , reversing @xmath201 to @xmath338 ( with @xmath43 ) replaces two edges contributing say @xmath339 to the total weight with four edges contributing @xmath340 , _ i.e. _ , @xmath339 again . similarly , reversing @xmath201 to @xmath40 ( with @xmath341 ) diminishes the contribution to the weight from @xmath339 to @xmath342 , and deleting @xmath58 diminishes it by @xmath339 . finally , replacing some positive ( _ resp . _
negative ) subword with an equivalent positive ( _ resp . _ negative ) word preserves the weight . as each letter in a braid word contributes at least @xmath75 in the weight , we deduce that the length of any word obtained from @xmath4 using right reversing and monotone equivalence is bounded above by the weight of @xmath4 .
the latter is the sum of @xmath2 fibonacci numbers between @xmath337 and @xmath333 , where @xmath343 is at most @xmath344 .
one easily checks that the worst case is when @xmath4 consists of @xmath345 negative letters and @xmath345 positive letters with weights @xmath333 , @xmath346 , , @xmath347 . using the very rough estimate @xmath348 for @xmath349 we obtain an upper bound of @xmath350 on the weight of @xmath4 .
the previous results leave open all questions about simultaneous left and right reversing
. we shall conclude with a conjecture which does appear to be true it is confirmed by extensive computer experiments and which would still imply a linear upper bound on the space complexity of the handle reduction algorithm .
[ c : newconj ] let @xmath4 be an @xmath8 strand braid word of length @xmath2 .
let @xmath37 be another such word which is obtained from @xmath4 by a sequence of reduced commutation relations and reduced word reversings .
then @xmath37 is of length at most @xmath9 , where @xmath1 is a constant depending only on @xmath8 .
we have no good guess , however , what the constant @xmath1 should be
. we do know that it is not @xmath353 , as the @xmath5 strand braid word @xmath354 of length 8 can be transformed into a word of length 52 ( which is larger than @xmath355 ) . in order to prove conjecture [ c : newconj ]
, it might be useful to consider bestvina s product structure on the flag complex @xmath356 which is closely related to the cayley graph of @xmath0 @xcite .
bestvina showed that there is a natural homeomorphism @xmath357 , where @xmath358 is another complex which satisfies a certain weak non - positive curvature condition , and is conjectured to be @xmath359 .
we think now of a braid word as a path in @xmath356 , and of our transformations of braid words as deformations of the path that preserve its endpoints . then banning positive equivalences of reidemeister iii
type amounts to forbidding the most obvious way of deforming a path in the @xmath358-direction .
in other words , applying only commutation relations and word reversings means deforming the path mainly in the @xmath360-direction . | it has been conjectured that in a braid group , or more generally in a garside group , applying any sequence of monotone equivalences and word reversings can increase the length of a word by at most a linear factor depending on the group presentation only .
we give a counter - example to this conjecture , but , on the other hand , we establish length upper bounds for the case when only right reversing is involved .
we also state a new conjecture which would , like the above one , imply that the space complexity of the handle reduction algorithm is linear .
this paper was motivated by attempts to estimate the complexity of the handle reduction algorithm in braid groups @xcite , _ via _ a detailed study of _ word reversings_. word reversing is a general combinatorial method for investigating monoids and groups specified by explicit presentations @xcite . in good cases , typically in the case of braid groups @xcite and , more generally , garside groups @xcite , it provides algorithmic solutions to the word problem , as well as an efficient way for proving properties such as cancellativity or existence of least common multiples in the monoid or quadratic isoperimetric inequalities in the group
. however , many natural questions about word reversing remain open , even in the basic case of the standard presentation of artin s braid group @xmath0 .
there are two types of word reversing , namely the left and the right one . in the case of @xmath0 and
, more generally , in the case of artin tits groups of finite coxeter type , garside s theory implies that every sequence of right reversings must terminate , and it gives an upper bound on the length of the final word thus obtained ; however , it says nothing about the length of the _ intermediate _ words and about many related questions .
also , very little is known about what happens when both the left and the right types are used in one reversing sequence . in particular , we raised @xcite [ q : main ] does there exist a constant @xmath1 such that the length of every freely reduced braid word obtained from a length @xmath2 word by using left and right reversing plus monotone equivalence
precise definitions are given below is bounded above by @xmath3 ? a positive answer would have implied a linear upper bound on the space complexity of the handle reduction algorithm in braid groups , and indeed a positive answer was carelessly proposed as a conjecture in @xcite .
the aim of this paper is , on the one hand , to answer question [ q : main ] in the negative , by proving [ p : lrm ] let @xmath4 be the @xmath5 strand braid word @xmath6
. then arbitrarily long freely reduced words can be obtained from @xmath4 using left and right reversing and monotone equivalence .
in fact , the result of proposition [ p : lrm ] can even be strengthened by requiring that all involved words contain no commuting pattern like @xmath7 .
on the other hand , we shall establish some positive results , namely : [ p : bound ] let @xmath4 be an @xmath8 strand braid word of length @xmath2 .
+ ( i ) every word obtained from @xmath4 using right reversing has length at most @xmath9 , with @xmath10 . + ( ii ) every positive negative word obtained from @xmath4 using right reversing and monotone equivalence has length at most @xmath11 , with @xmath12 . + ( iii ) every word obtained from @xmath4 using right reversing and monotone equivalence has length at most @xmath13 , with @xmath14 .
the upper bounds of proposition [ p : bound]@xmath15 and [ p : bound]@xmath16 are certainly not optimal , but they seem to be the first ones in this direction . as for proposition [ p : bound]@xmath17 , we notice that @xmath18 has to grow at least linearly with @xmath8 , as right reversing the word @xmath19 leads to a positive negative word of length @xmath20 . at the end of the paper we shall propose an alternative conjecture which does appear to be true , and which would still imply a linear bound on the space complexity of the handle reduction algorithm . before giving the technical definitions ,
we explain in some more detail the connection of our results with the handle reduction algorithm @xcite and @xmath21-definite forms of braids .
it is known that every braid word is equivalent _ modulo _ the braid relations to a @xmath22-definite word , _
i.e. _ a word in which at least one of the letters @xmath23 does not occur .
this fact is one of the two key points in the construction of a canonical ordering on braids @xcite .
handle reduction is a combinatorial method that solves the isotopy problem of braids and produces @xmath22-definite forms .
although extremely efficient in practice , the method remains partly mysterious and its exact complexity is unknown : the only upper bound proved so far is exponential , very far from statistical evidence .
even more frustrating is the lack of control on the length of the words appearing in the process : the only proved result is an exponential upper bound , while all experiments indicate that their length is bounded by @xmath9 , where @xmath2 is the length of the input braid word , and @xmath1 in a constant which appears to be growing linearly with the number of strands @xmath8for four strands , the choice @xmath24 seems sufficient , and as the example of the words @xmath25 ( with @xmath26 according to the parity of @xmath8 ) demonstrates , @xmath1 needs to grow _ at least _ linearly with @xmath8 .
now , since handle reduction is a compound of reversing and monotone equivalence , an affirmative answer to question [ q : main ] would have given the expected linear bound for the length of the words appearing in handle reduction .
as a corollary , it would have shown that , for fixed @xmath8 , every braid word of length @xmath2 is equivalent to a @xmath22-definite word of length @xmath27 .
let us mention that the latter statement has been proved recently in @xcite using a deep result about train tracks @xcite .
the current results leave the questions about handle reduction open .
however , handle reduction is in fact a compound of a more restricted set of operations , namely reversings and commutation relations , so we would be satisfied if the length of words remained bounded under iterated applications of these two operations this is exactly the modified conjecture stated at the end of the paper . |
it is generally believed that the top quark , because of its large mass , will be a sensitive probe into physics beyond the standard model ( sm ) @xcite .
in particular , its couplings to the gauge and higgs bosons may show deviations with respect to the sm predictions . in the sm
the flavour - changing neutral ( fcn ) couplings @xmath5 , with @xmath6 , vanish at tree - level by the gim mechanism , and the @xmath7 and @xmath8 ones are zero as a consequence of the unbroken @xmath9 symmetry .
the @xmath10 couplings also vanish due to the existence of only one higgs doublet .
these types of vertices can be generated at the one - loop level , but they are very suppressed by the gim mechanism , because the masses of the charge @xmath11 quarks in the loop are small compared to the scale involved . the single top production branching may be estimated roughly by @xmath12 @xcite , and the calculation of the branching ratios for top decays mediated by these fcn operators yields the sm predictions @xmath13 , @xmath14 , @xmath15 @xcite , @xmath16 @xcite , gev . ] and smaller values for the up quark . however , in many simple sm extensions these rates can be orders of magnitude larger . for instance , in models with exotic quarks @xmath17 can be of order @xmath18 @xcite .
two higgs doublet models allow for @xmath19 , @xmath20 , @xmath21 @xcite , and in @xmath22 parity - violating supersymmetric models one can have @xmath23 , @xmath24 , @xmath25 @xcite .
top fcn decays into a light higgs boson and an up or charm quark can also have similar or larger rates in models with exotic quarks @xcite , with more than one higgs doublet @xcite or with supersymmetry @xcite .
hence , top fcn couplings offer a good place to search for new physics , which may manifest if these vertices are observed in future colliders .
in addition , the study of fcn couplings provides model - independent information on the charged current couplings and the unitarity of the ckm matrix @xcite . here
we will focus on fcn interactions involving the top , a light charge @xmath26 quark @xmath27 and a neutral gauge boson @xmath28 . at present
the best limits on @xmath5 couplings come from lep 2 , @xmath29 @xcite , and the best limits on @xmath7 couplings from tevatron , @xmath30 @xcite .
they are very weak but will improve in the next years , first with tevatron run ii , and later with the next generation of colliders
. the cern lhc will be a top factory . with a @xmath31 production cross - section of 830 pb , at its 100 fb@xmath32 luminosity phase it will produce @xmath33 top - antitop pairs per year .
in addition , it will produce @xmath34 single tops plus antitops via other processes @xcite .
this makes lhc an excellent machine for the investigation of the top quark properties .
the search for fcn top couplings can be carried out examining two different types of processes . on the one hand
, we can look for rare top decays @xmath35 @xcite , @xmath36 @xcite , @xmath37 @xcite or @xmath38 @xcite of the tops or antitops produced in the sm process @xmath39 . on the other hand
, one can search for single top production via an anomalous effective vertex : @xmath40 and @xmath41 production @xcite , the production of a top quark without or with a light jet @xcite , and @xmath42 production @xcite . in these cases
the top quark is assumed to decay in the sm dominant mode @xmath43 .
one can also search for like - sign @xmath44 production @xcite and other exotic processes .
the tesla @xmath0 collider with a centre of mass ( cm ) energy of @xmath45 gev has a tree - level @xmath31 production cross - section of 0.52 pb , and produces only @xmath46 top - antitop pairs per year with its expected luminosity of 300 fb@xmath32 .
however , @xmath0 colliders are cleaner than hadron colliders . for instance , the signal to background ratio @xmath47 for rare top decays can be 7 times larger in tesla than in lhc .
but the sensitivity to rare top decays is given in the gaussian statistics limit by the ratio @xmath48 , and the larger lhc cross - sections make difficult for tesla to compete with it in the search for anomalous top couplings . in this paper we show that the use of beam polarization in tesla substantially enhances its sensitivity to discover or bound top anomalous fcn couplings and allows to improve some of the expected lhc limits up to an order of magnitude @xcite .
we first study the single top production process @xmath49 , mediated by @xmath5 or @xmath7 anomalous couplings @xcite
. then we study rare top decays in the processes @xmath50 , with subsequent decay @xmath51 . in all cases we take into account the charge conjugate processes as well : we sum @xmath52 production , and we consider @xmath51 or @xmath53 .
single top production and top decay processes are complementary .
single top production is more sensitive to top anomalous couplings but top decays can help to disentangle the type of anomalous coupling involved ( @xmath5 or @xmath7 ) if a positive signal is discovered .
we consider the planned cm energies of 500 and 800 gev , and for both we analyse the cases : ( _ i _ ) without beam polarization , ( _ ii _ ) with @xmath54 @xmath2 polarization , and ( _ iii _ ) with @xmath54 @xmath2 , @xmath55 @xmath3 polarization . the paper is organized as follows . in section [ sec:2 ] we describe the procedure used to compute the signals and backgrounds and to obtain the limits . in section [ sec:3 ]
we analyse single top production . in sections [ sec:4 ] and [ sec:5 ]
we consider top decays @xmath36 and @xmath35 , respectively . in section [ sec:6 ]
we summarize the results and draw our conclusions .
in order to describe the fcn couplings among the top , a light quark @xmath27 and a @xmath1 boson or a photon @xmath56 we use the lagrangian @xmath57 where @xmath58 .
the chirality - dependent parts are normalized to @xmath59 , @xmath60 , @xmath61 .
this effective lagrangian contains @xmath62 terms of dimension 4 and @xmath63 terms of dimension 5 .
the couplings are constants corresponding to the first terms in the expansion in momenta .
the @xmath63 terms are the only ones allowed by the unbroken gauge symmetry , @xmath9 . due to their extra momentum factor
they grow with the energy and make large colliders the best places to measure them .
for single top production we study the process @xmath49 mediated by @xmath5 or @xmath7 anomalous couplings ( see fig .
[ fig : feyn1 ] ) .
we will only take one anomalous coupling different from zero at the same time .
however , if a positive signal is discovered , it may be difficult to distinguish only from this process whether the anomalous coupling involves the @xmath1 boson , the photon or both . on the other hand , in principle it could be possible to have a fine - tuned cancellation between @xmath1 and @xmath64 contributions that led to a suppression of this signal . for top
decays we study the sm process @xmath50 , followed by antitop decay mediated by an anomalous @xmath5 or @xmath7 coupling ( see fig .
[ fig : feyn2 ] ) .
this gives the signals @xmath65 and @xmath66 , and the observation of the final state distinguishes @xmath5 and @xmath7 couplings . in the @xmath67 , @xmath65 and @xmath66 signals the top is assumed to decay via @xmath68 , with @xmath69 .
for the @xmath65 signal we only consider the @xmath1 boson decays to electrons and muons . for the @xmath70 signal
we calculate the matrix element @xmath71 . for the @xmath65 and @xmath66 signals
we calculate @xmath72 and @xmath73 , respectively .
these matrix elements are evaluated using helas @xcite and introducing a new helas - like subroutine iov2xx to compute the non - renormalizable @xmath63 vertex .
this new routine has been checked by hand . in all cases
we sum the contribution of the charge conjugate processes . for the @xmath74 signals
there is an additional contribution from @xmath70 production plus radiative emission of a @xmath1 boson or a photon .
this correction is suppressed because it does not have the enhancement due to the @xmath75 on - shell , and is even smaller after the kinematical cuts for the signal reconstruction .
the background for the @xmath70 signal is given by @xmath76 production with @xmath77 decay to electrons and muons .
the leading contribution to this process is @xmath78 production with @xmath79 hadronic decay , but it is crucial for the correct evaluation of the background after kinematical cuts to take into account the 7 interfering feynman diagrams for @xmath80 . taking all the interfering diagrams for @xmath81 into account does not make any appreciable change in the cross - section .
the backgrounds for the @xmath65 and @xmath66 signals are analogous , @xmath82 and @xmath83 , with 46 and 44 diagrams , respectively .
these three backgrounds are evaluated using madgraph @xcite and modifying the code to include the @xmath77 decay . to simulate the calorimeter energy resolution
we perform a gaussian smearing of the charged lepton ( @xmath84 ) , photon ( @xmath64 ) and jet ( @xmath85 ) energies using a realistic calorimeter resolution @xcite of @xmath86 where the energies are in gev and the two terms are added in quadrature . for simplicity
we assume that the energy smearing for muons is the same as for electrons .
note that more optimistic resolutions would improve our results .
we then apply detector cuts on transverse momenta @xmath87 and pseudorapidities @xmath88 @xmath89 the cut on pseudorapidity corresponds to a polar angle @xmath90 .
we reject the events in which the jets and/or leptons are not isolated , requiring that the distances in @xmath91 space @xmath92 satisfy @xmath93 .
we do not require specific trigger conditions , and we assume that the presence of high @xmath87 charged leptons will suffice .
after signal and background reconstruction , which will be analysed in detail for each of the processes discussed , we require a @xmath94 tag on the jet associated to the decay of the top quark to reduce the backgrounds .
we require the @xmath94 tagged jet to have @xmath95 ( polar angle @xmath96 ) and energy @xmath97 gev .
we assume a @xmath94 tagging efficiency of @xmath98 , and mistagging rates of @xmath99 for charm and @xmath100 for lighter quarks @xcite .
these are average numbers appropriate for the @xmath101 kinematical distributions we will obtain later .
after kinematical cuts , for each of the cases studied we obtain two types of limits on the anomalous coupling parameters @xmath102 , @xmath103 , @xmath104 .
below we outline the procedure used .
the correct statistical treatment of signals and backgrounds is specially necessary in our study since the backgrounds are very small , sometimes much less than one event even for high integrated luminosities . assuming that no signal is observed after the experiment is done , _ i.e. _ the number of observed events @xmath105 equals the expected background @xmath106 , we derive @xmath107 confidence level ( cl ) upper bounds on the number of events expected @xmath108 .
we use the feldman - cousins construction for the confidence intervals of a poisson variable @xcite evaluated with the pci package @xcite . on the other hand , we can obtain the smallest value of @xmath108 such that a positive signal is expected to be observed with @xmath109 significance , assuming that the number of observed events for @xmath109 ` evidence ' @xmath110 equals @xmath111 . for a large number of background events , the poisson probability distribution can be approximated by a gaussian of mean @xmath106 and standard deviation @xmath112 .
the requirement of @xmath109 significance is then simply @xmath113 .
however , this is seldom the case for our study , where the backgrounds are very small . in such case
, we use the estimator based on the @xmath114 number ( see for example @xcite ) . the number @xmath115 is defined as the probability of the background to fluctuate and give @xmath116 or more observed events .
@xmath110 is then defined as the smallest value of @xmath116 such that @xmath117 , corresponding to three gaussian standard deviations .
another possible estimator for the evidence of a signal can be built using feldman - cousins intervals . for a fixed value of @xmath106
, we define the number of observed events for @xmath109 evidence @xmath110 as the smallest value of the number of observed events @xmath116 such that the @xmath118 cl feldman - cousins intervals do not contain zero .
these two estimators for the evidence of a signal can be shown to be equivalent , and for @xmath119 both give similar results to the gaussian approximation @xmath113 . in our analysis
we find that usually @xmath109 evidence limits are numerically larger than @xmath107 upper limits , but this is not always the case .
for very small backgrounds @xmath109 limits are smaller , what means that the potential to discover a new signal is better than the ability to set upper bounds on it if nothing is seen .
this behaviour can exhibit fluctuations resulting from the discreteness of poisson statistics , but in general and comparing with lhc , the tesla discovery potential is better than the potential to set upper limits .
the process @xmath49 gives better limits on the top anomalous couplings than top decays . however , it has the disadvantage that the final state @xmath120 does not distinguish the type of coupling involved .
the background is @xmath121 production with @xmath122 and a jet misidentified as a @xmath94 .
we take only one type of fcn couplings different from zero at the same time , and we evaluate three signals : ( _ i _ ) with @xmath5 @xmath62 , ( _ ii _ ) with @xmath5 @xmath63 , and ( _ iii _ ) with @xmath7 couplings .
their cross - sections depend slightly on the chirality of the anomalous couplings .
the chirality - dependent parts can be written as @xmath123 , with @xmath124 .
for definiteness , we set @xmath125 , @xmath126 in our evaluations .
the results are the same setting @xmath127 and @xmath128 . for a cm energy of 500 gev and the three polarization options discussed the cross - section for @xmath62 couplings
differs @xmath129 setting @xmath130 , and for @xmath63 couplings it differs @xmath131 .
the signals are reconstructed as follows .
the neutrino momentum @xmath132 can be identified with the missing momentum of the event
. the longitudinal missing momentum can also be used , and @xmath132 is reconstructed without any ambiguity .
the @xmath77 momentum is then the sum of the momenta of the charged lepton and the neutrino . in the case of @xmath70 production ,
the invariant mass of the @xmath77 and one of the jets , @xmath133 , is consistent with the top mass , and the other jet has an energy @xmath134 around @xmath135 .
of the two possible pairings , we choose the one minimizing @xmath136 and require a @xmath94 tag on the jet associated to the top quark .
the kinematical distributions for @xmath133 and @xmath134 for the signals and the background at a cm energy of 500 gev without beam polarization are plotted in figs .
[ fig : tc - mt ] and [ fig : tc - ec ] . the @xmath101 distribution is shown in fig .
[ fig : tc - eb ] .
another interesting variable to distinguish the signals from the background is the two - jet invariant mass @xmath137 .
the @xmath138 background is dominated by @xmath78 production with @xmath139 , and the @xmath137 distribution peaks around @xmath140 , as can be clearly seen in fig .
[ fig : tc - mw2 ] .
a veto cut on @xmath137 can eliminate a large fraction of the background but makes compulsory to calculate correctly the cross - section to include all the diagrams for @xmath80 . also of interest
are the total transverse energy @xmath141 and the charged lepton energy @xmath142 in figs .
[ fig : tc - ht ] and [ fig : tc - el ] .
the kinematical distributions with polarized beams are very similar except the @xmath141 distribution . in this case polarization
decreases the peak of the background around @xmath143 . to enhance the signal significance we perform kinematical cuts on these variables
however , we find that the veto cut on @xmath137 is unnecessary in single top production since the requirement @xmath144 gev and the kinematical cut on @xmath133 practically eliminate the peak in the @xmath137 distribution .
a cut on @xmath134 is unnecessary because this variable is kinematically related to @xmath133 , and we prefer to apply a cut on @xmath133 to show the presence of a top quark in the signal . for simplicity
, we choose to apply the same cuts for the three signals and the three polarization options , but different for cm energies of 500 and 800 gev .
we choose the cuts trying to maintain the independence of the cross - section on the chirality of the coupling .
obviously , our results could be improved modifying the cuts for each type of coupling and each polarization option .
we now discuss the results for 500 gev and 800 gev in turn .
the kinematical cuts for 500 gev are collected in table [ tab : tc1 ] , and the cross - sections before and after cuts for the three polarization options in table [ tab : tc2 ] .
we normalize the signals to @xmath145 , @xmath146 , @xmath147 and sum the charge conjugate processes . for the chiralities with @xmath148 the cross - section after cuts differs @xmath149 for @xmath62 couplings and @xmath150 for @xmath63 couplings .
we see that lhc and tesla complement each other in the search for top fcn vertices .
the @xmath62 couplings to the @xmath1 boson can be best measured or bound at lhc , whereas the sensitivity to the @xmath63 ones is better at tesla . for photon vertices , lhc is better for @xmath151 and tesla for @xmath152 .
the complementarity of lhc and tesla also stems from the fact that tesla will not be able to distinguish @xmath5 and @xmath7 couplings in the limit of its sensitivity , whereas lhc will because final states are different and distinguish between them . on the other hand
, the good charm tagging efficiency expected at tesla will be able to distinguish @xmath153 and @xmath154 couplings looking at the flavour of the final state jet , what is more difficult to do at lhc .
we thank f. del aguila , k. mnig and a. werthenbach for a critical reading of the manuscript and s. slabospitsky for useful comments .
jaas thanks the members of the theory group of desy zeuthen for their warm hospitality during the realization of this work .
this work has been supported by a daad scholarship and by the european union under contract htrn ct200000149 and by the junta de andaluca . 99 for
a review see m. beneke , i. efthymipopulos , m. l. mangano , j. womersley ( conveners ) _ et al .
_ , report in the _ workshop on standard model physics ( and more ) at the lhc _ , geneva , hep - ph/0003033 g. mann and t. riemann , annalen phys . * 40 * , 334 ( 1984 ) ; j. i. illana , m. jack and t. riemann , tesla physics note lc th2000007 ( 2000 ) , hep - ph/0001273 ; j. i. illana and t. riemann , phys
* d63 * , 053004 ( 2001 ) g. eilam , j. l. hewett and a. soni , phys .
rev . * d44
* , 1473 ( 1991 ) b. mele , s. petrarca and a. soddu , phys . lett . *
b435 * , 401 ( 1998 ) f. del aguila , j. a. aguilar - saavedra and r. miquel , phys .
lett . * 82 * , 1628 ( 1999 ) d. atwood , l. reina and a. soni , phys . rev . *
d55 * , 3156 ( 1997 ) j. m. yang , b. young and x. zhang , phys . rev .
* d58 * , 055001 ( 1998 ) f. del aguila and m. j. bowick , nucl . phys .
* b224 * , 107 ( 1983 ) ; k. higuchi and k. yamamoto , phys .
* d 62 * , 073005 ( 2000 ) s. bejar , j. guasch and j. sola , hep - ph/0011091 j. yang and c. li , phys . rev . * d 49 * , 3412 ( 1994 ) ; j. guasch and j. sola , nucl
b562 * , 3 ( 1999 ) ; g. eilam , t. han , j. m. yang and x. zhang , hep - ph/0102037 f. del aguila , m. prez - victoria and j. santiago , phys .
b * 492 * , 98 ( 2000 ) ; f. del aguila , m. prez - victoria and j. santiago , jhep * 0009 * , 011 ( 2000 ) v. obratzsov , talk given at ichep 2000 , osaka , july 2000 aleph collaboration , cern - ep-2000 - 102 , phys .
* b * ( in press ) ; see also the contributed papers of the delphi collaboration , l3 collaboration and opal collaboration to ichep 2000 f. abe _
[ cdf collaboration ] , phys .
lett . * 80 * , 2525 ( 1998 ) t. stelzer , z. sullivan and s. willenbrock , phys . rev . *
d58 * , 094021 ( 1998 ) t. m. tait and c. p. yuan , hep - ph/9710372 ; t. m. tait , phys . rev . * d61 * , 034001 ( 2000 ) t. han , r. d. peccei and x. zhang , nucl
b454 * , 527 ( 1995 ) t. han , k. whisnant , b .-
young and x. zhang , phys . rev . * d55 * , 7241 ( 1997 ) t. han , k. whisnant , b .-
young and x. zhang , phys . lett .
* b385 * , 311 ( 1996 ) j. a. aguilar - saavedra and g. c. branco , phys . lett . *
b495 * , 347 ( 2000 ) f. del aguila , j. a. aguilar - saavedra and ll .
ametller , phys . lett .
* b462 * , 310 ( 1999 ) ; f. del aguila and j. a. aguilar - saavedra , nucl
b576 * , 56 ( 2000 ) m. hosch , k. whisnant and b .-
young , phys . rev .
* d56 * , 5725 ( 1997 ) t. han , m. hosch , k. whisnant , b .-
young and x. zhang , phys . rev . *
d58 * , 073008 ( 1998 ) y. p.
gouz and s. r. slabospitsky , phys . lett . *
b457 * , 177 ( 1999 ) j. a. aguilar - saavedra , hep - ph/0012305 , phys . lett .
* b * ( in press ) v. f. obraztsov , s. r. slabospitsky and o. p. yushchenko , phys . lett . *
b426 * , 393 ( 1998 ) ; t. han and j. l. hewett , phys . rev . * d60 * , 074015 ( 1999 ) ; s. bar - shalom and j. wudka , phys . rev . *
d60 * , 094016 ( 1999 ) e. murayama , i. watanabe and k. hagiwara , kek report 91 - 11 , january 1992 t. stelzer and w. f. long , comput .
commun . * 81 * , 357 ( 1994 ) s. bertolucci , talk `` calorimetry '' , given at the _ ` 5th workshop of the 2nd ecfa / desy study on physics and detectors for a linear electron - positron collider ' _ , obernai 16 - 19 october , 1999 ; see : + http://ireswww.in2p3.fr/ires/ecfadesy/talks/bertolucci/bertolucci.pdf c. damerell , talk `` vtx concept '' given at the _ ` 7th workshop of the 2nd ecfa / desy study of physics and detectors for a linear electron - positron collider ' _ , desy hamburg , 22 - 25
september , 2000 ; see : + http://www.desy.de/@xmath155ecfadesy/transparencies/det_damerell.pdf g. j. feldman and r. d. cousins , phys . rev . * d57 * , 3873 ( 1998 ) j. a. aguilar - saavedra , comput .
. commun . * 130 * , 190 ( 2000 ) g. cowan , statistical data analysis , oxford university press , 1998 | we present a comprehensive analysis of the sensitivity of the tesla @xmath0 collider to top flavour - changing neutral couplings to the @xmath1 boson and photon .
we study single top production and top decay processes , and we consider the cases without beam polarization , with only @xmath2 polarization and with @xmath2 and @xmath3 polarization .
we show that the use of the latter substantially enhances the sensitivity to discover or bound these vertices , and for some of the couplings the expected lhc limits could be improved by factors @xmath4 for equal running times .
ft128/01 + lc
th2001067 + hep - ph/0102197 + february 2001 + * probing top flavour - changing neutral couplings at tesla * j. a. aguilar
saavedra + _ departamento de fsica terica y del cosmos , universidad de granada + e-18071 granada , spain _ + t. riemann + _ deutsches elektronen - synchrotron desy + platanenallee 6 , d-15738 zeuthen , germany _ + |
Subsets and Splits